Building a Pinwheel Skeleton Clock

Discussion in 'Clock Construction' started by Allan Wolff, Nov 12, 2010.

  1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
  1. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    This is the first installment on building a tabletop size pinwheel skeleton clock. This introduction may be a little dry, but I promise to include exciting action shots of actual clock parts being made in future updates.

    Since I doubt many will actually build this clock, I do not intend to provide dimensioned drawings as part of this build. However, I will provide them on request as they become available if you send me a PM. The reason for this is simple copyright. I have seen many offers for books and plans for sale on auction sites where the seller simply copied other people's work from the Internet and burned a CD. I am happy to give the plans away; it just bugs me to see someone trying to profit from someone else's hard work. My form of copyright is simple. Along with my name, I will add your name and email address to the footer of each page and send you your very own personalized PDF file. So with that said, let's get started.

    I saw this clock on the cover of the 25th anniversary edition of "The Art and Craft of the Clockmaker" published by Derek Roberts Antiques. Its symmetry and graceful curves really caught my eye. Building the clock will present a number of new clock building challenges for me. I have never silvered a dial, cut solid steel pinions and I currently have no idea how to make the brass pedestal. I will figure them out when the time comes. I'm sure to learn a lot in the process. This will also be the first clock constructed without the aid of a “how-to” book. All dimensions will be estimated from photos or obtained through the design process utilizing CAD drawings and test builds.

    Tools and Construction Methods
    No exotic or expensive equipment is used to construct this clock. All machine work is done with only a bench grinder, drill press and Taig lathe. The lathe is equipped with a second head for milling, but does not have CNC or threading capability. The dividing plates and cutters used for machining wheels and ratchets are all homemade. The remainder of the work is done by hand.

    Standard materials will be used to construct the clock, primarily brass and steel. Every attempt will be made to make all of the required parts; within reason. Plate screws and other pieces that are obviously visible will be constructed, but there is no reason to spend countless hours making 0-80 screws or a fusee chain that are simply impractical to make in the home shop. I applaud any purists that wish to pursue this level of detail.


    Most clock construction books follow a similar sequence of construction. The frame is usually constructed first and then the mechanism is built progressing up the train from the mainspring to the escapement. Although this is most convenient in terms of construction, it is doubtful that the original clock was built in that order. This project will follow the order of my actual construction, starting with the escapement.

    The Escapement
    The original clock used a unique spring pallet escapement but I have decided to use a pinwheel arrangement. Having never built or even owned a pinwheel clock, the escapement will be built first to prove the design actually works and provide an indication as to the amount of power needed to drive the clock. From this the mainspring size and fusee dimensions can be derived.

    Typical pinwheel arrangements place all of the pins on the same side of the wheel and stagger the pallets using different length arms to obtain the necessary clearance and drop. On this design, the pins will be located on alternating sides of the wheel allowing the pallet arms to be equal in length. Although uncommon today, some of the very first pinwheel escapements used this staggered arrangement of the pins.

    The Train
    Several factors influence the train design. A tabletop size requires the use of a short, half-seconds pendulum resulting in a beat rate of 120 beats per minute; or 7200 beats per hour. Additionally, the clock should run for 8 days on a single wind. Finally, pinions with 12 or more leaves are desirable to reduce power loss. Experience with previous clock construction and repair projects showed that any imperfections with low count (typically 8) leaf pinions causes a tremendous loss of power as the pinion and wheel mesh at a single point. 12 leaf or higher pinions should have two points of contact and thus smoother operation. A spreadsheet was used to develop several gear combinations. None of the combinations could achieve the required runtime with a three-arbor system without using high tooth count wheels. (My index plate only goes up to 96) The four-arbor combination selected used the following wheels and pinions.

    Escape: Wheel=50; Pinion=12; RPD=3456; RPH=144
    Fourth: Wheel=72; Pinion=12; RPD=576; RPH=24
    Third: Wheel=72; Pinion=12; RPD=96; RPH= 4
    Center: Wheel=72; Pinion=12; RPD=16; RPH=.67
    Main: Wheel=96; RPD=2

    An appealing characteristic of this train is all of the pinions and three of the wheels are the same size allowing common parts to be machined together. This is especially convenient since the pinions will be cut from a single setup rather than individually constructed as lantern style pinions. A 0.8 module wheel and pinion size results in a time train that fits nicely within the allowable space of the frame.

    A Computer Aided Design (CAD) application was used to create a side and front view of the time train. See the attached drawing. Dimensions are obtained from this scale drawing to ensure the train will fit within the frame. On the front view, simple circles represent wheels and pinions with red for the pitch circle diameter (PCD) and black for the wheel blank, or tooth tip diameter. Green represents the motion works that is mounted on the outside of the front plate. On the side view, rectangles represent the wheels, pinions and arbors. This view is convenient for verifying the wheels will clear the arbors and also assists with general layout of the wheels and pinions on their respective arbors. A few important dimensions are called out on the drawing such as the distance from the winding arbor to dial center and overall time train length. The distance of the pallet arbor from the escape arbor is determined geometrically and will be discussed in the next installment when the escapement is built.

    You may have noted that a key requirement for all clock designs has not yet been addressed; the minute hand must rotate once per hour and preferably in a clockwise direction. Typically, the minute hand is driven directly by one of the arbors. For this design, the third arbor will extend through the front plate and hold a 13-leaf pinion. The pinion will drive one of the 52-tooth motion works wheels and divide the third arbor’s 4 revolutions per hour (RPH) to the necessary 1 RPH. This appears to be similar to the design of the original clock.

    The motion works utilizes the traditional English design that achieves the 12:1 ratio through a 96-tooth wheel driven by an 8-leaf pinion. Since the pinion drives the wheel, power loss is minimal and is not a concern like it is in the time train where the wheel drives the pinion. A .5 module gear size keeps the dimensions of the motion works wheels and pinions to a manageable level.

    Sorry to run so long, but I wanted to get all of this preliminary stuff out of the way so we can start making parts. I will try to post a new installment each week. Comments and questions are certainly welcome.

    Allan Wolff 76084.jpg
     
  2. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Pinwheel Skeleton Clock - Introduction

    After that long-winded post, I will keep this one short. As I stated in another thread, I am over a year into the build on this clock, but am presenting the construction in the order that it occurred. To wet your appetite and give you a better idea of the escapement used in this clock, I have posted a short video of it in operation on Youtube. Here is a linkhttp://www.youtube.com/watch?v=VaTW11mBPj4. Or just go to Youtube and search for pinwheel escapement.

    Allan
     
  3. Scottie-TX

    Scottie-TX Registered User
    Deceased

    Apr 6, 2004
    936
    36
    0
    Retired
    Mesquite, TX
    Country Flag:
    Region Flag:
    Re: Pinwheel Skeleton Clock - Introduction

    I applaud you on your choice of pinwheel escapement - perhaps one of the easier ones to design and fabricate. I say that not from experience! How did you decide on pallet arm length? Pallets appear to be very thin and perhaps a small curvature for locking? Is there an impulse face on the pallets or does all the impulse come from the pins, like a brocot?
     
  4. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Pinwheel Skeleton Clock - Introduction

    I will get into the details of building the escapement in the next installment, so this is a great lead-in. I drew up an escapement with some dimensions that fit the tools I had available and then calculated the remaining dimensions to satisfy the mechanical requirements. That sounds rather arbitrary so let me explain further.

    I used a 3" diameter bench block as a form to bend the pallet. That set the pallet arm length equal to the radius of the bench block or 1.5". The second variable is the distance from the center of the escape wheel to the pins. This distance is 1" since I selected a 2" diameter escape wheel. Both of these dimensions, 1.5" and 1", are somewhat arbitrary as long as you don't get too carried away and make the wheel too big or pallet arms too long that they do not fit in the space available.

    For a deadbeat escapement, the pins need to contact the pallet at a 90-degree angle to the rotation of the escape wheel. This angle is determined by the distance between the escape arbor and pallet arbor. This distance is the hypotenuse of a right triangle so it equals the square root of 1.5" squared + 1" squared or 1.8". (did I mention there would be some math involved in this project? :=) A drawing is attached to help explain. That is the theory. In practice, the exact length of the pallet arm is difficult to control, so the length of the finished arms are measured and that dimension is used in the calculation.

    To answer your other questions, the pallets are 1/16" thick and the pallets are curved to match the pallet arm length. The pallets do have an impulse face which requires even more math!
    Thanks,
    Allan 76177.jpg
     
  5. Scottie-TX

    Scottie-TX Registered User
    Deceased

    Apr 6, 2004
    936
    36
    0
    Retired
    Mesquite, TX
    Country Flag:
    Region Flag:
    Re: Pinwheel Skeleton Clock - Introduction

    I appreciate your explanation and don't mind the math until differentials and integrals are broached.
    So then it would seem (?) some impulse is provided by the impulse faces and some impulse provided by pin radius? What impulse angle did you choose - two degrees?
     
  6. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Pinwheel Skeleton Clock - Escapement

    The escapement for this clock will be built first. The style of pinwheel used is rather uncommon in that the pallet arms are equal in length and the pins are placed on alternate sides of the escape wheel. By making the pallet arms the same length, the impulse on each pallet is equal. Pinwheel escapements with unequal pallet arms technically have an imbalanced impulse, although the imbalance is not likely to affect the operation or accuracy of the clock. Alternating the pins on each side of the escape wheel does provide a substantial improvement in the amount of drop available and reduces the need for extremely tight tolerances when making the escapement.

    Two pallet arms are cut from 1/16" O-1 oil hardening ground flat stock. Each arm is cut slightly oversize with a jeweler’s saw and filed to shape. The rounded portion can be filed by eye, but better results are obtained by using filing buttons. The filing buttons are like washers made from a piece if 1/2" drill rod with a center hole the same size as the pallet arms. The buttons are then hardened so a file skates across the surface. The arms are clamped in a vise between the buttons with a drill bit used as an alignment pin and filed all around until the file reaches the buttons. The result is a nice round shape. Save the buttons as they can be reused on future projects.

    The surfaces of the pallet arms are smoothed at this time to a 600-grit finish. Some touch up may be needed to remove marks when the tabs are bent, but it is easier to prepare the surface as much as possible while the part is flat and easy to hold.

    The tabs of each pallet arm are now bent to form the pallet surfaces. This is accomplished by using a 3" diameter round steel bench block as a bending form. An adapter is made that fits the center hole of the bench block and cut to length such that the end sits 1/16" above the bench block surface. A pip 1/16" high is then machined to a diameter that fits snugly into the center hole of the pallet arm. The adapter fits somewhat loosely into the bench block hole. The tab on the pallet arm must be heated before bending due to the sharp radius bend required. To prevent the bench block from absorbing heat away from the pallet arm while being heated, a spacer is placed between the pallet arm and bench block as shown in the photo.

    The following sequence must be done quickly because the pallet arm will cool rapidly. The tab and lower end of the arm are heated to a dull red with a propane torch. The spacer is then removed and the pallet arm is pushed against the bench block with a block of wood. The tab is then hammered over the edge of the bench block with a rawhide mallet. Rawhide is used to minimize marks in the soft, hot metal. The tab is hammered so it conforms to the round contour of the edge of the bench block. It was necessary to reheat the tab to complete the shaping process as the bench block quickly absorbs heat from the pallet arm. The bend area is then carefully inspected. It will be necessary to re-make the part if any cracks are found.

    The next step will be to grind the pallet face. 76646.jpg 76647.jpg 76648.jpg 76649.jpg
     
  7. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Pinwheel Skeleton Clock - Escapement

    With the pallet bent, the 45 degree impulse face can now be cut. It is necessary to compensate for the additional angle from the curved tab which works out to be approximately 9 degrees. So, the total angle from the pallet arm mounting hole to the tip of the impulse face is 54 degrees.

    A jig held in the lathe tool post holder is set so the pallet arm is 54 degrees offset from the side of a grinding wheel. A protractor is used to set the jig to the correct angle. The jig consists of a piece of aluminum with two pins. One pin fits through the pallet arm center hole and acts as a pivot. The other pin stops the arm from rotating. The lathe bed is covered to protect it from grinding dust. The pallet is moved into contact with the grinding wheel using the carriage until approximately ¾ of the impulse face is ground. The face is checked to make sure it is being cut straight and adjustments are made as necessary. Grinding and checking continues until a sharp edge is obtained. The same operation is repeated for second pallet.

    The impulse face is then polished with progressively finer grades of sandpaper. The sandpaper is glued to a round disk and spun in the lathe similar to the grinding wheel. 600-grit paper will leave a mirror finish on the impulse face. The curved top of the pallet is polished by holding sandpaper against the side of the bench block and rubbing the pallet back and forth against the sandpaper with the round end centered by the same adapter used when the pallet was bent. Sanding in this method retains the correct radius on the top of the pallet.

    The pallet is then hardened by heating to bright orange and quenching in oil. The pallets are then immediately tempered by placing them in the kitchen oven at 350° F. for one hour. The remainder of the pallet is then sanded all over to a 600-grit finish. Final polishing to a mirror finish will be done after the clock is completely assembled. Automobile wax is applied to the entire pallet arm except the top of the pallet and impulse face, which should receive a light coat of clock oil. The wax will prevent any rust from occurring until the final polish. 76788.jpg 76789.jpg 76790.jpg
     
  8. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Pinwheel Skeleton Clock - Escapement

    The escape wheel is fairly simple to construct. The main consideration is to be sure the outside edge and holes for the pins are concentric with the center hole. A disk of 1/16" brass is turned to a diameter of 2.156". A milling spindle is mounted on the cross slide and offset 1" from the center of the wheel. The milling spindle is simply a second Taig headstock that is driven with a flex shaft by the lathe motor. An index plate is mounted to the lathe headstock and holes are spot drilled around the disk with a small center drill bit. See photo. Each hole is then drilled through with a #54 bit. This bit should make a .055" diameter hole for a press fit of the music wire pins. A test hole can be drilled in a piece of scrap brass to verify the bit will drill the proper size hole for a tight fit of the pins. The milling spindle is then reset for a 3/16" offset from center and 3 holes are spot drilled to attach the collet. The wheel is then removed and the 3 holes are drilled through with a #50 drill bit.

