Finding original roller table position

Discussion in 'Watch Repair' started by karlmansson, Jul 10, 2016.

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

    karlmansson Registered User

    Apr 20, 2013
    2,505
    86
    48
    Doctor
    Linköping, Sweden
    Country Flag:
    I've picked up some useful hints and techniques for determining which way a hairspring collet need to be rotated to overcome out of beat but I still have not found a sure fire way of replacing a roller table with precision in its original position.

    Any out of beat due to roller table misplacement can be easily corrected by the hairspring collet but the poise will be thrown off. The distances from roller jewel and balance staff are so small that a few degrees misplacement can easily go undetected, at least by me. Re-poising a wheel due to
    proper placement of the roller table seems destructive to me.

    Do you guys have any techniques for replacing with precision? Angular that is.

    Best regards
    Karl
     
  2. karlmansson

    karlmansson Registered User

    Apr 20, 2013
    2,505
    86
    48
    Doctor
    Linköping, Sweden
    Country Flag:
    This seems to me as a run of the mill procedure. Is everyone just guesstimating the position or am I asking a stupid question? Is it self evident?
     
  3. ddweave

    ddweave Registered User

    Nov 21, 2011
    5
    0
    1
    I usually make a mark on the rim of the balance wheel with a (removable) marker to denote the position of the roller jewel. I have seen some balances where a previous repair person made a file mark (permanent).

    I also mark the position of the hairspring stud on the top side of the balance.

    These things aren't perfect, but will get close enough for a start.

    If I receive a balance already disassembled, I put the roller jewel 90 degrees from the balance arms on the light side of the balance.

    Doug
     
  4. karlmansson

    karlmansson Registered User

    Apr 20, 2013
    2,505
    86
    48
    Doctor
    Linköping, Sweden
    Country Flag:
    Thanks for weighing in Doug!

    This is how I normally do it as well. Problem is the markings on the balance rim will both be removed during regular cleaning and be imprecise both in structure (too wide) and in location (gauging by eye) come overcome the problems I'm asking about. I guess there are markers with tips down to 0.2mm or so but placing that line is still going to be a problem. I'm still thinking it would amount to +-5 degrees. What are your results on poise using this method?
     
  5. Ticktinker

    Ticktinker Registered User
    NAWCC Member

    Jul 7, 2015
    1,067
    15
    38
    Male
    Construction Inspector
    Houston Texas
    Country Flag:
    Region Flag:
    Karl,
    Isn't there a way to do this on a lathe?
    Place a jeweling stake tool in a collet, in the head stock, and place the roller and jewel facing into the tool,
    Place a tiny straw, like from a plastic ear swab (Q tip) into a collet on the tailstock end, push your balance staff into the opening in that.
    With this sort of method, perhaps you can align the stake tool to an expansion gap in the collet, and find a way to make some sort of index mark on the opposing balance assembly.
    I look forward to hearing how you get through this.
    Good luck.
    David.
     
  6. karlmansson

    karlmansson Registered User

    Apr 20, 2013
    2,505
    86
    48
    Doctor
    Linköping, Sweden
    Country Flag:
    I like the idea David! I don't have a collet holding tailstock, unfortunately...

    I'm thinking some sort of jig for the jewelling or staking tool. Something to rest a balance arm against as a reference. Maybe a disc with graduations on a stake making it easier to read at which the roller is installed in relation to one of the balance arms. Something like the microstella tool from Rolex, not relying on gravity of course. Just a way to read angles with a higher resolution.

    https://www.cousinsuk.com/content/common/images/categoryproductsku/R8717_Pic1_cmyk.jpg
     
