the quest to get a mechanical 400 day to stay within a minute or so a year.

Discussion in '400-Day & Atmos' started by victor miranda, Jan 19, 2017.

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

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    #1 victor miranda, Jan 19, 2017
    Last edited: Jan 19, 2017
    ok... I have a big mouth...
    at this point I expect to crash and burn, and with any luck it will be entertaining.


    I said "I just want a mechanical clock that stays within a minute or so a year. minimal fuss." (I meant a 400 day clock)
    in the thread "400 day mainspring testing" url --> https://mb.nawcc.org/showthread.php?139326-400-day-mainspring-testing
    I have started saving for an atmos... the years of agony waiting will add to my joy of ownership.

    After some searching and planning,
    I have decided a quartz clock with a mainspring to run just the pendulum it the best bet.


    If I could tink up an easy around the center wheel, I would more seriously consider it.


    onward.
    If any one knows how the heat treatment on an atmos is done, I'd certainly give that a trial.
    so far I am at temp compensating pendulum. There are patents to read and consider.
    Any one here ever made a pendulum from invar? or tungsten?
    how did it go?

    here is another other... issue?
    how can you tell you are close?
    one escape click is 5 seconds
    and that is about one month's worth of error...


    I suppose I can mark one ear of the escape wheel black so I can see it.
    It should be at the top every 6 minutes.
    If you start at the top at the hour mark...
    it should be top at the top of the hour. right?


    any considerations I am leaving out?


    victor
     
  2. harold bain

    harold bain Registered User
    NAWCC Member Deceased

    Nov 4, 2002
    40,850
    140
    63
    Male
    deceased
    Whitby, Ontario, Canada
    Country Flag:
    Region Flag:
    IBM made an invar pendulum, and I have a master clock running in my garage that is good for about a minute a year with one. Has electric wind.
     
  3. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    IBM?
    a 400 day clock pendulum?

    cool either way.

    victor
     
  4. harold bain

    harold bain Registered User
    NAWCC Member Deceased

    Nov 4, 2002
    40,850
    140
    63
    Male
    deceased
    Whitby, Ontario, Canada
    Country Flag:
    Region Flag:
    I guess I misunderstood you, and thought you just wanted an accurate clock. It is a seconds beat pendulum, and a (more or less) 6 foot case
     

    Attached Files:

  5. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    :-D that is a thing o' beauty!

    victor
     
  6. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,666
    63
    0
    Calif. USA
    One minute a year is tough for any pendulum clock.
    If you get 1 minute a month on a 400 day your doing good.
    There are about 525600 minutes in a year. Even an uncompensated
    quartz clock is pushing it to get that good.
    It is about 2 parts per million.
    If you time for 6 minutes, you need to measure to 0.0002 seconds.
    Even with a stop watch, most don't do much better than about 0.2 seconds.
    The best way to do it is to keep a log. You keep track of how much
    you adjust it, how long it was since last adjusted and what the error
    was.
    You use the prior information to determine the next adjustment.
    You can use a tool called a microset but it doesn't know the average
    changes in temperature the clock will see.
    Careful logging is the only way to fine tune a clock.
    Tinker Dwight
     
  7. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    hi Tinker Dwight,

    I have a running log of my clock work (sorry for that pun)
    I'll start a page for the adjustments to keep them in one place.

    and thanks for pointing out a good direction to getting this done.

    victor
     
  8. Harry Hopkins

    Harry Hopkins Registered User
    NAWCC Member

    Nov 16, 2011
    543
    37
    28
    Retired Instrument Technician
    Mason, IL
    Country Flag:
    Region Flag:
    Victor, I admire your goal to see how accurate you can adjust a 400 day clock. I am not one of those who says that 400 day clocks are not good timekeepers.. I have about 20 of them running in my house and the ones that I am picky with are accurate to within a couple of minutes a month or better. The rest I just reset the time every week or so. I do keep a log of all my adjustments near each clock along with the last time I wound it. The problem I see with your testing is that modern 400 day pendulums simply are not made with that sort of precision in mind, there are just too many moving parts and for the most part no tension to hold them in the last position. It is even difficult to keep from changing their characteristics when you stop them to make a small adjustment. I think a disk pendulum would be much easier to adjust for long term accuracy. Even with a disk pendulum you are subject to outside influences... like temperature and barometric pressure although I know some disagree with that.