    The wheels spokes are next crossed out with a jewelers saw. I used 5 spokes, but any number can be used. Sanding and polishing the wheel will be very difficult once the pins are installed, so I polished and lacquered the wheel before inserting the pins. This required careful handling of the finished wheel to prevent damaging the lacquer.

    A length of music wire is next prepared by sanding with 600 and 1200-grit paper followed by Simichrome polishing paste. A Dremel cutoff wheel is mounted and the tool is placed in the bench vise with the wheel .220" from the vise. The end of the wire is rounded by rotating it against the side of the cutoff wheel. With the rounded end of the wire placed against the vise, the wire is rotated against the edge of the cutoff wheel to cut a groove in the wire. Do not cut completely through or the pin will likely be lost. See photo. The pin is then snapped off. Each pin is then set aside and the next pin is rounded and cut. 50 pins are needed, but a few extra pins are made just in case some are lost. A pin vise is then used to hold each pin while pressing the back side of each pin against the cutoff wheel to remove the pip. The pin is the rotated while the edge of the pin is lightly touched against the cutoff wheel to remove any burrs and to produce a slight bevel to aid insertion into the escape wheel.

    The escape wheel is laid flat on a soft wood block and the pins are inserted into every other hole on one side of the escape wheel. The pins are tapped in with a small hammer. A touch of Loctite can be used to ensure the pins do not come loose. The wheel is flipped over and positioned near a corner of the block so 4 or 5 pins are on the block. The position of the pins are marked on the block and the block is then drilled for each pin to allow the wheel to lay flat on the block. See photo. The remaining pins are then inserted in the other side of the wheel.

    The wheel and crutch collets will be made next. 77474.jpg 77476.jpg 77477.jpg
     
  9. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Pinwheel Skeleton Clock - Escapement

    The escape wheel collet is machined from 1/2" brass rod. The end is faced and the center arbor hole is spotted with a small center drill. The hole is drilled a little over 1/2" deep with a #37 drill bit and then reamed to the final dimension of .106". The end of the rod is reduced to form a spigot that fits into the escape wheel center hole. The wheel is test fitted to the collet often while small cuts are taken to achieve a snug fit. A left hand lathe tool or graver is used to clean out the corner if the spigot and make the hub slightly concave so that only the outer edge of hub contacts the wheel. The sleeve portion of the collet is then formed with multiple plunge cuts with a parting tool. The depth of these cuts should leave the sleeve slightly larger than the final dimension so a continuous clean up cut can be made with the parting tool to achieve a uniform finish. The back edge of the hub is rounded over with a file.

    An index plate is mounted on the headstock and the milling spindle is mounted on the cross slide. The spindle is offset 3/16". The wheel mounting holes are then spot drilled and then drilled through with a #56 drill bit. The drill bit is then replaced with an 0-80 tap and the spindle is turned by hand to tap the holes. The first photo shows the holes being tapped and the final shape of the collet. The collet is then finish to a 600-grit finish before parting off.

    The crutch collet is machined in the same manner as the escape collet except it has no wheel spigot and only 2 mounting screw holes are drilled and tapped.

    The pallet and escape arbors are machined to the same dimensions. These are made from blued pivot steel or music wire. A collet is recommended for holding the work as the arbor will need to be removed for measuring and machining the 2.375" distance between the shoulders. If a collet of this size is not available, a bushing can be made from a 1/2" length of brass rod with a center hole drilled and reamed to .106". A slot is then cut down the side of the bushing to allow it to squeeze around the work piece. The taig lathe collet, bushing and homemade reamer is shown in the second photo. This bushing is fine for driving the arbor but may not run perfectly true. To ensure the pivot runs true, a steady rest is used to hold the end of the arbor as it is machined. After the pivot is machined to a diameter of .051", the corner of the pivot is cleaned out with a pointed graver and reduced it to the final .049" dimension with a pivot file or very fine (#6) flat file.

    The pivot is polished with a fine Arkansas stone and oil as shown in the third photo and then burnished to bring up a mirror finish and work harden the pivot. The arkansas stone is white and makes it easy to see the gray trail left as material is removed from the pivot. An even trail indicates the pivot is perfectly flat. The second pivot is turned in the same way with the shoulder spacing as close to 2.375" as possible, but not less. The arbor is removed periodically and checked with a dial caliper. When the proper distance of 2.375" between the pivots is achieved, the second pivot is brought to final size, polished and burnished. 77952.jpg 44764.jpg 77954.jpg
     
  10. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Pinwheel Skeleton Clock - Escapement

    The pallet spacer serves several key purposes. It holds the pallet arms at the correct distance from the escape wheel and also positions the pallet arms to obtain the necessary amount of drop for the escapement. A 1" length of 1/2" diameter brass rod is chucked in the lathe. The end is faced and a cleaning pass is made over the outside. The center hole is center drilled, drilled and reamed to .106" for a slip fit over the pallet arbor. The surface of the spacer is then finished to 600-grit. The spacer is parted off at .456" in length. This distance was determined from actual measurements of each pallet width (approximately .134" each. This will vary depending on the bend radius of the pallet tab, thus measurement is required.) the thickness of the escape wheel (1/16") and clearance between the pallets and escape wheel ( 1/16" each side). Using a short length of .106" drill rod as a guide, one of the pallet arms is positioned on the end of the spacer and the two mounting holes are transferred from the pallet arm to the spacer. The pallet arm is then removed and the spacer holes are drilled to approximately 1/4" deep with a #56 drill. Both holes are tapped 0-80. The mounting holes on the other side of the spacer will be drilled based on the escapement test later. The spacer is shown in the third photo between the two pallet arms.


    The crutch is formed from a strip of 1/16" brass sheet. A jeweler’s saw is used to cut the outline. As with the pallet arms, the crutch is filed to shape and sanded to a 600-grit finish while it is still flat. The crutch is clamped in the bench vise with the rounded end and a 5/16" length of the arm exposed. The vise is fitted with copper jaw liners to prevent marring the brass. The exposed part of the crutch is bent forward using a wooden block and hammer to form a tight bend. The crutch is raised 1/2" and the exposed part is bent back in the same way. Finally, the crutch is raised in the vise again so all but the foot is exposed and bend the part forward. The dimensions of the bent crutch arm are not critical and the finished crutch may vary slightly without causing any problems. I fact, The crutch shown in the first photo will be replaced with an adjustable one to make it easier to set the beat.

    This completes construction of all of the parts needed to test and adjust the escapement. A temporary assembly is constructed from scrap material. The only dimensions that require particular attention are the plate spacing and the distance between the pallet and escape arbors. The distance between the plates should be 2.390" to accommodate the 2.375" arbors with an additional 1/64" for end shake.

    The arbors are spaced so the escape wheel pins contact the pallets perpendicular with the pallet arms. Given that the pins are 1" from the escape arbor and the distance from the pallet arbor to the top of the pallet is 1.468". The distance between the arbors is the hypotenuse of a right triangle and its length is found by the square root of 1 squared plus 1.468 squared = 1.776". The test assembly is shown in the second and third photos.

    If the collets do not fit tightly on the arbors, they can be temporarily held in place with low-strength 222MS Loctite. The pallets and escape wheel are mounted near one end of their arbors with the unattached pallet arm nearest the plate. This allows a drill bit to clear the pallet arbor when spotting the remaining mounting holes. A 1/2" hole is drilled in one plate centered 1/2" below the pallet arbor to allow the crutch to project outside the plate. The initial alignment of the crutch and pallet arms is such that the crutch hangs straight down while an escape pin is centered on one of the the impulse faces. This alignment may need to be adjusted several times to put the escapement in beat. A temporary pendulum is made from a store-bought suspension spring and pendulum wire assembly approximately 9" in length with a brass weight attached to the end. The suspension spring is held by a piece of aluminum angle with a slot to accept the suspension spring. A plastic or nylon spacer provides a slop-free fit between the pendulum wire and the crutch foot.

    The pallets are aligned with the impulse faces overlapping approximately 1/3 of the way as shown in the fourth photo. More overlap provides more lock, but this re-quires more drive power and a wider pendulum arc. Less overlap produces a smaller pendulum arc and minimizes circular error, but the escape pins still need to land on the flat top of the pallet, not the impulse face.

    The remaining pallet arm is attached to the spacer with white Elmer’s glue. Super glue can be used, but the white glue offers some degree of adjustment before it completely sets up. The pallet arm is attached and the glue is allowed to set for approximately 15 minutes. A small amount of clock oil is applied to the pallet impulse faces and arbor pivots. The escapement is test run using light finger pressure on the escape arbor to drive it. The crutch and pallets are adjusted to obtain an even beat. The pallet arm alignment is adjusted so the escape pins just fall on the top of the pallet. Some over swing of the pendulum is expected. The glue is allowed to set up for several hours before proceeding.

    The escapement is run for several revolutions of the escape wheel to verify that all pins lock and drop correctly and the impulse is sufficient to keep the pendulum swinging. The pendulum arc should be fairly small. The pallet arbor is removed and the remaining mounting holes are transferred to the pallet spacer using the pallet arm as a guide. The pallet arm is then removed with a light tap on the side with a plastic or rawhide hammer. the mounting holes are then drilled #56 and tapped 0-80. The pallet arm is then attached and the escapement assembly reassembled and run again to verify it still runs as before. There should be enough slack in the pallet arm mounting screws to make slight adjustments to the pallet overlap if necessary.

    Now that the escapement design has been confirmed, the remaining parts of the clock can be constructed.
    The escapement can be seen running on this short video taken later in the construction process after some of the gearing had been completed.

    http://www.youtube.com/watch?v=VaTW11mBPj4 77962.jpg 77963.jpg 77964.jpg 77965.jpg
     
  11. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,666
    63
    0
    Calif. USA
    Re: Pinwheel Skeleton Clock - Escapement

    Hi
    Really nice work. How do you counter balance the pallets?
    I may have missed it in the quick reading.
    Tinker Dwight
     
  12. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Pinwheel Skeleton Clock - Escapement

    The pallets are quite light and no counter balance is provided at this time. However, you are correct in noting the need for a counter balance as the pallet assembly does tend to favor one side due to the offset configuration. I am thinking that when I remake the crutch to add a beat adjustment, the adjustment mechanism can be located on the opposite side of the crutch to help balance the assembly.

    Thanks for pointing out the balance issue.
    Allan
     
  13. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,666
    63
    0
    Calif. USA
    Re: Pinwheel Skeleton Clock - Escapement

    Hi
    The other thing I noticed was the crutch/pallet arbor
    centering relative to the pendulum. To minimize the
    slipping of the crutch along the penulum shaft, the
    centers of the arcs must be matched.
    A guess would be the center of the spring.
    It would be better to temporarily attach a piece of
    piano wire to the top of the solid pendulum shaft
    and watch or better yet photo graph the swing.
    The apex of the X formed by the wire at both ends
    of the swing should be the best location to put the
    arbor bearing.
    This would minimize any rubbing of the crutch.
    Tinker Dwight
     
  14. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Pinwheel Skeleton Clock - Escapement

    Tinker,
    Your observation of locating the pallet arbor at the bending point of the suspension spring is very good. For the test assembly, I put the pendulum in motion and "eye-balled" the bending point of the spring. The temporary suspension mount was then located to line up the bending point with the pallet arbor. The method you describe of attaching an extension to the top of the pendulum shaft sounds more accurate. I will give it a try when the time comes to locate the final mount.
    For the test assembly, a small Teflon bushing is used between the crutch and pendulum wire so a little rubbing is tolerated with very little friction. The final assembly will be brass on steel, so rubbing will need to be minimized.

    Thanks again for the tip. The primary purpose of posting this build is to share information; in both directions.
    Thanks,
    Allan
     
  15. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Cutting Pinions for the Pinwheel Skeleton Clock

    Lantern pinions were used on the 2 previous clocks that I built, but the original clock upon which this one is based used cut pinions so I thought I would give it a try.

    A 12-leaf pinion requires a cutter designed specifically for this module and leaf count. The homemade cutter was built according to the method described by David Creed as posted on the Yahoo MLhorology group. The cutter was mounted on a second Taig headstock used as a milling spindle held with the milling attachment on the crosslide similar to the configuration shown in the first photo. Note that this photo shows a wheel being cut, but the configuration is the same. The variable speed lathe motor was used to drive the second headstock via a 1/4" flex shaft; the orange item in the photo. The motor must be started slowly to prevent the flex shaft from twisting up. The tailstock could not be used to support the end of the rod with the milling attachment in place, so a 1" diameter brass bar served as a makeshift support as shown in the second photo. A test run was made with a piece of 1/2" diameter hot rolled mild steel rod. After reducing the diameter to the proper dimension, slots are cut with a 1/32" slitting saw to the full depth of the tooth in a single pass. The pinion cutter was then run through the slot at full depth producing a nicely formed leaf. A successful but misleading test!

    The final pinions need to be hardened for optimum durability, but the hot rolled steel does not contain enough carbon to be sufficiently hardened. Therefore, a piece of oil hardening drill rod was used for the actual pinions. Even in its annealed state, drill rod is much harder than mild steel. After cutting approximately 1/2" of the first slot, the slitting saw lost its edge and was ruined. The pinion cutter was then run down the slot and quickly became dull. The flex drive shaft also jumped and vibrated significantly. Time to re-group.