  7. Ticktinker

    Ticktinker Registered User
    NAWCC Member

    Jul 7, 2015
    1,067
    15
    38
    Male
    Construction Inspector
    Houston Texas
    Country Flag:
    Region Flag:
    Another Idea I have is:
    A staking anvil elevated so a jeweling stake may be placed and held close to the top face of the anvil.
    Here is the tricky part, print out or scratch a grid on the anvil like a circle bisected twice (into 4 parts).
    Parallel lines laid out at equal distances from one of the dividing lines. (To provide a set of guide lines for where the balance arm should rest as it is set down onto the roller) Position The staking tool to hold the roller just above the surface.
    Then maybe stab the assembled staff and wheel down onto the waiting roller. (on a Qtip) (official watchmaker tool x3000) :)
    Then the matter is laying the marked point of the balance down onto the cross hair. (draft a cross-hair and the parallels with a very hard, sharp drafting pencil)
    Some rodico may be needed if your roller stake is a bit over size as mine would be...
    On the image you show; I would have concerns about unwanted rotation of the parts I want to squeeze together with it. Is there some sort of mechanical consideration for that?
    Per your Idea of graduations; I have been wondering where to find or how to create a very sharp image like a hair spring with a collet on it. this would be to do hairspring shaping work. Follow the curling shape and true a hairspring.
    Another 2 cents,,,
    Dave.
     
  8. Ticktinker

    Ticktinker Registered User
    NAWCC Member

    Jul 7, 2015
    1,067
    15
    38
    Male
    Construction Inspector
    Houston Texas
    Country Flag:
    Region Flag:
    Karl I have no tailstock at all, and I have been considering getting a 90 degree plate that I would mount jigs or other things to hold work on to work along with my head stock...
    I have looked about for a tailstock and even been in contact with a expert of sorts, but they have not come back with what I need.
    Tail stock for a Rivett WW lathe.
     
  9. praezis

    praezis Registered User

    Feb 11, 2008
    413
    15
    18
    Germany
    Country Flag:
    Region Flag:
    #9 praezis, Jul 13, 2016
    Last edited by a moderator: Dec 31, 2017
    Hello Karl,

    when you claim "but the poise will be thrown off", about what dimensions are you discussing, what do you think?

    I calculated once, if gap and jewel may compensate each other. Result: they do only a bit, not fully. But the poise error is small, much less than adding the thinnest washer to a balance screw. I suppose, you cannot detect a different poise error from turning the roller by 5 degrees only.

    It would be a nice task to test the poise error with roller jewel at 0° and at 180° (worst case). Amplitude must be the same in both tests and <200° .

    Added: tool to mark the stud position (I have a similar tool but never used it).

    Frank

    tool_stud.jpg

    tool_stud.jpg
     
  10. karlmansson

    karlmansson Registered User

    Apr 20, 2013
    2,505
    86
    48
    Doctor
    Linköping, Sweden
    Country Flag:
    Thank you Frank!

    Well, if it is of no greater consequence I guess I have my procedure right there: don't worry about it :).

    My reason for asking is that I've noticed some of the balances where I've replaced staffs are out of poise after replacement. I don't know if i should chalk this up to poor roller placement or the result of previous dynamic poising.
     
  11. Skutt50

    Skutt50 Registered User

    Mar 14, 2008
    3,375
    88
    48
    Male
    Gothenburg
    Country Flag:
    I tend to agree with Frank. A minor deviation will hardly be notisable.

    The reason you find some balances out of poise can be e.g. not poised properly after previous repair, a balance arm hole that is slightly too large resulting in a small shift of mass center when staking, home made balance arbor that is not properly centered, just to mention some.

    Do you demagnetize the balance before poising? (I replaced a staff years ago and could not get it properly poised. After demagnetizing the balance, no problem!)

    You could possibly check the poise before removing the balance arbor to see if the problem was already there or if you turned it out of poise!
     
  12. DeweyC

    DeweyC Registered User
    NAWCC Member

    Feb 5, 2007
    1,493
    391
    83
    Watchmaker
    Baltimore
    Country Flag:
    i agree with this, especially the idea that the you can always check the prior work before removing the staff. Sadly, for years popular books were promoted that encouraged learners to always poise the balance, w/o first making it clear that the balance was perfectly poised at the factory and that the only way it could come out of posie was in handling during repair. It certainly did not gain or lose mass!

    First things to verify are if the roller is "close enough" and truing the balance in flat and round. If the watch has been untouched since the factory, this should suffice. Use a bell jar over the posing tool; get the balance to stick in 8 positions.

    On previous repairs it is messier. You have to look for evidence of what was done previously and make some trials.

    While checking poise using an undamaged section of the staff that needs to be replaced is a way to baseline to see if you make things worse, it is somewhat superflous. You are still going to have to repoise almost any baance that has been restaffed since the factory. For modern solid balances, I would avoid removing any material (effect on hor vs vert amplitude has been discussed previously) and would concentrate on repositioning the roller. If too much mass is removed, the watch rates will not equalize between horz and vertical.
     