    Good luck with your testing and keep us up to date.
     
    Berry Greene likes this.
  9. doug sinclair

    doug sinclair Registered User

    Aug 27, 2000
    14,364
    48
    48
    Calgary, Alberta
    Country Flag:
    Region Flag:
    If you ever achieve such accuracy, you'll be the only person in recorded history to do it! So be certain to let us know, but only after one year of dead accuracy, without touching the clock!
     
  10. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    hi doug sinclair,
    I have read that the atmos clock can do within a minute a year.
    I thought there is little reason an ordinary 400 day clock can't be close.
    so I did a bunch of reading and searching and looking.

    from all I can tell, the quartz clock pretty much stoped all mechanical clock
    accuracy research, As a type of clock the torsion type did not get much effort in any case.

    I like 400 day clocks. Charming and quiet and goes a looong time between windings.
    If any of mine were able to stay close to time... I would not be trying this.
    they all tend to surge or lag, with a cold snap or warm week...

    I was hunting for a company that made replacement pendulums

    Picture an art deco streamlined and temperature compensated bob to swing under your
    favorite 400 day. You can point at it and say "that's my fancy"

    that hope is gone...

    One of the things about the internet is that people make all kinds of claims.
    If I get close, you will not be the only one finding it hard to believe.
    .... get in line behind me...

    I have a couple of parts clocks from e-bay and I have found a small sheet of invar.
    assuming I avoid removing a finger tip with the tin snips
    I hope I'll have a bob made in a week or two.

    I have to get some pictures up.
    and I guess I have to build some wall platforms for to display the victims.

    Hi harry hopkins,
    I have to thank for saving me a bunch of work testing mainsprings.
    I had plans to make a remontoire to test if that made any difference.
    'cause I know 400 day clocks have the occasional power problem.

    victor
     
  11. lesbradley

    lesbradley Registered User
    NAWCC Member

    Dec 20, 2007
    1,037
    4
    38
    IT analyst
    Fairford, England
    Country Flag:
    Region Flag:
    I have a Grivolas which I acquired last summer. It took me a couple of months to get it regulated but I have not touched it since September last year and it now continues to stay within half a minute either side of perfect time. Will be interesting to see how long it can maintain it. Certainly the best 400 day clock I have had for accuracy.
     
  12. Harry Hopkins

    Harry Hopkins Registered User
    NAWCC Member

    Nov 16, 2011
    543
    37
    28
    Retired Instrument Technician
    Mason, IL
    Country Flag:
    Region Flag:
    #12 Harry Hopkins, Jan 22, 2017
    Last edited by a moderator: Jan 23, 2017
    Victor, To get the accuracy you desire are you willing to sacrifice beauty over form? Les is achieving some pretty impressive accuracy with his Grivolas but it has a disk pendulum so it is a little more stable. If you are using a 4 ball pendulum you could improve your chances of success by adding tension to the pendulum as I am showing in the pictures below. With the rubber band added I think it would be wise to always be adjusting from fast toward slow as you would always be pushing the balls into more tension.
     

    Attached Files:

  13. Harry Hopkins

    Harry Hopkins Registered User
    NAWCC Member

    Nov 16, 2011
    543
    37
    28
    Retired Instrument Technician
    Mason, IL
    Country Flag:
    Region Flag:
    To all: I apologize for the large pictures in my last post. Not sure what I did wrong, I will be more careful with future posts.
     
  14. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    Hi Harry Hopkins,
    Hmmm.... am I willing to sacrifice beauty over form?

    let us explore the answer to that after I get my pendulum made...

    I like the band. darned good idea!

    victor
     
  15. Harry Hopkins

    Harry Hopkins Registered User
    NAWCC Member

    Nov 16, 2011
    543
    37
    28
    Retired Instrument Technician
    Mason, IL
    Country Flag:
    Region Flag:
    I guess I should have said are you willing to sacrifice beauty for function...