    A tachometer was temporarily attached to the milling spindle to check the speed of the cutters. It was found to be rotating at approximately 300 rpm. This speed was fine for the one-inch pinion cutter on mild steel, but much too fast for drill rod. The 1.75" diameter slitting saw needed to be slowed down to approximately 88 RPM and the pinion cutter should be run at approximately 150 RPM to achieve the proper surface feet per minute cutting speed. The milling spindle drive arrangement was also changed to correct the flex shaft problems. Since the Taig lathe is not heavy, it can easily be moved to obtain the configuration shown in the third photo. The 1" motor pulley drives the 4" flywheel of the milling spindle, providing a 4:1 speed reduction. The lathe motor is loosely fastened to the lower plywood board allowing the weight of the motor to keep the belt tight. The sequence of cutting the pinions was also changed. A slitting saw .025" wide is used to cut the initial slots in 2 passes, the first at 1/2 the depth and the second at a few thousandths shy of full depth. Lots of cutting oil was used for cooling and lubrication. The feed rate of the saw needs to be very slow due to power and rigidity restraints of the lathe, taking approximately 2 minutes for each 2" long cut. A new slitting saw is then used to widen the slots to .036" in a single pass. Excessive vibration prohibited making a full depth pass with the pinion cutter. A first pass was made at a depth of .075" and the second pass was then be made at the full depth of .091". After many hours of work, the result is a pinion rod approximately 2.5" inches long from which the individual pinions are cut. The pinion rod can be seen in the second photo. If I ever cut pinions again, this will be my excuse to buy some bigger machinery!

    The pinions will be finished in the next update. 78997.jpg 78998.jpg 78999.jpg
     
  16. leeinv66

    leeinv66 Moderator
    NAWCC Member

    Mar 31, 2005
    9,470
    116
    63
    Male
    Industry Specialist
    Launceston Tasmania
    Country Flag:
    Region Flag:
    Re: Cutting Pinions for the Pinwheel Skeleton Clock

    Thank you Allan, for this very interesting post! I am another lurker that has been following your posts, but have had nothing useful to add. To date I have built a few movements using donor gear trains from my parts bin. The idea of cutting my own wheels and pinions has always seemed beyond my abilities, but your post has started to change my mind. Thank you! Now if I could just get my head around setting up a cutting system like yours:)
     
  17. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Cutting Pinions for the Pinwheel Skeleton Clock

    Peter,
    My setup for cutting wheels and pinions is only one of many possible configurations. It was built on what I had at the time which was the Taig lathe. There is a video made by W.R. Smith called "Wheel Cutting, Pinion Making & Depthing for Clockmakers and Modelmakers" that can be purchased on Smith's website or rented from Smartflix. This video shows a number of different configurations that I found interesting and helpful towards understanding the process.

    When I built my first clock, I probably spent half of the time making tools to make parts for the clock; index plates, wheel cutters, etc. I made many mistakes (still do!) learned a lot (still learning) and always enjoy the process of creating something, even if it is sometimes just a a pile of swarf.
    Thanks,
    Allan
     
  18. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Cutting Pinions for the Pinwheel Skeleton Clock

    We now have a long pinion rod from which the actual shorter pinions are cut. Each pinion is drilled prior to removing it from the pinion rod. Three pinions 1/4" long are needed for the third, fourth and escape arbors. Each pinion is drilled separately to minimize the tendency of the drill bit to wander when drilling a deep hole. Each pinion is first drilled with a #38 bit to a depth of 1/4" and then opened to the final size with a #36 reamer. A smooth slip fit on the arbor is required for attachment with Loctite during final assembly. The pinion for the second arbor is 5/16" long and the center hole step drilled and reamed to a final diameter of 3/16". The larger diameter center hole allows the pinion to be mounted on the second wheel collet close to the wheel as shown in the second photo.

    After each pinion is drilled, the pinion rod is slowly rotated in the lathe while a Dremel tool with a cutoff wheel grinds through the leaves. Attempting to cut the leaves with a parting tool will bend or break them. When the body of the pinion is reached, the lathe is rotated in reverse while the body of the pinion is cut with a jeweler’s saw until the drilled hole is almost reached. Sawing the rod while rotating it ensures a straight cut. The lathe is then stopped and the remainder of the cut is completed with the jeweler’s saw in the normal fashion. If the cut is completed with the lathe running, the saw will catch when it breaks through and the saw blade will be broken. Any burrs are removed from the center hole and each end of the pinion is sanded flat to a 600-grit finish.

    The pinions are hardened to improve wear characteristics. Unfortunately, the pinions contain sharp corners and sections of various thicknesses that make the parts susceptible to cracking and deforming during the heat treatment process. Therefore, steps must be taken to minimize these potential problems. To even out the heat distribution, the pinion is placed in a wire cage. The cage is made of soft iron wire (rebar tie wire) that is formed by wrapping it around the remaining pinion rod. 3-4 inches of wire are left on one end which serves as a handle. A small flat washer is placed at the bottom of the cage, the pinion is inserted and another washer or flat disk is placed on top of the pinion. The washers and wire cage deflect the direct heat from the torch and help heat the pinion evenly. These components are shown in the first photo. To further help prevent cracks during quenching, the quench oil is heated to between 90 & 140 degrees F as recommended by Machinery’s Handbook. I simply set the quench oil container on an electric hot plate. Any oil will work, but I prefer mineral oil. It is clean, does not smoke, quenches well, is non-toxic, (it is actually sold as a laxative!) inexpensive and readily available at any drug or grocery store.

    To prevent decarburization (loss of carbon from the metal) and scale build up, the wire cage and pinion assembly are dipped in a solution of denatured alcohol and boric acid (available as roach killer). The solution should be the consistency of pancake batter. The boric acid forms a barrier around the part and prevents oxygen from reaching the surface. The recommended temperature for hardening O-1 drill rod is 1475 °F, which corresponds to a dull cherry red color. However, since the pinion cannot be easily viewed while inside the cage, a magnetic test is also used to indicate when the correct temperature is approaching. For O-1 drill rod, The Curie temperature, or temperature when it becomes nonmagnetic, is approximately 1440 ºF. The assembly is slowly and evenly heated by rotating the cage in the torch flame. The boric acid and alcohol produce a bright green flame as the alcohol is burned off; wish I had a picture of that! When the wire cage begins to glow a magnet is brought near it periodically. When the assembly is no longer attracted to the magnet, it is heated a few seconds more and then quench in the oil. This method is not as precise as a heat treating oven, but works pretty well. The pinion is then removed from the wire cage and washed it in hot water. In the first photo, you can see that the pinion came out pretty clean while the top of the disk scaled very badly since it was directly exposed to the torch flame and the air.

    After cleaning, the pinions are tempered in the kitchen oven at 325-350 degrees F for one hour to prevent cracking and toughen the steel. The pinion ends are cleaned up with fine sandpaper. Each pinion is then polished with a Cratex tapered edge rubberized abrasive wheel mounted in a Dremel tool. The second photo shows a Cratex wheel sitting on the vise. The pinions are held in a small vise while being polished. Only light pressure is required as even the very fine grit Cratex wheels cut surprisingly fast! The photo also shows how the second pinion is mounted directly on the collet. All of the other pinions will mount on the arbors.

    The wheels will be made next. 79518.jpg 79519.jpg
     
  19. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Cutting Wheels for the Pinwheel Skeleton Clock

    Before discussing the actual process of cutting wheels for the skeleton clock, I would like to show some of the tools required. All of the tools I use are home made. If I could go back and start over, I would probably buy some of the tools rather than make them as will be pointed out shortly.

    First up are the cutters. A selection are shown in the first photo. The upper left is a .8 module, single point flycutter made according to instructions provided by John Shadle on his OnlineClockBuilding website. The cutter in the lower right is similar except it is a .5 module cutter used for the motion works wheels. The upper center is a .8 module multi-tooth wheel cutter and the upper right is a .8 module cutter for 12 leaf pinions. These are made according to instructions posted by David Creed on the Yahoo MLHorology group. All of these cutters work very well but they take a lot of time and care to make. The greatest difficulty I have with making cutters is getting both sides of the cutter symmetrical. Any discrepancies will show up as mis-shaped teeth on the wheel. In hind sight, I would purchase a multi-tooth cutter from PP Thornton since I seem to primarily use one cutter the most for .8 module wheels. They are expensive, but it would have saved a lot of time and trouble. On the positive side, the experience of making cutters has equipped me with the ability to make one-off cutters such as pinion cutters as they are needed. The rod-shaped cutter in the lower left is a .5 module 12-leaf pinion cutter used for one pinion of the motion works. Because of the small size of the cutting tip, this cutter can only be used on brass. Attempting to cut steel would quickly break off the tip.

    The next item needed is a way to hold the wheel blank while the teeth are being cut. The second photo shows two arbors that fit on the Taig lathe. The one on the left is for wheels over 1-inch. It has a center shoulder that fits snuggly into the center hole of the wheel blank. Several adapters are also shown for holding wheels with larger center holes. The arbor on the right has been turned down to hold wheels down to 3/4-inch. The cap screw serves to center the wheel blank. In typical use, a wheel blank is sawn from a sheet of brass. It is then mounted on the arbor, turned to size and the teeth cut without being removed during the entire process.

    The third photo shows a method I like to use on large wheels such as the great wheel. Here a block of wood is mounted to a faceplate. The wood is faced so the mounting surface runs true in the lathe. The wheel blank is then fastened to the wood with screws. Some care is required to make sure the spokes can be laid out around the screw holes. This method of mounting provides several benefits. The wood block provides backup material to the blank. More importantly, the center hole and outside diameter can be machined with one setup, ensuring the teeth run perfectly true with no runout.

    The fourth photo shows several other items needed to cut wheels. I use a dividing plate mounted on the outboard side of the lathe headstock. The dividing plate provides a sure and fairly fool-proof method of turning the wheel blank from one tooth to the next. The downside is you are limited to cutting wheels that have the same number of teeth as the dividing plate holes. Wheels with fewer teeth can be cut by skipping holes, but you are still limited to counts that are evenly divisible from the holes available. If a wheel is needed with a tooth count that is not available, a new ring of holes must be drilled. This takes time and must be done carefully to get accurate spacing of the holes. Given the opportunity to start over, I would purchase a good quality rotary table.

    The fourth photo also shows two other necessary components. An arbor is needed to hold the wheel cutter. This one is very similar to the wheel arbors shown in the second photo except it has a larger mounting bolt and is tapered so the arbor will clear the wheel blank. Also needed is a milling spindle to drive the cutter. Here, a second Taig headstock is mounted to the milling attachment on the cross slide. This combination works very well; allowing the spindle to move up/down to set the tooth depth, in/out to center the cutter over the wheel, and left/right to move the cutter into the wheel to cut the tooth. The milling spindle is driven by a flex cable attached to the lathe motor. Not an ideal drive method, but it is versatile and easy to set up. This photo shows the 72-tooth wheels being cut.

    I will go into the actual wheel cutting process on the next update. 80162.jpg 80163.jpg 80164.jpg 80165.jpg
     
  20. leeinv66

    leeinv66 Moderator
    NAWCC Member

    Mar 31, 2005
    9,470
    116
    63
    Male
    Industry Specialist
    Launceston Tasmania
    Country Flag:
    Region Flag:
    Re: Cutting Wheels for the Pinwheel Skeleton Clock

    Thank you for another great posting Allan!!! I'll start with my first dumb question. How do you workout where the holes in the dividing plate need to be? Does it matter where on the dividing plate they are drilled as long as they are evenly spaced?
     
  21. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Cutting Wheels for the Pinwheel Skeleton Clock

    Peter,
    Even spacing of the holes is the most important criteria for the index plate. How much space is not as important. If they are not evenly spaced, the wheel teeth will not be the same thickness. If you look closely at the plate in the 4th photo, you will see that the spacing of the 96 holes on the outside ring are closer together than the inside rings, the smallest of which has only 8 holes. The only requirement is that the holes have some space between them so the holes do not touch. Of course, you want the rings with the fewest holes closest to the inside of the index plate, saving the outside rings for the larger hole counts.
    This thread has some additional information on ways to make index plates.
    https://mb.nawcc.org/showthread.php?t=43632&highlight=index+plate
     
  22. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Cutting Wheels for the Pinwheel Skeleton Clock

    This clock requires three 72-tooth wheels. For a .8 module wheel, a 2.355" blank is needed. I cut the blank about 1/8" oversize to compensate for centering errors and sawing variations. Each disk is marked out with a dividers and cut from a sheet of 1/16" thick, 356 engraver’s brass. The center hole is drilled and reamed to size. If the lathe arbor is long enough to hold all three blanks at once, machining the diameter and cutting the teeth together can save a great deal of time. My arbor could accommodate only one wheel and a backing plate, so the wheels were machined separately. This is the arbor on the left shown in photo 2 of the previous post. The backing plate is a scrap disk of aluminum or brass about the same diameter or slightly smaller than the wheel blank and about the same thickness. The backing plate helps reduce vibration and deformation on the back side of the teeth during cutting.