  13. karlmansson

    karlmansson Registered User

    Apr 20, 2013
    2,505
    86
    48
    Doctor
    Linköping, Sweden
    Country Flag:
    The mass removal affecting differences between hor and vert was new to me at least! I remember that we had a discussion about amplitude and resonance in regards to mass but not differences between positions. I though it was the distance between regulator pins that affected that difference the most?

    Correct me if I'm wrong, but a watch that says "Adjusted to six positions" will not have a perfect poise in the balance when poised statically, will it? A balance that has been dynamically poised will have been adjusted by adding or removing weight from a perfect, static poise.
     
  14. Harvey Mintz

    Harvey Mintz Registered User
    NAWCC Life Member

    Jun 7, 2002
    524
    9
    18
    Male
    Software Engineer, retired
    South East Florida
    Country Flag:
    Region Flag:
    I keep on reading all these complicated methods of positioning the roller table and putting the watch in beat, and I'm simply astounded that you guys all like to do it the hard way. So, here's how **I** do it:
    (1) for watches with 2 arms on the balance, I position the roller jewel at 90 degrees from the arms (this has a lot more to do with being able to see things than anything else.
    (2) I put the balance cock into place without the balance
    (3) I position the balance over the cock so there's a straight line between the pallet fork pivot, the roller jewel and the balance pivot, and the balance pivot is directly above the balance jewel.
    (4) I sight down the hairspring stud to make sure it's directly over the stud hole. If it's off a bit, I rotate it until it **is** directly over the stud hole.
    (5) I assemble the balance to the cock, making sure the hairspring is between the regulator pins.
    (6) I install the cock/balance assembly and tighten the cosk screw.

    Assuming (which is sometimes dangerous unless you've properly checked) the shape of the hairspring is good, the watch will be in beat.

    The reason this works is that lining up the roller jewel in a straight line between the balance pivot and the pallet fork pivot gives you the neutral point of the pallet without actually removing power from the spring. At the neutral point, the hairspring should have no tension on it (which is what it will have, seeing how it's not actually attached to anything at the stud end), so if the stud is directly over the stud hole, the watch will be in beat.

    Remember, everything must be cleaned, oiled and the hairspring shaped properly for this to work - but if the hairspring isn't shaped properly, the watch won't run well anyway! This is just an easy way to assemble and set the beat on an otherwise good condition watch without struggling.

    (There are, of course, movements that are difficult to do this procedure on - the undersprung early Walthams, and any full plate watch that's studded to the place with an arm are examples of movements where this is difficult to do).
     
  15. DeweyC

    DeweyC Registered User
    NAWCC Member

    Feb 5, 2007
    1,493
    391
    83
    Watchmaker
    Baltimore
    Country Flag:
    #15 DeweyC, Jul 15, 2016
    Last edited: Jul 15, 2016
    Karl,

    When you change the resonance curve ("Q") you reduce the efficiency of the oscillator (balance assembly). Once you remove too much mass (whether in static or dynamic) , you can never equalize the rates between horiz and vert. because the power is not there to drive the oscillator to the amplitude needed. This became clear to us in class using Glucydor balances.

    You are correct, one adjustment for isochronism is the spacing at the regulator pins. But the amplitude has to be there and what if you have a free sprung watch?

    Static poise is the starting point for bimetallic balances with overcoil springs and new (factory new, never used) monometallic balances. It ensures you have the solid parts poised before you add the balance spring. Dynamic poise allows you to account for the changing center of mass as that spring develops. When you remove the spring of a dynamically poised assembly, the static poise will be "off".

    It is likely even slightly off with an American RR watch that has been properly maintained throughout its life due to positional adjustment. I do not think RR watches were dynamically poised (as we call it today) other than as an effect of turning the quarter screws for positional timing. When you think it through, they were doing dynamic posing, just the long way around.

    Thank you, I may have confused people around that point.

    But dynamic poise is not as simple as it may seem. You need to test 8 positions, and the pattern of increasing fast and decreasing slow rates should be orderly (fastest rate opposite slowest rate). If they are "random", there is something to be fixed first or you will be chasing the heavy spot all around the balance. Also, you need to maintain an amplitude in all the positions (I use 190; others choose 250). If the watch cannot keep the amplitude, then you have something else going on that needs to be addressed.