    So are you making your own pendulum?
     
  16. John Hubby

    John Hubby Principal Administrator
    Staff Member NAWCC Star Fellow NAWCC Life Member

    Sep 7, 2000
    12,077
    165
    63
    The Woodlands, TX
    Country Flag:
    Region Flag:
    Harry, wasn't your fault and I have removed the big pix. The photo insertion tool is acting up since Friday afternoon, apparently due to a short shutdown and reboot of the server and something has gone off the rails. We are working on it but the fix can't be checked until Monday morning when staff return to work at Columbia.

    In the meantime, if anyone posts photos they likely will do the same thing as happened to Harry. You can delete the big pix by editing your posts immediately after you click "Submit Quick Reply" by clicking on "Edit Post", then highlight each photo and click "delete" on your keyboard and it will go away. The small thumbnails will stay and can be opened if desired to see larger photos. Thanks to all for your patience and help while this is being repaired.
     
  17. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    #17 victor miranda, Jan 23, 2017
    Last edited: Jan 23, 2017
    Hi Harry Hopkins,

    yes I have made a pendulum from a sheet of invar 36.
    the sheet arrived about a week ago.

    I have the bob in a shatz 53 clock now.
    it is clicking at about 5.3 seconds a beat.

    I have to get a photo up here. perhaps on the morrow.

    victor
     
  18. Harry Hopkins

    Harry Hopkins Registered User
    NAWCC Member

    Nov 16, 2011
    543
    37
    28
    Retired Instrument Technician
    Mason, IL
    Country Flag:
    Region Flag:
    Victor, I didn't read close enough before to understand you were making your own pendulum. I am anxious to see what you have designed as this is also an interest of mine. What will be your method of being able to make tiny rate adjustments to the pendulum?
     
  19. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    Hi Harry Hopkins,

    I am torn between using a file and a die grinder....

    my plan was to get a ring up and going and then hang some weights on it.
    once close, lightening a weight or two.
    then.... I await changes in weather...

    victor
     
  20. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    #20 victor miranda, Jan 24, 2017
    Last edited: Jan 24, 2017
    ok not sure where this will appear.
    here is the pendulum I made in a clock it will fit...

    I spent awhile tonight trying to find how how much to add or remove to change the time
    adding a 12mm long rod of tungsten welding rod will add .05 seconds to a beat.

    I was looking for a way to slow it a bit.
    victor
     

    Attached Files:

  21. Harry Hopkins

    Harry Hopkins Registered User
    NAWCC Member

    Nov 16, 2011
    543
    37
    28
    Retired Instrument Technician
    Mason, IL
    Country Flag:
    Region Flag:
    Looks good Victor, Which plate number is your movement? Have you decided how to add weight to slow it down? It looks like it is probably a 6 second/beat clock.. Are you still running at 5.3 seconds/beat?

    By the way.. I love the base of your clock and the 4 leveling legs. I will have to be on the lookout for a Schatz with that base.
     
  22. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    :-D
    Hi Harry Hopkins
    there is one on e-bay right now with dome and another with wall sconce.

    I got that one as a 'parts' clock. the paint on the dial is crackled. (I have been looking for a solution to that issue)
    I have another movement of that type and it was not ready for the pendulum, yet more parts...
    and I have to build the dome/box my plan is glass sheets and all thread with brass sleeves.

    it is a schatz 53 type. It should be 6 sec/beat? oooops... I'll have to recount that...
    been adjusting the wrong way.

    I am now planning to add some plates and tungsten rods.

    well what is life without going the wrong way a few times?

    victor
     
  23. Harry Hopkins

    Harry Hopkins Registered User
    NAWCC Member

    Nov 16, 2011
    543
    37
    28
    Retired Instrument Technician
    Mason, IL
    Country Flag:
    Region Flag:
    #23 Harry Hopkins, Jan 24, 2017
    Last edited: Jan 24, 2017
    Just making an educated guess but I think your back plate is 1013A.. If so your clock would keep correct time at 6 seconds/beat or 10 beats/minute.
     