    The diameter of the blank is turned down to 2.355" and the edge is coated with layout blue. The lathe is then setup as shown in the 4th photo of the previous post. The arbor holding the wheel is removed and a pointed arbor is installed on the lathe headstock. The cross slide is then adjusted so the cutter is centered as shown in the first photo. If the cutter is off-center by even a little bit, the teeth on the wheel will lean. Later, when the ratchet wheels are cut, we will intentionally offset the cutter for this very reason. With the cutter centered, the gib screws are tightened to lock the cross slide in place. The arbor and wheel blank are then re-installed and the milling attachment is raised so the cutter does NOT make a full cut. The cutter is backed off so it clears the blank and the lathe is started and brought up to approximately 300 RPM. Faster speeds can be used since the engraver's brass is pretty easy to cut. The carriage is moved towards the blank at a slow and steady rate until the cutter passes completely through the blank. The carriage is then moved away from the blank until it is well clear of it. The wheel is then indexed to the next position and another pass is made with the cutter. This will leave a partially formed tooth with a flat top. The blank is indexed back to the first pass and the milling attachment is lowered slightly. Cuts are made through the same two slots that were cut the previous time. The flat top of the tooth should be getting thinner as it takes on the curved part (addendum) of the tooth. The process of lowering the milling attachment and re-cutting the slots is repeated until there is just a sliver of layout blue visible on the top of the tooth. At this point, the gib screws on the milling spindle are tightened to lock it in place. The remaining teeth are now cut to full depth in a single pass. When the last tooth is cut, index the wheel to its original position and run the cutter through the first slot cut. If the cutter does not remove any brass, this is a good indication that the wheel was cut successfully. If the cutter does remove any material, something slipped or vibrated out of position during the machining operation and the wheel will likely need to be scrapped.

    After the teeth are cut, and while the wheel is still mounted in the lathe, the centerlines of the spokes are marked on the wheel using the index plate to space them evenly. Like the escape wheel, these wheels have five spokes. The milling spindle is then remounted for drilling and the three mounting holes are drilled, also using the index plate to space them evenly. The wheels are then removed from the lathe arbor and coated with layout blue on one side. The spoke outlines are then lightly scribed on the blued side and the wheel is crossed out with a jeweler’s saw.

    A difficulty that is often encountered with wheel crossing is getting the spokes symmetrical in spacing and shape. I find it helpful to draw the wheel and spoke outlines in CAD and print an actual size drawing. The wheel can then be laid on the drawing as a visual aid to indicate where adjustments are needed during the filing process. This is shown in the second photo.

    Proper files make a big difference in the effort required to shape and smooth the wheels. The files should be sharp and reserved for use only on brass. Initial shaping is done with a #2, 4-inch crossing file for curved surfaces and a #2, 4-inch barrette file for flat surfaces. A #4, 4-inch crossing file is used for final shaping and initial draw filing. A #6 crossing needle file is then used for draw filing to achieve a finish that is equivalent to 400-grit paper. After filing, the wheels are then cleaned in denatured alcohol to remove any remaining layout dye and progressively sanded to a 600-grit finish.

    A collet is needed for each of the wheels. These are identical to the escape wheel collet.

    The great wheel will be cut next. 80926.jpg 80927.jpg
     
  23. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Cutting Wheels for the Pinwheel Skeleton Clock

    The great wheel is cut using the method shown in the third photo of the first post. A block of wood is fastened to a small face-plate and turned to a diameter to match the great wheel blank. Facing cuts are made across the block to ensure that it runs true. A 1/2" center hole is then drilled into the block for clearance when the wheel center hole is bored.

    The 96-tooth great wheel is turned from a blank with a final diameter of 3.111". A 3/16" thick disc of 356 brass is center punched and the diameter is rough-cut oversize with a jeweler's saw. How much oversize will depend on the accuracy of the sawing and how well the disk can be centered on the wood block. I made my blank approximately 3.25". Three holes are drilled approximately 1.25" from the center and equally spaced around the blank. These holes will be removed later when the wheel is spoked. The blank is now placed on the wood block and approximately centered with the aid of the tailstock center. Wood screws are used to fasten the blank to the wood block.

    The diameter of the blank is now turned to the final diameter. I had to use riser blocks on the headstock and toolpost of the Taig lathe to obtain enough clearance to reach the edge of the blank. The teeth are cut using the same setup as the 72-teeth wheels. The wood block is sacrificial and easy cut by the gear cutter. After the teeth are cut, the center hole is drilled. Starting at 1/8" the hole is step drilled to a diameter of 1/4" and then opened to an "almost final" diameter of .305" with a small boring bit. By cutting the teeth and boring the center hole with-out disturbing the setup, the wheel is guaranteed to run true. The hole is then enlarged to its final diameter with a 5/16" reamer. Although the center hole could be bored to its final size, using a reamer to finish the great wheel and maintaining wheel center hole ensures they are exactly the same diameter.

    The wheel is removed from the wood block and rubbed on sandpaper to remove any burs. The wheel must now be machined to make room for the maintaining spring. This is accomplished by removing half of the thickness of the wheel where the spokes will later be cut in an operation known as trepanning. The wheel is mounted on an arbor and material is removed with a round nose or Vee-shaped lathe bit. The corners are then cleaned out with a small boring bar as shown in the first photo.

    After the trepanning is complete, the wheel is removed and coated the flat side with layout dye. The spokes are marked and crossed out in the manner as as previous wheels with the exception of the slot area at the end of one spoke. This slot will accept a pin from the maintaining wheel. A 1/8" hole is drilled at each end of the slot and the remainder is cut out with a jeweler’s saw and filed to shape. A second spoke is drilled to accept a pin. This pin and the maintaining wheel pin will provide a means of attaching the maintaining spring. Finally, a hole is drilled and tapped in the hub for a screw that will hold the wheel retaining washer. A pin is cut approximately 1/64" shorter than the wheel thickness (3/16") and fastened in the wheel spoke hole with Loctite 609 so it is flush with the flat side of the wheel and slightly below the rim of the wheel on the trepanned side. The wheel is finished to a 600-grit finish and is shown in the second photo.

    The maintaining wheel is next. 81592.jpg 81593.jpg
     
  24. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Cutting Wheels for the Pinwheel Skeleton Clock

    The maintaining wheel is made from 1/8" brass. As with the great wheel, a slightly oversize blank is mounted to a wood block and the diameter is machined to 2.8". The center hole is drilled and reamed exactly like the great wheel. There are 96 ratchet teeth cut on this wheel. When cutting wheel teeth, great care is exercised to make sure the cutter is on center so the teeth do not lean. For ratchet teeth the cutter is deliberately offset so the front face of the ratchet leans slightly to help hold the click (or pawl in this case) in the groove. The setup is shown in the first photo with the cutter offset about 1/8" to the right. The depth of the notch is determined by trial and error, and it should not be very deep. A shallow notch helps minimize the noise made by the pawl as it drops in each notch when the wheel rotates during normal operation. A small flat spot may be left at the tip of each tooth. The notches shown in the second photo are approximately 1/32" deep.

    After spoking the wheel, several holes are drilled for fastening various other components. The hole is drilled and tapped to hold the stud for mounting the fusee ratchet click. Two small 0-80 holes are drilled and tapped for mounting the click spring and another hole near the edge is for the pin which protrudes out the back and through the slot on the great wheel. The third photo shows the completed maintaining wheel. The side facing up goes towards the great wheel.

    The wheel is finished to 1500 grit and the pin is fastened with 609 Loctite. A finer finish is applied to the maintaining wheel so polishing is the only remaining step. That is because additional sanding will be difficult with the pin in place. The fusee click, spring and maintaining spring will be made later.

    The mainspring barrel is next. 81594.jpg 81595.jpg 81596.jpg
     
  25. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Mainspring and Barrel for the Pinwheel Skeleton Clock

    The mainspring barrel for this clock is somewhat larger that on other fusee clocks I've seen. Dimensions obtained from scaling photos of the original indicate the outside diameter of the barrel is 3". Since the proportions look good on the clock and work well with the chosen mainspring, I will stick with a 3" barrel. The mainspring will be covered in more detail shortly.

    A 3" long piece of 3" outside diameter, 1/8" wall thickness brass tubing is used. (A 2.5" long piece would work fine, but the cost was the same for 3" from the supplier.) The extra length allows all of the machining to be done with a single setup. The barrel is made somewhat wider than necessary to allow using mainsprings up to 1.5" in width and to fill the space available between the plates.

    The tubing is mounted on the lathe using the inside jaws of a 4-jaw chuck and set it to run true with a dial indicator. I could not grip the tubing from the outside since it is almost as big as the lathe chuck! The chuck jaws must not be overtightened or the tubing will be distorted. Only light cuts are taken to reduce the chance of the tubing slipping in the chuck. Most brass tubing is non-leaded brass so a better finish is obtained by using cutting oil. The end of the barrel is faced to true it up. The inside of the barrel is bored only as much as necessary to true it up for a length of 2" as shown in the first photo. The inside edge is slightly chamfered with the boring bit to prevent interference with the barrel flange.

    The barrel length is marked at 1.875" using lay-out dye and a dial caliper. An index plate is installed on the lathe headstock and the milling spindle is attached to the cross slide for vertical drilling as shown in the second photo. A center drill is used to spot three equally spaced holes around the outside diameter 1/16" from each edge of the barrel. Each hole is drilled through with a #52 bit. An additional hole is spotted and drilled with a #21 bit in the center of the barrel side for the spring hook. This hole is tapped with a #10-32 tap and countersunk on the inside of the barrel approximately 1/32" for the flat head screw that will serve as the barrel hook. I will make the hook a little later.

    A 1/16" hole is drilled centered between two of the barrel flange mounting screw holes on one edge for the fusee cable. The hole is spotted straight on and then the barrel is turned 30 degrees (one hole of a 12-hole index circle) and drilled through at an angle to reduce the bending stress on the cable. This is shown in the third photo. Finally, a light clean-up cut is made on the outside of the barrel to true it up and remove any scratches. The outside of the barrel is sanded to a 600-grit finish and then parted off to a length of 1.875". The inside edge of the barrel is then chamfered as before with a file or deburring tool. 81843.jpg 81844.jpg 81846.jpg
     
  26. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Mainspring and Barrel for the Pinwheel Skeleton Clock

    Disks for the barrel flanges are cut from a sheet of 3/16" thick brass sheet. The final diameter will be 3-1/8" so the disks are marked out and rough cut to 3-1/4" to allow a little extra material to compensate for inaccuracies in drilling the center hole or saw slip-ups. A coping saw works fine for roughing out brass and the blades are not near as fragile as a jeweler’s saw. After the disks are cut, the center hole is drilled and the disk is mounted on the same arbor used to cut the wheels. This arbor requires a small center hole. A 1/4" arbor would be better, but I did not have one at hand.

    With the disk mounted on the arbor, the outside diameter is turned down to 3-1/8". A 1/8" shoulder is then cut for a slip fit into the end of the barrel. When the shoulder diameter is correct, the corner is cleaned out with a graver or parting tool so the flange lip fits tightly against the edge of the barrel. The flange lip is then rounded over with a file and progressively sanded to a 600-grit finish.

    One flange will require a cutout for clearance of the fusee cable where it enters the barrel. A flange is mounted on the milling attachment and an end mill is mounted in a collet to machine the cutout. The setup to mill the cutout is shown in the first photo. This arrangement proved to be less than secure and when I started to take the first cut, the end mill spun the flange and left track marks around the edge of the shoulder. Fortunately these marks will not be visible when the barrel is assembled. I was able to complete the cutout by taking very light cuts. The second photo shows how the cutout lines up with the cable hole in the barrel.

    The barrel flange is removed from the arbor and installed it in a 4-jaw chuck using the outside jaws to grip the shoulder as shown in the third photo. The 4-jaw chuck is carefully adjusted until the barrel flange runs true, first on the face and then on the flange lip. The center hole of the flange is bored undersize to approximately .360" with the last cut being light and slowly advanced to obtain the best finish possible.

    Next up, the flange holes will be brought to final size in a rather unorthodox manner. 81856.jpg 81857.jpg 81858.jpg
     
  27. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Mainspring and Barrel for the Pinwheel Skeleton Clock

    The center holes of the barrel flanges serve as bearing surfaces on the barrel arbor so a good fit and finish are desired. Typical clock construction techniques would use large cutting and smoothing broaches to size, burnish and work harden the hole. Unfortunately, broaches of this size are expensive and not widely available. Instead, a simple but seldom seen technique (in horology anyway) called "ballizing" or "ball sizing" is used. A ball bearing is pressed through a slightly undersize hole to expand, smooth and work harden the surface. Some spring back will occur, so the barrel arbor will be machined to fit the hole in a later step. An arbor press or bench vise provides the power to push the ball through the hole. I prefer using the bench vise as it provides better control and does not jump when the ball goes through the hole.

    The first photo shows the flange with the cutout, a 3/8 inch chrome steel bearing ball and a 1/4" diameter steel pusher. The end of the pusher is lightly drilled to form a concave surface so the ball stays centered.
    In the second photo, a metal block on the right side of the flange has a 1/2" hole through it to receive the ball after it passes through the flange. The pusher should be no longer than necessary to reduce the tendency to twist and it should be carefully aligned so it pushes the ball straight through. A light coat of oil is applied to the surface of the hole for lubrication.

    The vise is closed, forcing the ball through the center hole. The 3/8" ball expanded the hole from .360" to .373" A great deal of force is required to push the ball through the hole and a burr will be raised on both sides of the flange as shown in the third photo. The burr is easily removed by facing each side of the flange followed by progressively finer grits of paper until a 600-grit finish is achieved. 82148.jpg 82149.jpg 82150.jpg
     
  28. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Mainspring and Barrel for the Pinwheel Skeleton Clock

    The barrel flange holes are located by installing a flange in each end of the barrel, making sure to line up the cable cutout in the flange with the cable hole of the barrel as shown in a previous post. The flanges are clamped in place and a dimple is made in the flange shoulder through each hole with a drill bit. This operation is shown in the first photo.