    Dynamic poising really became a standard with the widespread use of flat balance springs in high grade watches. Static poise is good enough in a watch with Breguet spring that is designed to develop concentrically and thus maintain a center of mass.

    Flat springs require dynamic posing for maximum performance because they develop unevenly. The availability of accurate timing instruments helped the trade with this and I think Greiner led the way.

    The old books never even discussed safe handling; just said every restaff requires poising without helping the reader think the issue through.

    Since there are numerous excellent examples of American RR watches even today, it is apparent that educated workers knew how to do the job without resorting to extreme alterations. I am convinced most of the altered balances we see are the result of people trying to learn by reading "gospels". There were many well respected schools that turned out skilled and proficient watchmakers even in 1900.

    So for static poise I do nothing for modern balances. If positional rates are bad, I would try the roller in the other two spots it could have been. Remember, if you are servicing a decent Swiss watch, the balance assembly was dynamically poised and there is very little you can do to screw that up on these. You can either seriously deform the balance (replace) or deform the spring (can only save if damage is simple). I am talking watches for performance here; not whether or not the watch can talk.

    On bimetallics, first check true and level; then look for obvious signs of alteration. Consider if it makes sense to move the roller 180 degrees.

    FWIW, I return bimetallics to static poise and then use dynamic poise procedure for positional adjustment. I would not disturb the poise on a monometallic balance until after checking positional rates and using pins or timing screws.

    Then if you understand the procedure, go on to dynamic poising for maximum performance. The owners manual for the Greiner Micromat had a very good explanation I think Witschi used to have a good tutorial as well, but they seem to have changed their website.

    I use the dynamic poising procedure to obtain positional adjustment on bimetallic watches as well.

    If a bimetallic balance obviously has signs of mass removal, I would not remove any more. And at some point, you have to consider a donor assembly if performance is preferred over originality.
     
  16. praezis

    praezis Registered User

    Feb 11, 2008
    413
    15
    18
    Germany
    Country Flag:
    Region Flag:
    Dewey,
    can you give figures?
    What difference is to be expected between roller jewel in one position and in the opposite (s/d)?

    Regards,
    Frank
     
  17. DeweyC

    DeweyC Registered User
    NAWCC Member

    Feb 5, 2007
    1,493
    391
    83
    Watchmaker
    Baltimore
    Country Flag:
    Frank,

    Given the wide variance of balance assemblies, this is not possible. It depends on the ratio of the moment of inertias of the roller and wheel. But, you can get an idea of this by looking at the dots under the balance where a three point cutter was used to remove material in poising. Very little mass can make a big difference, especailly in ladies sized watches.

    I suspect you already get this, but the real issue is for readers to understand that what appears to be an out of poise balance when restaffed may only be an artifact of trying to static poise a balance that was dynamically poised as an assembly. And if positional variances are excessive, to look first at roller placement and for a spring deformity.

    OTOH, starting over on a bimetallic balance with an overcoil spring will result in no permanent loss of done correctly; and may in fact improve the situation. Even here though, if you are confident in your technique, it is worthwhile to ignore static poise at first and to try the restaffed bimetallic assembly in the watch and check the positional variations. Then, if it is not within your standards, remove the balance and start at static poise (if variation is excessive; otherwise use dynamic poise tech (but only use tiing screws to adjust poise) to bring into tolerance). Then time by moving pairs of screws.

    Dynamic poising of modern balances is primarily a factory procedure and the warning is that the only way to impact poise on solid balances without quarter screws is mass removal, and there is a point at which too much mass is removed despite the fact you could still regulate the watch in a position. Modern qwatches with quarter screws (Rolex) can be repoised if needed; but given the robustness of these balances and the likelihood the worker on such pieces would have good technique, the need for a correction is less likely.

    Poising has been promoted by many authors as if it were a sine qua non of watchmaking, I have come to the conclusion that this is one of those situations where less is more; espcially if the worker has not thought through what happens in the life of a watch.

    The best advice for modern balances is to be certain you locate the roller in its original position when restaffing. And remember, marking the rim with 3 pt cutter removes material and not much is needed to change poise. When we were doing our work for tests, we even removed the ink from the marker we used for marking locations to avoid errors due to the mass of the ink.
     

Share This Page