  24. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,666
    63
    0
    Calif. USA
    Measure the amount the dial is off rather than guess.
    Adding weight or removing weight at the outer diameter
    has most effect.
    If you keep track of where and how much you change
    the rate, you can calculate what is needed to get close
    to the right amount.
    Tinker Dwight
     
  25. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    Hi Tinker Dwight,

    yeah I know I come across as a complete mess.
    please keep in mind I have posted a specification on how much metal got me a specific change.
    so I have a ball park. I'll test again once I get some weights cut and added.
    given how much guessing I've done so far, I think I am doing ok

    I may have miscounted the clicks to move the minute hand... I've been known to multiply 2 times 3 and write down 5, too.

    I want to know why I suddenly decided to adjust the clock to 5 second beats...
    this ain't my first Schatz 53.

    I tend to re-check my work so I doubt I'd have gotten too far off the rails.
    I still think it is funny. we may have lost a few rivets in that last air pocket... but we are still flying.

    I wonder if I have a weight scale handy hmmmm....

    hi Harry Hopkins,

    I'll get you a photo of the back plate and one of the pendulum with weights added
     
  26. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    Hi Tinker Dwight

    my apologies...I see I was not as specific as I thought I was.
    I thought I typed the diameter of the weldng rods.
    the tungsten welding rod is sold as 1/16 inch diameter...
    it is 1.5 mm so it is a bit shy on the 1.6 mm it should be.

    Harry Hopkins, I left my camera next to my computer. Pictures on the morrow.

    I reassembled the pendulum with all pieces I had before I started removing parts.

    it ran at 10.6 per click cycle two ticks, it is easier for me to see. Or 5.3 sec/tick
    I added 4 approx 15mm rods and I am now 11.04 sec/cycle or 5.50 sec/tick.
    from a timed minute run 10 ticks and 55.14 sec....

    and the 4 15 mm rods get me about .05 sec/tick/rod.
    I may have to find another way to add these rods.

    victor
     
  27. MartinM

    MartinM Registered User

    Jun 24, 2011
    2,964
    79
    48
    Male
    Medical Insurance Systems Analyst
    El Dorado, CA
    Country Flag:
    Region Flag:
    You can try searching for the name, "Mademoiselle".
     
  28. KurtinSA

    KurtinSA Registered User
    NAWCC Member

    Nov 24, 2014
    3,145
    96
    48
    Aerospace Engineer (Ret.)
    San Antonio, TX
    Country Flag:
    Region Flag:
    I would think there's no better adjuster than the three point style. With the four points, it's easy to get one corner off and the whole things rocks back and forth. Can't rock with three points.

    Kurt
     
  29. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,666
    63
    0
    Calif. USA
    I'm still not sure where the rods are placed.
    The effect on timing is proportional to the distance from
    that axis times the square root of the mass, added to the
    time.
    Think of it as being C + K*D*SQR(M)
    C is the time rate without adding.
    D is the distance of the mass from the arbor.
    M is the mass of the weight.
    K is determined by experiment.
    This doesn't account for the rotational inertia
    of the mass about its own axis but for small weights
    relative to the diameter of the pendulum it is a
    good working approximation.
    Since K is an experimental value, you can measure
    your rod in inches length in place of mass.
    Do note that thoriated tungsten welding rod is radioactive.
    treat filings and such with some care.
    Dwight
     
  30. Harry Hopkins

    Harry Hopkins Registered User
    NAWCC Member

    Nov 16, 2011
    543
    37
    28
    Retired Instrument Technician
    Mason, IL
    Country Flag:
    Region Flag:
    Thanks for the heads up on the 'Mademoiselle' name Martin. I have not run across one of those before.

    Kurt, I agree that the 3 point leveling is better but I just like the look of that particular model.
     
  31. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    Hi Harry Hopkins,

    oh it is a cutie. I have two. my Sweetie call the first one "your big girl"

    my second needs the correct finial on the dome.
    ... and a proper pendulum :-D

    a search of the bay for Schatz 400 clock will find one at 225 dollars.... (it ain't mine :()
    I can't justify buying it even if I want the wall mount.

    victor
     
  32. Harry Hopkins

    Harry Hopkins Registered User
    NAWCC Member

    Nov 16, 2011
    543
    37
    28
    Retired Instrument Technician
    Mason, IL
    Country Flag:
    Region Flag:
    I have 2 Schatz Barock's which are similar but I do not have the wall shelf for either.
     