    It may be tempting to drill the flange holes completely at this time, but this often results in broken drill bits when chips lodge in the hole at the barrel/flange interface. Instead, the flanges are removed from the barrel to complete the drilling. The holes in the flanges are tapped to receive #0-80 screws. As a final construction step, the 6 flange screw holes in the barrel are countersunk so that a flat head #0-80 screw sits flush with the surface. This is process is easily completed by slowly rotating a countersink mounted in the lathe and pressing the barrel by hand into the countersink as shown in the second photo.

    We can now obtain a measurement from the completed barrel that is needed to machine the barrel arbor. With the flanges installed, the inside distance between the flanges is measured with the tail end of a dial caliper as shown in the third photo. The measurement obtained is from the inside of one flange to the outside of the flange where the caliper is inserted. The thickness of the flange is subtracted to obtain the inside dimension. Alternately, the outside dimension of the flanges can be measured from which the thickness of both flanges is subtracted. Regardless of the method used, each flange thickness must be measured since facing cuts have been taken on each flange and they are no longer exactly 3/16" thick. Finally, 1/64" is subtracted from the inside dimension to allow for end shake of the arbor inside the barrel.

    We now have enough information to make the barrel arbor. 82153.jpg 82154.jpg 82155.jpg
     
  29. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Mainspring and Barrel for the Pinwheel Skeleton Clock

    Raw stock for the barrel arbor is a 4" length of mild steel rod 1" in diameter. I used drill rod because it was on hand, but machining hard steel of this size works the small Taig lathe to capacity. Although the finished length will be 3.25", 3/8" is added to allow for center drilling each end. The center drill holes will be cut off near the end of the machining process. Also, note that the arbor will be roughed out slightly oversize and turned between centers to the final dimensions.

    The rod is mounted in the 4-jaw chuck and set true with a dial indicator positioned close to the chuck. A steady rest is installed near the chuck and the arms positioned for a sliding fit before moving the steady to the unsupported end of the rod. The end is faced and drilled with a center drill to accept a dead center. The steady rest is removed and a dead center is installed in the tailstock. A live center would be better if you have one. I use a short piece of 3/16" drill rod turned to a 60-degree point and held in a drill chuck as shown in the first photo. The outside diameter of the exposed arbor is reduced to .910". Then the end of the arbor is reduced to .385" for a length of 1.25” as shown in the first photo. With the cutting tool against the shoulder just cut, the carriage is backed off .320" and the carriage stop is set. The remaining end of the end of shaft is reduced to .320". As before, the carriage is backed off, this time to .190" and the carriage stop is set as show in the second photo. The remainder of the shaft is reduced to .283". This concludes the roughing out of this end of the arbor. The arbor is reversed with the reduced end passing through the lathe chuck. The chuck should grip on the large diameter portion of the arbor, not the section that was just reduced. The chuck is adjusted so the arbor indicates true on the previously machined surface. Using a steady rest, the end is faced and center drilled to accept the dead center as before. With the dead center in place, the outside diameter of the arbor is reduced to .910" to match the previously machined portion. The carriage stop is set to leave 1.625" (the measured barrel inside dimension obtained previously) at the .910" dimension and the remainder of the shaft is reduced to .385". The carriage is backed off.440" and the remainder of the shaft is reduced to .320". This completes the roughing out of the barrel arbor and it can now be machined to its final dimensions.

    The barrel arbor is mounted as shown in the third photo, being extra careful to true the work in the 4-jaw chuck. The large diameter center section is verified that it measures is 1.625" (as measured) and adjusted as necessary. Next each barrel pivot is reduced for a tight fit into the barrel flanges. The pivots are the areas that where roughed out to .385". They should end up close to .373" (as measured after the flange holes were ballized), but the actual dimension is determined by test fitting to the barrel flanges. The arbor will need to be removed to test fit the pivot on the chuck side of the arbor. The arbor must be trued each time it is returned to the lathe.

    After the pivot diameters are correct, the pivot lengths are trimmed to obtain a distance of 2.375" needed to fit between the plates. This measurement is shown in the third photo. Note the barrel will not be centered on this dimension. It is located slightly closer to the front plate so the fusee cable will line up correctly. Now that the pivot lengths are correct, they can be finished in the same way as the previous pivots with a #6 file, Arkansas stone and burnisher. The barrel flanges should slip easily onto the pivots with very little wobble. The pivots that run in the plates can now be made, but since the plates do not yet exist, a 5/16" hole is drilled and reamed in a scrap piece of 3/16" thick brass. Each plate pivot is reduced for a tight slip fit into the test hole. The plate pivots are prepared in the same manner as the barrel pivots for a length of .2". I cover the pivots with masking tape to protect finish. The ends will be cut off ends to their final dimensions and the winding arbor flats formed in a later step.

    The center of the barrel arbor center section is located and drilled a with a #18 bit approximately 5/8" deep. The hole is tapped #10-32 and countersink approximately 1/16" deep. This will hold the mainspring hook which will be made next. 82554.jpg 82555.jpg 82556.jpg
     
  30. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Mainspring and Barrel for the Pinwheel Skeleton Clock

    The mainspring hooks in the barrel and arbor are made from 10-32 steel flat head screws. The head of the screw for the barrel hook should stand approximately .040" above the inside surface of the barrel. The head of the screw is ground on the bench grinder and test fitted until this height is achieved. The sharp edge is then removed with a file. Loctite 609 is applied to the screw threads and the screw is tightened by gripping the threaded end with a pliers or bench vise. After the Loctite sets up, the screw is cut off flush with the outside of the barrel with a jeweler’s saw. Some masking tape on the barrel around the screw area will help prevent scratching the barrel surface. The screw is carefully filed and sanded until it blends in with the barrel. The first photo shows the barrel hook before it has been screwed all the way in.

    A flat head screw is installed in the arbor making sure that the threads are deep enough to allow the head to fully seat against the arbor. A few threads can be cut off if the screw is too long. With the screw installed in the arbor, the head is ground down so the front of the screw is approximately .040" above the arbor. The back side should slope down to .020" to prevent it from distorting the mainspring when it is wound. The arbor winds clockwise so the front of the hook will be on the right when the arbor is viewed from the winding side as shown in the second photo. Note this photo was taken after the winding arbor was machined. I will cover how that is done later. When the final shape is obtained, the arbor hook is secured in place with Loctite.

    Next time, we will do some calculations in preparation for making the mainspring. 82577.jpg 82578.jpg
     
  31. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    #31 Allan Wolff, Jan 29, 2011
    Last edited by a moderator: Dec 26, 2017
    Re: Mainspring and Barrel for the Pinwheel Skeleton Clock

    Standard mainsprings are typically inexpensive and hardly worth your time to make one. Unfortunately, fusee mainsprings can be expensive and the current price for one in the size needed for this clock is approaching $80. This spring was made for less than half that amount. It is made from a strip of blue spring steel. A thickness of 0.020" and width of 1.5" are chosen as these appear to be relatively common dimensions for fusee mainsprings. Unfortunately, there is no formula that I am aware of that provides mainspring dimensions to obtain a given force. Any formula would be highly dependent upon the characteristics of the steel used and these characteristics are widely variable. Therefore, dimensions are chosen and the fusee will be adjusted to the extent possible based on actual measurements. If a stronger or weaker spring is needed, new material will need to be obtained. According to a rule-of-thumb found in a 1905 book (available from Google Books) called "Lessons in Horology" by Jules and Hermann Grossmann, it is customary to make the radius of the arbor equal to 1/3 the internal radius of the barrel to obtain the maximum number of turns from the spring. Using this ratio with an internal barrel radius of 1.375", the arbor radius should be .45" or 0.9" diameter. The same book also states the minimum arbor diameter should be 32 times the thickness of the mainspring. This is to avoid excess stress when the spring is wound, causing it to break or take a "set". Applying this rule yields a minimum arbor diameter of .64" so our arbor diameter of .91" provides an additional margin.

    The commonly used formula shown below gives the mainspring length that provides the maximum number of turns of the barrel.
    82724.jpg
    Where:
    L = the length of the mainspring
    R = inner radius of the barrel
    r = radius of the arbor
    t= mainspring thickness

    Plugging in the values of our components we get:
    82725.jpg = 132.6 inches

    The blue spring steel supplier sells the material in 10 ft. or 70 ft. lengths. Rather than order excess material that may need to be replaced if the mainspring size does not work out, a 10 ft. piece is ordered. At 120", this spring is too short to obtain the maximum number of turns. Just how many turns we will lose with the shorter spring requires a more complex equation; (also from the Grossmann book):
    82726.jpg
    Where:
    N = the number of turns
    r’ = the inner radius of the mainspring coils when the spring is fully unwound against the barrel, which is given by the equation:
    82727.jpg

    At L=120" 82728.jpg = 1.061

    Plugging this into the turns equation gives:
    82729.jpg = 10.98 turns (I don't know why this image is so small, sorry.)

    At L=133" 82730.jpg = 1.022

    Again, plugging this into the turns equation gives:
    82731.jpg = 11.05 turns

    This shows the 120" spring will provide substantially the same number of turns as a 133" one; 10.98 vs 11.05. When it comes time to measure the force of the spring, we can count the actual number of turns and verify our math.

    The mainspring calculations given here may be more detail than some clockmakers care to bother with. However, these equations will be useful if the builder chooses to use a different mainspring.

    With the math out of the way, we can proceed with making our mainspring.
     
  32. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Mainspring and Barrel for the Pinwheel Skeleton Clock

    A 10-foot long, 1.5" wide, .020" thick strip of blue spring steel was purchased for the mainspring. It is unwound, cleaned and inspected for cracks or rust. The steel is provided in a hardened and tempered state and must be annealed at the ends to drill the holes. Approximately 4" of one end is heated to a dull red and cooled slowly by gradually removing the heat source. The corners are easy removed with tin snips with final shaping done on the bench grinder. A file is used to remove any burs or rough edges caused by the grinder. The location for a 5/16" hole is center punched and drilled approximately 3/4" from the end. Any burs need to be removed. Clamping the steel to a 7/8" bar as shown in the first photo. The bar is held vertical in a bench vise. Wrap the steel around the bar while heating the spring steel with a propane torch to form the inner loop. The inner loop should consist of approximately 3/4 of a turn. The other end of the spring is annealed and drilled in the same manner but is left straight. Both ends of the mainspring are cleaned with a brass or steel brush to remove any scale or loose particles.

    Winding a spring of this size should always be done with a mainspring winder. Previous experience showed that my winder was not robust enough and a new .9" winding arbor would be needed to match the barrel arbor of this clock. Instead, the Taig lathe is used to safely wind the mainspring. A handle is fitted to the headstock pulley and the barrel arbor is mounted in the 4-jaw chuck with the outboard end supported by a center held in the tailstock. A 10-32 screw with two nuts and a washer are used to fasten the free end of the mainspring to the cross slide. With this setup, shown in the second photo, the mainspring is cleaned and then wound and released while dry. This initial wind provides a "feel" for how the spring will behave before the additional complication of oil is added. With the spring relaxed, it should be cleaned one last time and then lubricated. I used Slick 50 as a synthetic lubricant, but any oil or grease designed for mainspring service should work fine. After lubricating, the spring is fully wound, released and wound again to distribute the lubricant. The spring is held in place with a retaining sleeve made from a galvanized pipe fitting. The sleeve needs to be big enough to fit over the fully wound spring and small enough to fit inside the barrel with a little room to work. A 2.5" sleeve was used and is shown in the third photo. The free end of the mainspring is bent to a slight curve to matches the barrel. The spring is then installed in the barrel, rotating it to catch the hook.

    Make sure the spring coils in the same direction as the cable hole as shown in the 4th photo.
    The arbor is reinstalled in the mainspring and the entire assembly is mounted back in the winder. With the barrel firmly in hand, the spring is wound until the retaining sleeve can be removed. (Have someone remove the retainer for you since you will have one hand on the winder and the other on the barrel.) Finally, the spring is let down into the barrel. 83079.jpg 83080.jpg 83081.jpg 83082.jpg
     
  33. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Mainspring and Barrel for the Pinwheel Skeleton Clock

    Now that the barrel arbor has completed its service as a winding arbor, final machining can be completed. The arbor is mounted in the lathe with the long end towards the tailstock. Traditionally, the end is filed square with a file rest. A quicker and easier method is to mill the flats with an end mill as shown in the first photo. There is quite a bit of overhang, so light cuts are taken to prevent chatter. The index plate is used to hold the arbor in position while it is milled. Each end of the arbor is then sawed off and filed to the final dimensions, removing the center drill holes in the process.