  33. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    a photo of the pendulum after I added a few weights.
    the tungsten rods are the pair over toward the right.
    the spikes center are holding the pendulum together.

    attachment.jpg

    I have it going timed at 11.8 sec/tick-pair or 5.9 sec/tick

    I think I may need to re-visit the balance of the pendulum....
    the center post is not straight up and down.

    ... iffin it was easy every one would be doing it.

    victor
     

    Attached Files:

  34. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,666
    63
    0
    Calif. USA
    I was thinking about the equation I'd posted. It it wrong because it
    doesn't include the change in mass in the square root.

    To your problem at hand, the balance.
    There are two kinds of balance, static and dynamic.
    Although, you'd think they were the same, they are not.
    A rotating object will find a different center of rotation
    than the CG if the masses are not evenly distributed.
    Tinker Dwight
     
  35. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    Hi Tinker Dwight,


    I looked over your first equation and decided
    I did not have enough real numbers to plug in to make it work.


    also and this is not something I am proud about
    I was thinking I hope I can avoid the math and algebra
    by adding enough mass to the curtain ring to get the clock to desired time.


    If you consider that my strategy so far was to make a pendulum
    that was as large a diameter as can fit under this type of clock
    and then regulate what ever way was needed after that,
    I have managed to avoid calculations. ( I hope...)


    I really want to see if invar36 solves the variable time regulation of my 400 day clocks.




    I tend to be leary of additives so my tungsten is packaged as 'pure'


    I went looking for my scale... once I find it I'll share those numbers
    the diameter of the hoop is 2.85 inches It is made of three legs 3 inches long
    and 1/2 inch tall. I'd guess the rods are less than 1 grams each or less.


    I added 10 rods to go from 5.6 to 5.9 per tick.


    so there is a curve to the additions. I expected to hit 6 seconds.


    victor
     
  36. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,666
    63
    0
    Calif. USA
    It is because the mass to time is a square root. My error in
    thinking as well.
    T = K sqr( OldMass + addedMass )
    as an example Sqr( 5 ) = 2.236
    Sqr( 5+1 ) = 2.449
    Sqr( 5+2 ) = 2.646

    The first difference is .213 while the second is .197
    So, the mass being added has less and less effect
    as the mass is increased.
    The equation for a balance wheel shows this.
    If you had the difference in the rates for different
    lengths, you could calculate the effective mass of the ring
    and spokes.
    You could use the lengths of the rods as the mass.
    Do you have the values?
    Of course, having more exacting numbers will give better
    answers. Using the dial for a few hours will improve the
    results.
    With the right numbers to work with, we should be able to
    get close enough to use something like a simple screw
    and nut to fine tune it.
    Tinker Dwight
     
  37. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    the rods are about 14mm long.

    I assembled and hung them in pairs.

    victor
     
  38. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,666
    63
    0
    Calif. USA
    You need to do at least 2 experiments with different lengths of rod.
    You also need to run the clock for a longer time. Measuring to 1/10
    second for 6 seconds is not a good enough measurement.
    You can run the clock for a couple of hours and note the time
    Actual and dial to at least 1/4 minute.
    That will give you a rate to 1 part in 500.
    Measure the rods with a caliper or micrometer. Note the total
    length for each of the two experiments.
    Remember, 1 minute a month is one part in about 40,000.
    Your not going to get that goods counting swings with a
    stop watch. Take advantage of the clock. It can run for a couple
    hours, just log the results.
    Tinker Dwight
     
  39. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    Hi Tinker Dwight,

    I need to share a bit of the rest of this.
    my plans were along the lines things have gone so far.

    I have the pendulum going at 12 sec/cycle or 6 sec a tick
    and now I am telling you minutes I have it timed at 3 minutes --> 3:00.23
    and another timed run at --> 3:00.03
    so yep I am at the end of stopwatch and Mark I eyeball and Mark I thumb.
    I have it going on a noted and timed run starting at 12:08 this morning.

    This is where I usually watch it for a week and make sure I have no flutter or power issues.
    After all, this is a new to me clock...

    I have a photo of the victims .er, parts I planned to use to make frankinclock.
    and I have not yet assembled the ... umm ... creature.
    I can't decide which face to use. I like the plain one more, and I got the fancy for this project.

    I also have not made the shelf where my test/comparison clocks will run

    attachment.jpg


    about my precision...
    like the math and algebra, I tend to avoid it If at all possible.
    again I am not proud of that tendency.

    If I have to do it, I will find the precision, I just do not see that the 400 day clock pendulum
    requires me to adjust to fine and small portions of a milimeter, yet.

    I plan to use a pair of rods as my adjustment point.
    I expect I'll file at them alternately over the next month
    to get closer to accurate.

    If it helps, I know I have to visit the pendulum balance again.
    I had the thing spinning centered and now it swings through a small arc.
    the time is good, however.

    victor
     

    Attached Files:

  40. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,666
    63
    0
    Calif. USA
    My only thought is that it is always nice to have a target value.
    One can then add a small window of error. Filing and watching can
    be significantly reduced.
    The more accurately this is known, the less frustrating it will
    be.
    It looks to be great fun anyway.
    Please keep us informed on the project.
    Tinker Dwight
     
  41. MartinM

    MartinM Registered User

    Jun 24, 2011
    2,964
    79
    48
    Male
    Medical Insurance Systems Analyst
    El Dorado, CA
    Country Flag:
    Region Flag:
    If anyone's willing to play along, I'd like to discuss the design of the pendulum victor has crafted.
    In my mind, this design would seem to be among the worst for addressing temperature compensation.
    The weight is almost entirely distributed at the rim of the pendulum and, being a ring, it will expand greatly in diameter with temperature.
    Large weights, suspended on a straight rod would be the best design I could envision without getting cute with levers and/or other compensation mechanisms.
    That said, there must be a reason the Atmos uses the same general type of design as victor.
    Thoughts?
     
  42. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    #42 victor miranda, Jan 27, 2017
    Last edited: Jan 27, 2017
    relating an un expected conversation:

    last night I was proudly showing off my creation.
    my lovely wife says to me, Are you going to brass plate it to make it pretty?

    .... ummm.

    I had no such plans ( how does one brass plate things anyway?)

    I got to thinking, ( not easy to picture )
    If it turns out that a bob made of invar fixes all the timing or regulation problems
    of an ordinary 400 day clock is anyone (other than me) going to build one?

    I find it hard to believe I am the first to make a torsion pendulum from invar.
    I also can't find any info on anyone else who has given it a trial.
    dry research papers are not found on the 'net very often I know, I just thought I find
    a pointer or two.

    the three ideas kinda lead me back to what Harry Hopkins mentions about beauty vs function.

    For the most part, 400 day clocks are intended to be pretty.

    If this works, I'll be looking into making one that is pretty.
    that takes care of me and I do wonder.

    Except for the 'barock' owners where the pendulum is largely hidden,
    is any one going to shift their 400 day clocks over to invar pendulums?
    Invar looks like steel or stainless steel.

    Hi Tinker Dwight,

    I know I could have done a bunch of calculations
    to get a good estimate ahead of my building anything for this.
    My experience is that when I do that, I get to outcomes I dislike
    most times I build nothing because I create an "impossible to build it"
    and second most often I find my skills and tools do not get me what I planned.

    I did plan, however.
    I decided I wanted as large a ring as I could reasonably fit.
    I measured my various clocks...
    The large ring was to made small weight adjustments effective.
    I hoped the ring might keep atmospheric problems to a minimum,
    even If I doubted they were a timing problem.

    an unexpected problem with the ring is it seems kinda sensitive to balance.
    While I think balance may be a problem, it might might be more of a
    handling issue... another way of saying don't move the weights! when
    attempting to adjust the regulation.

    Most of the weights are snug on the ring, so I think I will not have too many
    problems with them shifting around.