    The barrel ratchet is made from 1/8" mild steel. Brass could be used, but the silver color of the steel gives a little extra visual appeal. Similar to the wheels, the raw disk is cut slightly over size disk from the sheet stock. A center hole is drilled and the blank mounted on a 3/4" diameter arbor with a machine screw. The diameter of the blank is then reduced to 1". An attempt was made to cut the teeth with a fly cutter. The fly cutter previously worked well for a brass ratchet, but the sharp point of the cutter became rounded after the first few passes through the steel. Instead, a slitting saw works much better. A 0.025" thick slitting saw blade is mounted on the milling attachment and centered over the ratchet arbor. The saw is then offset 0.052" to the right to give each tooth a 6 degree undercut. The saw is lowered until it just touches the edge of the disk and then lowered further to cut a slot 0.068" deep. The setup is shown in the second photo, looking from the tailstock. As with the pinions, slow speed and cutting oil are required to preserve the slitting saw. 26 slots are cut to form the face of each tooth. The number of teeth was chosen so the geometry of the teeth is obtained without re-indexing the ratchet. I cannot show you the math since experimentation with a CAD drawing was used to determine the number of teeth and dimensions to offset. After the first cut is complete, the saw is offset an additional .341" to the right and cuts to a depth of .129" as shown in the third photo. Some adjustments to these dimensions may be needed so the second cut intersects the first cut without hitting the face of the tooth. The hub of the ratchet is then dished out, purely as a decorative feature. The hub is then smoothed and polished with Cratex abrasive wheels mounted in a Dremel tool while the ratchet is slowly rotated in the lathe. Burs are removed from the back of the ratchet with a file or parting tool before removing the ratchet from the arbor. The center hole of the ratchet is then roughed out to a square with a jewelers saw and filed to its final shape for a slip fit over the square end of the barrel arbor. The completed ratchet is shown in the fourth photo. Final polishing is still needed. An overcut can be seen at the base of each tooth face. This trade-off was made in order to obtain the maximum tooth height and a square corner at the face of the tooth. Alternately, the saw would have left an angled cut in the corner that would prevent the click from seating properly and requiring additional filing to square up the corner. 83368.jpg 83369.jpg 83370.jpg 83371.jpg
     
  34. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Mainspring and Barrel for the Pinwheel Skeleton Clock

    The barrel click is cut from 1/8" thick mild steel. The actual shape of the click is somewhat arbitrary and can be adjusted to suit your taste. However, the geometry has a critical dimension that must be observed or the click will not stay in the ratchet wheel tooth. I found this out because my first click would not stay in the ratchet. It was amazing how little force was required to pop it out of place. Here's what happened. Refer to the bottom drawing of the first photo. A solid red line is drawn from the center of the click pivot point through the tip of the click. A perpendicular dashed line is drawn towards the center of the ratchet. When the click is too long, the perpendicular line is on the wrong side of the center of the ratchet wheel. This results in two forces. The blue arrow indicates a force towards the click. The second force shown by the magenta arrow is directed away from the pivot causing the click to lift out of the ratchet tooth. The top drawing uses a shorter click which results in the perpendicular line falling on the other side of the ratchet wheel center. The resulting forces now have a downward component, shown by the green arrow, that helps hole the click in the ratchet tooth. Something to watch out for when working with ratchets and clicks.

    A click spring will also be made, but this will not occur until the plates are available so the spring can be matched to the shape of the plate.

    A special screw is made to hold the click in place. It has a shoulder that provides a bearing surface that is slightly thicker than the click. This allows the click to rotate with the screw tightly fastened to the plate. The screw and click are shown in the second photo.

    The next major step in building the clock is to construct the fusee. The purpose of the fusee is to match the mainspring to the clock. First, we need two pieces of information; the force of the mainspring and the power required to run the clock. We have most of the parts necessary to measure the mainspring, but we need to build a few more parts to measure the power needed to run the clock. These are the second,third and fourth arbors.

    The fourth wheel arbor is identical to the escape wheel arbor, so I will not go over that process again. The second wheel arbor is also very similar except it is a little thicker, made from 1/8" music wire (instead of .106") and the pivots are slightly larger as well.

    The third wheel arbor will extend through the front plate and drive the motion works. A standard pivot is cut on one end and a mild steel sleeve is used on the other end to establish the distance between the plates. The final length of the arbor will be determined later, so we will leave the end of the arbor a little long and cut it off at that time. The sleeve is reamed for a slip fit on the arbor and held in place with Loctite 609. This arbor with the wheel, collet and pinion installed is shown in the third photo.

    Next time, we will measure the force of the mainspring so the fusee can be calculated. 83615.jpg 83616.jpg 83617.jpg
     
  35. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Fusee for the Pinwheel Skeleton Clock

    The fusee matches the power needed to drive the clock with the power available from the mainspring. A fully wound spring will deliver more power than a spring that is only wound a turn or two. Typical clock construction books will specify a certain mainspring size and fusee shape, but often do not go into detail about how these dimensions are determined. For this project, the force available from the mainspring and the force required to run the clock will be measured. The fusee shape will then be determined to match the two forces.

    A 10-foot length of 3/64" wire rope serves as the fusee cable and is fed through the barrel and knotted to keep it from pulling through the hole. The barrel is then mounted between the temporary plates and secured firmly in a bench vise. The ratchet and click are installed to keep the barrel arbor from turning. A digital fish scale is attached to the free end of the cable to measure the force available from the mainspring as shown in the first photo. A flat tape measure (not shown) is fastened to the bench vise and held next to the scale to determine the 1-inch intervals where readings will be taken.

    The cable is pulled until the mainspring is fully wound and the length is noted on the tape measure. The spring is then slowly allowed to return. It is necessary to keep an eye on cable as it winds back onto the barrel so it does not overlap or run off of the edge of the barrel. The wind/unwind process is repeated several more times to distribute the mainspring oil and condition the spring.

    This would be a good time to locate an assistant to write down the data as you take the readings. The first reading will be taken with the spring fully wound (cable pulled all the way out) and subsequent readings are taken at 1-inch intervals as the cable winds back onto the barrel. This is the direction the spring will be working under normal operation. If the readings are taken as the cable is pulled out, the higher force required to wind the clock will be measured. Make sure to keep an eye on the cable as it winds back on the barrel so it does not run over the edge of the barrel (Do you get the feeling that something like this happened to me? Well, it did and I can tell you it makes quite a large bang when the spring releases all at once! No damage done though.)

    Readings taken from the hand held scale proved to be somewhat erratic. A ratchet or other mechanical means of steady support would provide more stable readings. However, the data is usable by making a few manual adjustments and using some averaging techniques. The data is entered in a spreadsheet that I will provide next time. Since the first reading is taken with the cable fully extended, the values are entered from the bottom up. The scale readings in pounds and ounces are entered and then converted to just ounces. These raw ounce values are plotted as the magenta line shown in the second photo and used as initial starting values. The average of nine values is calculated and plotted on the same graph as the blue line. Even with the averaging, the curve is still influenced by the erratic scale readings and manual intervention is required. This "fudging the numbers" may appear rather arbitrary, and it may well be. However, it seems no more arbitrary than calculating the curve mathematically where a single force constant is applied across the entire range of the mainspring. 83931.jpg 83932.jpg
     
  36. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Fusee for the Pinwheel Skeleton Clock

    In order to measure the power required to run the clock, the fusee arbor is needed to mount the great wheel. The fusee blank is also needed to hold the great wheel and serve as a hub for the weight cord. The fusee blank will be cut oversize and turned between centers to its final dimensions in a later step. The blank is a piece of 1 7/8" brass rod cut to approximately 1.6" in length. Both ends are faced and the blank is drilled through and reamed to 5/16". The arbor is cut from 5/16" drill rod 4 3/16" long and drilled on each end with a center drill so the arbor can later be mounted between centers. The arbor and center hole of the fusee blank are cleaned with acetone to remove any grease. The fusee blank is installed with 609 Loctite approximately 1.22" from one end of the arbor. After the Loctite cures, the assembly is mounted in the 4-jaw chuck and supported at the other end with a dead (or live) center as shown in the first photo. (The photo shows masking tape placed over the arbor to protect the pivots.)

    Light cuts are made to true up the ends and surface of the fusee blank. The final diameter is noted; 1.856" in this case. This dimension will be used later to calculate the fusee shape. Each end of the arbor is machined to create the pivots that will fit into a 1/4" ballized hole. A 1/4" ballized test hole in a piece of scrap brass serves as a gage.

    The assembly is removed from the lathe and drilled and tapped to #4-40 on the surface of the fusee blank near the long end of the arbor. This hole can be seen on the left side of the blank in the first photo. The great wheel is slipped over the short end of the arbor and the location of the retaining washer screw hole is transferred to the fusee blank. This hole is drilled and tapped to #4-40 and the great wheel is fastened to it with a #4-40 screw as shown in the second photo. These #4-40 holes in the fusee blank are temporary. One hole will be removed when the fusee blank is cut to shape and the fusee ratchet will cover the other hole.

    We now have all of the parts needed to assemble the train in the temporary assembly used to test the escapement. A homemade depthing tool designed by John Wilding is used to determine the placement of the arbors. See the third photo.

    The spacing is the same between the escape, fourth, third and second arbors since the pinions and wheels are identical. Because the escape arbor is already placed from the Escapement test, work should proceed from there down the train. Each pivot hole is drilled slightly undersize, then enlarged and finished with cutting and smoothing broaches. The spacing between the great wheel and second pinion is different from the others and a separate depth measurement is required. The fusee arbor hole will be drilled to .238". Drills this large tend to wander from their original starting point so it is good practice to scribe several concentric circles around the center mark with a small dividers and slowly increase the size of the drill bit. After each increase in size, the new hole is checked against one of the scribed circles to make sure it is properly centered. The hole is recentered with a file if necessary before moving up to the next drill size. The fusee arbor holes are then brought to their final size of 1/4" with the ballizing technique used for the barrel flanges. The burr raised by the ball is removed with a file and the plates are checked to ensure they are straight.

    A hole is needed in the base of the temporary plate assembly to pass the weight cord. This hole should be offset from the arbor centerline by the radius of the fusee blank, .928" in this case and the same distance from the inside of the plate as the temporary #4-40 screw hole. A 1/2" diameter hole provides ample room for clearance.

    The wheels, pinions and arbors are installed in the temporary assembly, but the pallets are not installed at this time. The location of each wheel and pinion is determined on the arbor, similar to the arrangement shown in the first post of this project. Each part is fastened to the arbor with low-strength Loctite so the parts can easily be disassembled later. Since the second pinion is already fastened directly to the second wheel collet, this assembly can be left floating on the arbor if desired. The Loctite is allowed to set up and then clock oil is applied to the pivots. The train is test run using finger pressure to drive each wheel. Adjust any pivot holes or the plate spacing where binding occurs. If every thing is correct, the train will coast for 5-10 seconds when given a good spin from the great wheel. The pallet and pendulum assembly are now installed and test run by driving the train with finger pressure.

    A weight cord is attached to the temporary screw in the fusee blank. The digital scale is attached to the other end of the cord. A container is suspended from the scale and scraps of metal are added to adjust the weight. (No expense is spared for these test components. :D A plastic milk carton is used to for the weight container and the test stand is cobbled together from scrap lumber.) The amount of weight is adjusted until the clock runs well with a small amount of over swing after each escape pin lands on the pallet. For my clock 7 pounds of weight was required. This includes the weight of the scale. The clock may be run for a period of time to ensure proper operation and enjoy the moment of seeing your creation come to life. However, the low-strength Loctite should not be considered a permanent bond so do not leave the assembly unattended.

    Now that we know how much force is needed to run the clock, we can complete our calculations for the fusee. 84730.jpg 84731.jpg 84732.jpg 84733.jpg
     
  37. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    #37 Allan Wolff, Feb 26, 2011
    Last edited by a moderator: Jan 17, 2018
    Re: Fusee for the Pinwheel Skeleton Clock

    From a drive weight of 7 pounds (112 oz) and a fusee blank radius of .928”, the torque required to drive the clock is 112 X .928 = 103.88 inch-ounces. (Remember, these values are from measurements taken for this particular clock. Different components will provide different values.) The radius of the fusee needs to be matched with the force of the spring so it always delivers 103.88 in-oz of torque to the clock. We also want to operate in the linear portion of the mainspring curve. From the spring force graph shown in the first post, the average plot is fairly straight from about 12 inches through 70 inches, so 13 inches is chosen as the starting point. This tells us that we need 13 / 3π = 1.4 turns of preload for our 3” diameter mainspring barrel. With a starting point of 13 inches of cable, we can use a spreadsheet to perform the remaining calculations. You may want to open the spreadsheet at this time and follow along as I explain the next steps. The measurements from the first post are entered in columns A, B and C. Column D converts the pounds and ounces into ounces so we can work with one unit of measure. Column E is also ounces with some of the numbers manually adjusted. Recall this adjustment was made to smooth out my shaky readings from holding the scale by hand. Obvious dips and peaks are manually averaged by observing a plot of the values. Column F then takes these adjusted values and performs a 9-sample moving average of the adjusted values in column E to provide additional smoothing. Column G is the radius of the fusee required to deliver 103.88 in-oz of torque from the force available from the mainspring at that point. For example, at 13 inches of cable, 148 ounces of force is provided by the mainspring, so 103.88 / 148 = .701 inches for the fusee radius. Column H is the cumulative number of turns the fusee will make based on the previously calculated radius. One inch of cable and a radius of .701” equates to 1/(2πr) = 1/(2π(.701)) = .23 turns of the fusee. At the next inch increment, the new radius of .686” is used to obtain an increment of .23” and that increment is added to the previous value to obtain a cumulative total of .46 turns. This provides a stepped approximation to the actual number of turns of the fusee. Now that we have a method for determining how much the fusee turns for each incremental inch of cable, we can calculate how far along the fusee each increment will be. The fusee groove will be cut at 13 turns per inch so the groove will travel 1/13 = .077” per turn. Column I shows the amount of travel at each increment. This is calculated as .077 times the number of turns. For example, inch 14 is .077 X .23 = .017 for the fusee width. (Note: the spreadsheet calculates without rounding, so the numbers will vary slightly from those shown here. The rounding is not cumulative and is of no concern.)
    The fusee will be cut using an incremental step method as described in volume 1 of Guy Lautard’s Machinist’s Bedside Reader. This method entails making a series of plunge cuts with a parting tool leaving a step pattern that approximates the fusee shape. The steps are then filed off to create the final shape. Cuts can be made for every inch of cable increment, but it is more efficient to skip every other step at the expense of a little more filing. Column J is the step increment. This is the width of each step and determines how much the lathe carriage should be advanced from its previous position. It is the difference in fusee width in column I from the previous cut. For example the difference in fusee width from inch 15 and inch 17 is .072 - .035 = .037. Column K is the depth increment and determines how much the cross slide will be advanced from its previous position. It is the difference in fusee radius in column G from the previous cut. For example the difference in fusee radius from inch 15 and inch 17 is .675 - .652 = .023 (the table shows .022 in column K for inch 17, more rounding variation). The step and depth increment for inch 13 are highlighted in blue to indicate this is the starting position. The diameter shown in column L is simply twice the radius of column K plus the thickness of the fusee cable (.046”) and is provided as a convenience for checking dimensions while cutting. It is necessary to add the thickness of the fusee cable since diameter will be reduced by the same amount when the groove is cut. We now have all of the information needed to cut the final fusee shape.
    upload_2018-1-17_11-1-26.png
     

    Attached Files:

  38. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Fusee for the Pinwheel Skeleton Clock

    The incremental cut method used to form the fusee curve is more confusing to explain than it is to actually do. So, I will attempt to explain the process and then follow up with a bunch of photos.