    Right now I have an ugly clock and as such, it is nothing special.
    In a few months we get to see if this idea has any merit.
    If we have a well regulated fancy sitting next to it that can't keep the same time....
    we will know for sure.

    victor
     
  43. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    #43 victor miranda, Jan 27, 2017
    Last edited: Jan 27, 2017
    Hi MartinM,

    I agree that if the metal in the pendulum were brass it would expand
    and by as much, if not more than the things one often sees on Schatz 400 day clocks.

    the pendulum is made of what was sold to me as 'invar 36" which is an alloy
    of nickel and iron plus some other metals.
    it is supposed to be free of thermal expansion from below zero Fahrenheit to over 200.
    You are on one of the less certain reasons to make it this way.
    I was thinking how can I be certain about the expansion?
    I decided the ring is easy to build and uses the least material
    .... and it will test the expansion claim.
    I have read that one can temper it for even better performance
    and little explanation on how to get the various results claimed.

    enough of invar...
    I have read that Atmos uses a brass pendulum and the spring
    is of a metal similar as invar called elinvar. The spring is heat treated to reverse the effects of the expansion
    of the brass pendulum.

    .... oh the things I want to learn.

    victor
     
  44. Tinker Dwight

    Tinker Dwight Registered User

    Oct 11, 2010
    13,666
    63
    0
    Calif. USA
    First, one of the first things I learned about things expanding in temperature is
    that a ring will expand exactly as much as a solid disk of the same outside diameter,
    for the same temperature.
    So, lets see what the math would say.
    I have two pendulums of the same material. Both have the same moment of inertia.
    One is a disk and the other is a ring.
    The moment of inertia of a solid cylinder is 1/2 MR^2
    That of a thin ring is just MR^2
    For simplicity sake, lets set the mass of the ring at one and its radius as 1.
    The moment of inertia is 1.
    We'll say the cylinders radius was 1/2. making the mass of 8 to get the same
    moment of inertia for timing purposes.
    Lets say we heated them enough to expand 1% ( really quite hot ).
    The disk changes to 1.0201 for the inertia.
    The ring is 1.0201.
    It would seem that making a ring rather than a disk it more economical
    on material but makes no difference for temperature compensation.
    Tinker Dwight
     
  45. John Hubby

    John Hubby Principal Administrator
    Staff Member NAWCC Star Fellow NAWCC Life Member

    Sep 7, 2000
    12,077
    165
    63
    The Woodlands, TX
    Country Flag:
    Region Flag:
    Victor, your statement is correct. Firstly, the Atmos suspension spring is made of ELINVAR, invented by Charles-Edouard Guillaume in 1918. The original product has zero coefficient of expansion and a zero change in torsional resistance over a wide temperature range. Guillaume received the Nobel prize in physics for this effort, and Elinvar rapidly became the global standard for manufacture of precision watch hairsprings. It has held that position as the best possible material for precision timekeeping until the recent developments using fused quartz and ceramic hairsprings (which still may have issues).

    Jean_Leon Reutter selected Elinvar as the material of preference for his Atmos clock suspension springs. However, he quickly noticed that he could not achieve the expected accuracy of his clock designs and determined that expansion and contraction of the pendulum ring and suspension tube with changes in temperature caused the inaccuracy. He first tried the use of bimetallic temperature compensating strips placed inside the pendulum rings from around 1931 to 1934, with only limited success. Along with this he also studied the properties of Elinvar and discussed its treatment with Guillaume, who advised him that the torsional resistance could be modified by heat treatment. In 1934 Reutter started experiments to treat the Elinvar suspension spring material that led to its being able to completely offset both the radial and vertical expansion and contraction of the pendulum caused by temperature changes. This practice was refined by Jaeger-LeCoultre after they purchased the Reutter business at the end of 1935 and their refinement completed by 1939. They still use the exact same means to treat the suspension wire today, I have witnessed the actual equipment and process during a visit to the Le Sentier manufacture but cannot reveal what it is. That is the "secret" that enables Atmos clocks to be regulated to within one minute per year which I have demonstrated with two of my own clocks.

    By changing the pendulum material to Invar, which has a zero coefficient of thermal expansion, you "should" be able to come close to your goal by using either Elinvar or NiSpan C suspension springs provide they have not been heat treated to change their torsional resistance with temperature. My concern about the possibility of success is that the movements of 400-Day clocks are not at all comparable in quality and low frictional resistance to an Atmos and simple mechanical issues may overcome the inherent accuracy of your pendulum-suspension combination. I will also say that without going through all the efforts and innovations you are taking, I do have a few 400-Day clocks that keep time to less than 5 minutes per year.
     
    Berry Greene likes this.
  46. MartinM

    MartinM Registered User

    Jun 24, 2011
    2,964
    79
    48
    Male
    Medical Insurance Systems Analyst
    El Dorado, CA
    Country Flag:
    Region Flag:
    I'm going to have to try this before my brain will accept it.
     
  47. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    I put a proper as from Schatz pendulum on the Mademoisele.
    Added a new spring and sanded it to get the proper time.
    It is intended as my control once I get it to lets say within 2 minutes a month

    And I made a franken-clock for the invar pendulum. I forgot to take a picture... on the morrow...

    My last time for the invar was 4 minutes lost over 38 hours.

    I found my little chemical scale and can state one tungsten pair is 22 grams
    I took off a 2.5mm bit and the scale says the assembly is 15 grams...
    hmmm. I am thinking I am for getting a decimal point there is no way the little rods
    are even close to an ounce. must be 2.2 grams and 1.5 grams.

    am also going to find a way to check that scale...
    there is no way I took a .7 gram weight off it.
    it is what I wrote, and it has to be wrong.
    I am thinking .4grams at most. I'll start with replacing the batteries.

    Hi MartinM
    I am not certain either way about expansion being even no matter what.
    I know that we can measure expansion over a length, and that ring is 9 inches long.
    So if it expands or contracts it will affect the period of the suspension spring and bob.
    to use an old engineering term, close enough, I will be able to tell.

    victor
     
  48. Harry Hopkins

    Harry Hopkins Registered User
    NAWCC Member

    Nov 16, 2011
    543
    37
    28
    Retired Instrument Technician
    Mason, IL
    Country Flag:
    Region Flag:
    I have a couple of common 400 day 'parts' clocks on the shelf that are missing pendulums. If your experiment with the Invar proves successful I can see myself making a pendulum for one of these clocks.. in fact I have already started researching the machinability of invar. No reason a pendulum that looks like stainless steel can't also be pretty. I am thinking more of a disk pendulum with adjustable weights. I have not looked into the cost of invar yet so that might be a problem. More research to do....
     
  49. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    Hi Harry Hopkins,


    the only place I have found that lists prices seems to think
    I want large amounts and have deep pockets.


    so I won't mention them...
    the other place is on the bay, search for "invar 36".


    the stuff is my idea of pricy, you may feel different.


    after my changes last night the clock has gained 4 minutes in 12 hours
    so I must have changed more than just the one weight.
    I may have moved a weight or shifted one somehow.


    I will need to revisit how I am attaching the weight blocks.


    one of the reasons I made the pendulum the way I did
    was to get a sense of how sensitive to a shift in weight
    the system is.


    one of the reasons I did not make a screw type adjuster
    is that I could not get invar rod to use to make all-thread or screws.


    I did think up one way to make an fine adjuster that avoided expansion
    of the screw material, however, I wanted to get a sense of how 'fine' the adjustment
    needed to be before I built it.


    about machining invar... I read one description that seems to cover it,
    it is gummy.
    sharp tools are good. it acts like something between steel and the
    various stainless metals I have met. I have not tried threading it.


    another metal to consider is tungsten, it has a very low expansion rate.
    ...'course it is a little tough.


    victor
     
  50. victor miranda

    victor miranda Registered User

    Jan 13, 2017
    319
    30
    28
    computer programming stuff
    Baltimore
    Country Flag:
    Region Flag:
    I just found mcmaster-carr sells invar. so there is a good option.

    I also read that invar 'grows' over time.
    I kinda wished I could find out if that process also stops at some point.

    victor
     

Share This Page