    In the first photo (which is not drawn to scale), the start position on the upper right corresponds to the first row of data for columns J, K and L from the spreadsheet of the previous post where the diameter equals 1.449”. The carriage is then moved left .035” and another cut is taken .026” deeper than the previous. The carriage is then moved left .037” and another cut taken .022” deeper than the previous; and so on. The blue stair-step line in the first photo is the result. After coating the work piece with layout dye and filing until the dye is removed, the red line in the first photo remains which is the actual fusee curve.
    Here is the machining process.
    The fusee assembly is returned to the lathe with the long end of the arbor in the chuck just like it was when the fusee blank was trued in the second post. A dial indicator is used to verify that it runs true. Masking tape is placed over the exposed pivots to protect their surface. With a dial indicator mounted to measure carriage position, (lower left of the second photo) the right edge of the parting tool is offset .150” from the right edge of the fusee blank and the carriage is locked in place. This will be the shroud for the click and ratchet. A plunge cut is made with the parting tool until the diameter measures 1.449”. The reading on the cross slide dial is noted before backing out. Let’s say it is 8 for our example. The carriage is then moved to the left by the amount shown in column J for inch 15, .035” in this case. The next cross slide reading is calculated by adding the amount shown in column K, .026” in this case, to the previous cross slide reading of 8. For our example the new cross slide reading will be 26 + 8 = 34. Write this number down so it is not forgotten. Make another plunge cut until the cross slide reaches the depth of the previous cut and then continue until the cross slide dial reads the calculated value of 34. This process is repeated until the end of the table is reached. It may be necessary to remove excess material from the fusee blank if the parting tool cannot cut to sufficient depth. This is shown in the third photo. If this is required, make a note of where you left off. Also, remember to adjust your cross slide dial calculations when crossing zero. For example, the next cut will add 22 to the present cross slide reading of 34 equaling 56. Since the Taig dial rolls over at 50, subtract 50 from 56 so the new dial reading is 6. Again, write these numbers down since it is easy to lose your place. Photos 2, 3 and 4 show the fusee cutting progress. The final shape is achieved by coating the fusee with layout dye and then filing the surface with a half-round file until the dye is removed as shown in photos 5 and 6.

    Next up, the fusee grooves will be cut. 85351.jpg 85352.jpg 85353.jpg 85354.jpg 85355.jpg 85356.jpg
     
  39. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Fusee for the Pinwheel Skeleton Clock

    Virtually every source of information I have found on grooving a fusee uses a lathe with thread cutting capabilities. The Taig lathe is not configured to cut threads so another method must be used to move the carriage in sync with the spindle. If your lathe has thread cutting capabilities, feel free to skip ahead. The first photo shows a drive center made from ½” threaded rod with 13 threads per inch. This drive center is held in a 4-jaw chuck. The 60-degree point of the drive center is trued each time the center is installed on the lathe. A linkage made from a nut welded to a ¼” diameter rod connects the drive center to the carriage. The rod is secured to the carriage with machinist’s clamps. The cross slide nut is removed from the cross slide so the cutting tool can be fed into the fusee by hand.

    A special form tool shown in photos 2 & 3 is used to cut the groove. The main body is made from ¼” key stock drilled approximately ¾” deep to accept a section of .055” music wire. The profile of the body is shaped before inserting the wire. It is ground back on the right hand side so it does not rub on the fusee blank. It can be difficult to produce a groove of consistent depth when feeding the tool into the fusee by hand. To solve this problem, the left shoulder acts like a depth stop, preventing the tool from cutting too deep. The shoulder has a polished edge to provide a smooth surface to prevent scratching of the fusee. The protruding portion of music wire does the actual cutting. The end of a piece of music wire is rounded and ground to the center to make the cutter. The cutting end is sharpened and smoothed on an emery stone. The wire is fastened inside the cutter body with Loctite so it stands .025” proud of the shoulder. This is the depth we want to cut the groove.

    The fusee assembly is installed between centers with the large end towards the tailstock and verified with a dial indicator that it runs true. Lathe dogs are used to couple the fusee arbor to the drive center. Before cutting, the carriage is run from end to end to verify it will reach from one end of the fusee to the other and that the cross slide has enough movement to reach the largest and smallest diameters of the fusee surface. The motor belt is removed and a hand crank is attached to the headstock pulley. This setup is shown in the fourth photo.

    The starting position is marked with a felt pen at the large end of the fusee. Each cut will start at this mark. All cuts are made from the large end to the small. The cutter is fed into the fusee by applying light pressure to the cross slide by hand while the linkage moves the carriage from left to right. Go slow and take light cuts to prevent the cutter from digging in or chattering. This fusee required approximately 20 passes.

    After the groove is complete, the grooving assembly is removed and the fusee is mounted in the 4-jaw chuck and the large end of the fusee is trepanned to a depth of 1/8”. The rim is approximately 1/8” wide. This is similar to the machining done on the great wheel, so I will not show it again here.

    Next, the cable access hole is located in the trepanned area by placing the start mark of the fusee groove at the 11 o’clock position. A line is drawn straight down as shown in the fifth photo. The hole is drilled to a depth of approximately 1/4". I was lucky and this hole matched up with the temporary 4-40 hole used to fasten the great wheel for the previous test run. The cable hole is then drilled from the starting mark of the fusee groove to the access hole. This should be drilled at a slight angle as shown in the sixth photo to prevent breaking through the side of the fusee. This technique was suggested in one of W.R. Smith’s books. Be careful not to let the drill bit cut through the sidewall of the fusee groove. I started the hole with a small end mill since a drill bit wanted to wander down the angled fusee surface. The corner of the hole is rounded with a needle file where the cable will exit and begin tracking the groove. This will reduce bending stress on cable. The fusee is then remounted in the lathe and the small end of the fusee is trimmed to length, making sure there are at least 16 turns available on the finished fusee. Finally, the maintaining wheel and great wheel are installed on the large end and the position of the slip washer groove is marked. The groove is machined with a 1/16” wide parting tool. I will provide more information on that when we make the slip washer.

    The fusee can now be sanded to a 600-grit finish. The winding flats are machined on the fusee arbor in the same manner as the barrel arbor, except the finished flats should measure .175”. The smaller size for the fusee winding arbor is intended to prevent the winding key from fitting over the barrel arbor as both are located on the front of the clock. Accidentally winding the barrel arbor would require resetting the fusee preload. The ends of the fusee arbor can now be trimmed to their final length to remove the center holes.

    Next time the fusee ratchet, click and various other small parts will be made. 86039.jpg 86040.jpg 86041.jpg 86042.jpg 86043.jpg 86044.jpg
     
  40. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Fusee for the Pinwheel Skeleton Clock

    Because of its small size the fusee ratchet is turned from a piece of 7/8” diameter brass rod. The end of the rod is cut with the same 60-degree ratchet cutter that was used for the maintaining wheel. In this case, the teeth are cut as deep as possible, bringing them to a sharp point. The 6-degree under-cut is achieved by offsetting the cutter .045”. The three mounting holes are drilled with a #41 bit in the milling spindle mounted to the cross slide. The center hole is then drilled and reamed to size. The ratchet is then parted off to a thickness of 3/32” as shown in the first photo. If multiple ratchets are desired, cut the teeth further along the rod, drill the holes deeper and part off as many ratchets as needed. Each of the mounting holes is countersunk so a #2 flat head machine screw sits just below the surface of the ratchet. The ratchet is then slipped into position on the fusee arbor to transfer the position of the mounting holes to the fusee. The holes are drilled with a #50 bit and tapped #2-56. The ratchet then mounted to the fusee with three 3/8” long #2-56 flat head machine screws as shown in the second photo. Note that the fusee ratchet covers the access hole for the fusee cable. This was done intentionally so the cable will not interfere with the fusee components.

    The fusee click pivots on a stud made from 1/8” drill rod. The rod is threaded approximately 3/16”, then the die is reversed to chase the thread so the rod can be screwed all of the way into the maintaining wheel without any exposed threads. This allows a smooth bearing surface for the fusee click.
    The fusee click is located in the trepanned area of the fusee and space is rather tight so the click is cut from 3/32” tool steel to the approximate shape and then filed to fit. The pivot hole should be reamed to 1/8” for a close fit on the stud. Excess threads are then cut or ground flush with the maintaining wheel. The smooth section of the stud should be flush with the top of the fusee click. Test fit the fusee click into the end of the fusee and shape it with a file until the pawl fits into the teeth of the ratchet and does not interfere with the rim of the fusee.

    The shape of the fusee click spring is laid out on a piece of 1/32” brass sheet. The holes are drilled before sawing out the shape to provide a safe amount of material to clamp while drilling. After sawing, the narrow section of the spring is hammered to harden it and make it springy. It is then bent into a "C" shape. Determining the exact shape is also a trial and error process. The shape is adjusted so that it fits in the end of the fusee and provides the proper amount of tension on the click. When the final shape is achieved, the click spring is filed and sanded to a 400-grit finish. The click assembly is installed and verified that the ratchet action is satisfactory. The click and spring are shown installed on the maintaining wheel in the third photo. 86687.jpg 86688.jpg 86689.jpg
     
  41. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Fusee for the Pinwheel Skeleton Clock

    The maintaining spring shown in the first photo is made from music wire. Experimentation with several shapes and sizes indicated a wire diameter of .047” provided about the right “feel”. The actual thickness and shape may need to be adjusted for more or less force based on tests after the clock is reassembled. The loop on the free end of the spring will attach to the maintaining spring pin on the back side of the maintaining wheel. The spring as shown is fully relaxed and will come under tension when it is positioned at the slot on the top spoke. The second photo shows the spring as assembled between the great and maintaining wheels. The fusee rotates clockwise, further tensioning the spring as the pin moves to the right side of the slot under normal operation. During winding, the fusee rotates counterclockwise while the maintaining wheel is held in place by a pawl. The maintaining spring continues to push the great wheel clockwise, providing power to run the clock while the fusee is wound.

    The slip washer is cut from a section of 1” diameter brass rod. The end of the rod is faced and a light cleanup pass is made on outside circumference. The edge is then filed to a 45-degree bevel. The center of the washer is located and a dimple is made with a center drill. The partially completed washer is then parted off to a thickness of 1/16”. The screw hole and clearance hole are spotted from the center dimple and all three holes are drilled through on the drill press. Removing the material between the larger holes with a jeweler’s saw then forms the slot. The saw marks are cleaned up with a file. The washer can be filed down if it is too thick. If the washer is too thin or the great wheel is loose, the washer can be domed to produce a tighter fit. This is accomplished by squeezing the washer against a pliable surface (such as leather or soft wood) and a hard curved surface as shown in the third photo. Here, a round drawer knob and soft poplar wood block are used in the bench vise to add a slight concavity to the slip washer. The curvature is increased or decreased as necessary until the great wheel moves freely with minimal play. The washer is completed by sanding to a 600-grit finish. The slip washer screw will be made later along with the plate screws.

    The maintaining wheel is sanded to remove any visible marks and bring it to the desired finish before fastening the maintaining spring pin and click stud in place with 609 Loctite.

    This completes the fusee and associated components and we now have enough information to begin work on the plates. 86779.jpg 86780.jpg 86781.jpg
     
  42. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Plates and Pillars for the Pinwheel Skeleton Clock

    It is finally time to cut out the plates!! Typically, clocks built from a “how-to” book build the plates first. This makes sense if you already know where you are going. However, when building from scratch, it is necessary to prove various concepts and the design may change as the project progresses. It is better to experiment on inexpensive strips of brass rather than a large, expensive sheet. The brass for these plates cost close to $200 two years ago.

    There is a large degree of artistic freedom that can be expressed in the plate design. The simple test frame used up to this point would serve perfectly well as a fully operational, yet very boring clock. The creators of the original clock developed a wonderfully elegant design that is worthy of duplicating. A photograph of the original clock was imported into CAD as a background image and each line on the left side was “traced” with a series of points connected to form lines that approximated the plate structure. The photo was then removed and the lines were cleaned up by moving misplaced points and substituting curves, circles or straight lines where appropriate. This proved to be a very tedious process. After the left side of the plate was completed, it was copied over as a mirror image to form the right side of the plate. This ensures the plate is symmetrical. Center holes for the pillars, mainspring barrel and fusee were added to the CAD drawing to aid in drilling the major holes. The gear train will run up the center of the plates and each pivot hole will be determined with the depthing tool.

    Before mounting the patterns, the brass sheet should be checked for flatness and adjust as necessary. Two copies of the plate are printed actual size on 11”X17” paper and laid out on 3/16” thick brass sheet. One end of the paper is taped to the brass to retain alignment. The paper is then folded back and the paper and brass are coated with a spray adhesive. Make sure the paper does not distort as it is folded back into position and adhered to the brass. After the adhesive sets, rough cut the major sections apart as shown in the first photo. Plate patterns laid out as shown to minimize waste. The black line shows where the two halves will be separated. A power saw and chain drilling was used to make this rough cut. Note the brass sheet rests on towels to protect it from scratches. Any scratches will need to be sanded out later and it is easier to avoid them in the first place.

    The outsides of the plates are cut to shape with a jeweler's saw. Because of the large number of curves and corners, it is inevitable that the saw will need to reach across large sections of the plate. For these situations, a fret saw with a 12-inch frame comes in handy. The fret saw holds standard jewelers saw blades. It takes some practice to become adept at working the long, heavy frame with the fragile blades. The fret saw and 4-inch jeweler's saw are shown in the second photo.

    After the outside is cut, the lower interior sections can also be removed since the frame structure in this area is very thick. The interior sections of the upper part of the plates will be left for the time being to prevent the thin, delicate parts of the frame from becoming bent during handling. This will also make material available in the event a component needs to be relocated during the later stages of construction.

    Approximately 15 hours and 20 blades were required to cut both plates as shown in the third photo. Ironically, most of the blades were broken backing out of a cut or when repositioning the work piece. Those equipped with suitable power equipment such as a variable speed scroll saw may choose to use it here.

    The outside edges can now be filed to remove the saw marks and make the platesmore comfortable to handle. Final finish work will not be completed until the entire plate is cut out. When working the edges, care should be exercised to keep them square and not round the corners.

    Next time I will show how I align the plates and drill some holes. 87752.jpg 87753.jpg 87754.jpg
     
  43. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Plates and Pillars for the Pinwheel Skeleton Clock

    The plates must be kept aligned with each other, or “registered” so the pivot holes line up correctly. Typically, clockmakers will drill small holes in the plates and insert taper pins to align them. This leaves unsightly holes in the plates. I use a method described by Steven Conover is his book "Building an American Clock Movement". This method uses alignment bushings made from ¼” drill rod drilled to pass a 6-32 screw and cut to a length just short of the thickness of both plates, 5/16” in this case. Drill and ream to ¼”, the lower left pillar hole on the front plate and lower right hole on the back plate. Align the plates with the patterns facing out so the holes match up and install an alignment bushing and fasten it with a screw, washers and nut. Clamp the plates together and install a second bushing in the upper right pillar hole. The alignment bushings should always be installed in the same holes for consistent results when registration is required. With two alignment bushings in place, the remaining pillar holes can be drilled and reamed. The barrel arbor hole can also be drilled and reamed to 5/16” at this time.

    The fusee pivots are also drilled and ballized at this time. You may want to spot the second arbor location from the fusee arbor with the depthing tool before drilling the fusee arbor hole as it is easier to determine the spacing from a small mark rather than a ¼” hole. Remember to draw concentric circles around the hole to confirm the drill does not wander while drilling. Also remember to remove the burr raised by the ballizing process.

    The pillars will be made next. 88244.jpg
     
  44. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    #44 Allan Wolff, Apr 9, 2011
    Last edited by a moderator: Dec 26, 2017
    Re: Plates and Pillars for the Pinwheel Skeleton Clock

    The pillar are made next. Six are required, 4 lower pillars and 2 slightly smaller ones for the upper section. Efficiency is increased when making multiple parts by performing the same operation on all parts before moving on to the next step.

    The pillars are 2.390” long from shoulder to shoulder to provide .015” (1/64”) end shake for the arbors. A 3.25” length of brass rod is cut and inserted end as far as possible into chuck. This may leave a substantial amount of overhang so light cuts are required to prevent chatter. The end is faced and the spigot is cut to diameter for a nice slip fit into a .250” reamed hole. To prevent rocking, the shoulder of the pillar is undercut slightly as shown in the first photo. A hole is center drilled and then drilled #36 to a depth of 3/4” and tapped #6-32. I hold the tap in the tail stock to ensure it goes in straight. The rod is reversed in the chuck and the second spigot is created so that the shoulder-to-shoulder distance is slightly long, approximately 2.4” works well. This process is repeated for all six pillars.

    The lathe chuck is removed and one end of the pillar is mounted in a 1/4” collet with the free end supported by a live center as shown in the second photo. This setup allows the pillar to be easily removed for measuring and returned to the lathe without re-centering. Face the shoulder to achieve the final length of 2.390” and undercut the end as before. This operations is repeated for the remaining pillars.

    A large pillar is installed and snugged tight against the collet. A parting tool is set up to cut .187” from the shoulder and the carriage is locked in place. Note the second photo shows the parting tool located at the dead center end of the pillar. This caused excessive pressure at the dead center and eroded the spigot threads. Because of this, the cut should be made at the headstock end and a live center should be used. Cutting at the headstock end also helps reduce chatter. A plunge cut is made to a diameter of .540”. Make a note of the cross slide reading. The pillar is then flipped end-for-end and the same cut is made at the other end of the pillar. It is faster to lock the carriage and flip the pillar than to reposition the carriage for each cut. Repeat the process for the remaining large pillars. Cut the small pillars in the same way except the diameter is cut to .415”.

    The center section of each pillar is machined so the diameter matches that made by the parting tool. The center section is further reduced to .500” on the large pillars and .375” on the small pillars leaving a .125” decorative band at each end.

    The ends can be rounded by filing or with a graver, but a profile tool is faster and will ensure that all of the radiuses are identical. The tool must have all grinding marks removed or grooves will be left in the finish. This tool was ground to rough shape on a bench grinder and then finished against a 1/4” diameter stone mounted in a Dremel tool. A sharp tool and slow speed will produce a very nice finish as shown in the third photo. The profile tool chatters easily, so I removed the lathe belt and turned the headstock by hand.

    The pillar is extended slightly from the collet and the surface is finished with a fine file before sanding each pillar to a 600-grit finish. It may be necessary to touch up the threads that were in contact with the live center. The completed pillars are shown in the fourth photo.

    The pillar washers will be made next.
    89077.jpg 89078.jpg 89079.jpg 89080.jpg
     
  45. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    #45 Allan Wolff, Apr 17, 2011
    Last edited by a moderator: Dec 26, 2017
    Re: Plates and Pillars for the Pinwheel Skeleton Clock

    The pillar washers and plate screws will be constructed at this time although wear and tear will be saved on these parts if store-bought washers and screws are used for test fit-ups during construction. Eight 1" diameter washers and four 7/8" diameter washers are needed. Unlike the pillars, we will complete each washer before starting the next since it is difficult to hold the washers once they are parted off of the main stock.
    A 2.5” length of 1” diameter brass rod is mounted and centered in the 4-jaw chuck. The end is faced, center drilled and drilled with a #27 bit to a depth of approximately 2 inches. The center hole is counter bored with a 3/8” end mill to a depth of .063”. The end mill leaves a flat bottom for the screw to sit on.
    The edge is chamfered to 45 degrees and the inner curved surface is roughed as shown in the first photo. The internal curved surface is then shaped with a profile tool. The second photo shows the profiling operation in progress. The work may need to be turned by hand due to the amount of overhang on the first washers. The profile cut is complete when the edges of the chamfer and inner radius meet. The washer is sanded to a 600-grit finish before parting off to .125” thick. You may want to go ahead and polish and lacquer the washers with a buff mounted in a Dremel tool before parting off. I chose to polish the parts later. This process is repeated for the remaining large washers. The same process is used to machine the smaller 7/8" washers. The same profile tool can be used to form the inner curve by rotating the tool slightly to increase the angle.

    The plate screws will be made next.
    89719.jpg 89720.jpg 89721.jpg
     
  46. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    #46 Allan Wolff, Apr 19, 2011
    Last edited by a moderator: Dec 26, 2017
    Re: Plates and Pillars for the Pinwheel Skeleton Clock

    Twelve screws are needed to fasten the plates to the pillars. After making these, you will never look at a store-bought screw in the same way again. It is amazing the amount of work required to make such a common part.

    The plate screws are machined from 3/8” diameter mild steel rod. An 8” piece of rod is held in a collet with approximately 3/4” protruding. Note that I bored out the Taig headstock to 3/8” and made a custom 3/8” collet. I have never understood why the stock Taig bore is only .343". The 4-jaw chuck can also be used although the stock must be re-centered for each screw. As with the pillars, steps were used to help speed up production. The end of the rod is faced and, with the tool bit against the face, a 1/2” thick spacer is used to set the carriage stop. The diameter of the rod is reduced to .370” for a length of 1/2”, just enough to remove the factory scale and scratches. A second spacer 3/8" thick is used to reset the carriage stop and the diameter is further reduced to .138” for a length of 3/8”. This portion is threaded #6-32 with the die mounted in a tailstock holder to ensure the threads run true. The end of the thread is lightly chamfered with a file. The screw is then parted off leaving the head .103” thick. This operation is shown in the first photo. Note the use of a homemade rear-mount parting tool is mounted on the cross slide along with the cutting tool. Not having to change tool holders really helps speed things up. Repeat these same operations for the remaining 11 plate screws. Three additional screws for the fusee iron mount, motions works bridge and pendulum suspension bracket can also be made at this time as the lathe setup is the same. The head of these additional screws is slightly smaller at .225" diameter.

    A short section of scrap rod is mounted in a collet, drilled and tapped 6-32. A shoulder is formed to allow full access to the screw head. This setup is used to finish the screw head as shown in the second photo. This is one of the three smaller head screws. A partially completed screw is installed in the fixture and tighten it with soft jaw pliers. Alternately, regular pliers can be used with a piece of brass shim stock wrapped around the screw head to prevent marking the work piece. A profile tool is used to round the screw head. The head is then sanded and polished to a mirror finish at this time. A Dremel mounted buff and Tripoli compound provide the final finish. the third photo shows the polished screw. If finishing were performed after the slot were cut, the edges of the slot would be rounded which is undesirable. The screw is removed from the fixture with the same care used to install it to prevent scratching the newly polished surface. Round and polish the heads of the remaining screws. The slot is cut with a slitting saw set up in the milling head as shown in the fourth photo. The dividing plate is attached to the headstock to prevent the screw from rotating while the slot is cut. Extreme care is required to cut the slot in the center of the screw. Even a few thousands offset will be visible to the naked eye. The carriage is advanced until the saw just contacts the center of the screw head. The cross slide is then backed off and the carriage is advanced .057” and locked in place. The screw slot the then cut in one pass by advancing the cross slide. Even though a slot is now available, the screw should be removed with soft jaw pliers to prevent any damage to the screw slot. Check the slot on the first few screws to make sure it is centered. A file can be used to remove the burr on the exit side of the slot. The completed screws are cleaned thoroughly and car wax is applied to prevent rust. To prevent damage to the screws and washers, set them aside in a safe place until final assembly. I use ordinary flat washers and store bought screws for test setups through the remainder of the build.
    90000.jpg 90001.jpg 90002.jpg 90003.jpg 90672.jpg
     
  47. mcandrew1894

    mcandrew1894 Registered User

    Sep 28, 2009
    29
    0
    0
    Country Flag:
    Region Flag:
    Re: Fusee for the Pinwheel Skeleton Clock

    Great thread Allen....keep it up!

    I'm familiar with and have used the incremental cut method for various curved surfaces. Have you tried Smith's Fusee cutter?

    Dave
     
  48. dickstorer

    dickstorer Registered User

    Oct 19, 2010
    725
    24
    18
    Male
    Retired Military
    Hampton, Georgia
    Country Flag:
    Region Flag:
    Re: Fusee for the Pinwheel Skeleton Clock

    Pardon me for a dumb question, what does "ballized" mean?

    Dick Storer
     
  49. carloclock

    carloclock Registered User

    Jul 15, 2009
    40
    1
    6
    Male
    retired
    Beautiful Liguria -Italy
    Country Flag:
    Region Flag:
    #49 carloclock, Apr 23, 2011
    Last edited: Apr 23, 2011
    Re: Plates and Pillars for the Pinwheel Skeleton Clock

    Hi Allan,
    why not to improve the look of the pillars quite simply?

    I guess that a skeleton clock deserves, other then an elaborate plate a similar elaborate pillar and to get a much more improved design, once that you respect your own philosophy "Efficiency is increased when making multiple parts by performing the same operation on all parts before moving on to the next step", is a very easy task .

    It is quite impossible for me to explain in English all the operation I made cutting the pillars for a clock of mine . I described them in a topic ( anatomia di una colonnina - pillar anatomy) in my own very small forum here ( it is written in Italian but pictures are self explicative and I hope they might be of some help for you and for other perspective enthusiast clockmakers.

    Regards.
    Carlo
    PS not sure I can share a forum link . Should not that be the case, I hope the Admin will forgive me.
     
  50. Allan Wolff

    Allan Wolff Moderator
    NAWCC Member

    Mar 17, 2005
    610
    34
    28
    Male
    Tulsa, OK
    Country Flag:
    Region Flag:
    Re: Plates and Pillars for the Pinwheel Skeleton Clock

    Carlo,
    The link you provided is perfectly within the rules of this message board. Using one of the available language translators available on the internet, I was able to read your site and found it very interesting. I look forward to reading some of the other posts.

    I chose to follow the same pillar design of the original clock, but I fully agree that these parts could certainly be more ornate than they are. I really like the design of your pillars and think your construction techniques are excellent. Please continue to share your ideas as that is what this message board is all about.
    Thanks,
    Allan
    ps: Nice Schaublin lathe. I am very envious.
     

Share This Page