# Positional Adjustment across horizontal and vertical positions

#### Slowmojoe

##### Registered User
Hi all,

I am having difficulty achieving good positional adjustments across horizontal and vertical positions. I followed the recommended procedure in Hans Jendritzki's Watch Adjustment and Dewey's article (e.g. wind at low amplitude, look for the heavy spots and adjust) and managed to achieve tight positional variance between the vertical positions. The problems kicked in when I measured the horizontal ones- they diverged significantly from the vertical positions. My best guess was that I compensated for heavy spots by lengthening the timing screws on the opposite side, which resulted in the balance becoming heavier along its rim.

Anyone care to share their dynamic poising procedure in order to avoid the above from happening? How do I ensure that horizontal and vertical adjustments converge rather than diverge? Do I just try to match up DD and PU positions and just "throw" all the error into the opposing PD position?

Jon

#### karlmansson

##### Registered User
Hey Jon! What is your delta across the vertical positions?

Do you have differing rates between the two horizontal positions or between the both horizontal and all six vertical?

If the horizontal positions differ you likely have an issue with either endstone, pivot or the balance and/or hairspring rubbing on something in one positions but not the other. All of these are usually shown through both rate variation and loss of amplitude.

If both horizontal are the same but they differ from the vertical you have a tougher nut to crack. Dewey has gone into this at length in the past but the gist of it is that is comes down to specifications from when the watch was made. Some watches cannot be adjusted to a good delta between horizontal and vertical.

What CAN be done is to alter the spacing of the regulator pins. This has to be done very carefully to maintain the proper position of the hairspring between the pins and the parallellism of the pins. What increasing the spacing does effectively is to alter the active length of the hairspring depending on the amplitude. A lower amplitude will mean it takes longer for the spring to contact the pins on each vibration and thus result in a slower rate. Vice versa, diminishing the spacing will mean less difference in the effect of the pins depending on amplitude.

So for a watch where amplitude is lower in the vertical positions (almost always the case in my experience) and higher in the horizontal the rate will most of the time be slower for the higher amplitudes and faster for the lower. To remedy that, if you widen the pin spacing the working length of the hairspring will be longer for a greater percentage of the vibration of the balance at lower amplitudes than at higher amplitudes (seeing as the spring will contact the pin at a given degree of wind/unwind of the spring and that degree will be a lower of the total at 300 deg than at 220 deg). The result will be slower rates at lower amplitudes compared to before. This is however at the cost of isochronism. You can only make this adjustment work at a given state of wind for any given watch. Making the watch perform differently at different amplitudes is the antithesis of isochronism.

Regards
Karl

Slowmojoe

#### Skutt50

##### Registered User
As Karl pointed out we need some numbers. It will help us understading your problem,

Just give us the BPS deviation in each of the standard 6 positions ........

In some situations one can play with the point-/flat-ness of the balanse pivots to even out differences in horisontal VS vertical positions.

#### John Runciman

NAWCC Fellow
NAWCC Member
Unfortunately I'm going to repeat more or less what everyone else said. It would be really nice to see numbers. Then if we dynamic poising we need even more numbers. Traditionally it's easier to dynamic poise at 180°. Now what I mean by numbers is I'm assuming you of a modern timing machine. Perfect example modern timing machine is the Chinese 1000 or the 1900. Not acceptable is phone apps and I'll just skip over why they don't work for a while. Then it be nice for numbers and hundred and 80° to show us dial up or dial down and DS when you moved the crown/in the positions the amplitude will drop to worry about it normally if you're obsessed it have four positions eight would be better for dynamic poising that means we need 10 sets of numbers eight for the pendant/crown positions and dial up and dial down.

Then I do curiosity how close together are your regulator pins? Plus of course when was the last time the watch was serviced. Then exactly which watch is this. Because unfortunately quality the watch also has an effect on all of us. Then it be really nice in addition to all those numbers I requested a low amplitude it be nice if you wind the watch up let it run about 15 minutes and six positions would be adequate for that. Then for numbers I want the rate the amplitude and the beat. Yes it sounds quite demanding but if you're doing dynamic poising you have to do these things.

Then one last question did you statically poise first before you Did the dynamic poising?

#### Slowmojoe

##### Registered User
Hi all,

Thanks for the responses so far! I am blown away by the quality of assistance rendered, clearly I have much to learn.

The watch in question is an Illinois grade 604 16s 17j, not railroad approved but I suppose capable enough of accuracy.

Karl, thanks for the information on isochronism, horizontal and vertical positioning- I will chase down those posts by Dewey. I will need to learn when to stop rather than beating the same dead horse!

John, you are absolutely right about having a systematic methodology for diagnosis. I am currently using a Chinese 1900. No known service history, watch does not seem to have been touched in a long time. Regulator pins are parallel, tight and close to the hairspring. Thanks for highlighting the need to track amplitude and beat error along with rate, I will do so in the future.

Unfortunately I am unable to satisfy everyone's need for numbers at this point- shortly after posting yesterday my DU/DD numbers went haywire, I think some dirt or magnetism on the hairspring. Instead of wasting everyone's time chasing phantoms, I'm going to scrub that sucker, demag and re-lubricate. Will keep everyone updated...

(P.S. enclosed a picture of the watch in question.)

#### John Runciman

NAWCC Fellow
NAWCC Member
No known service history, watch does not seem to have been touched in a long time.
One of the things you have to be careful with your dynamic poising is not to assume it's the magical fix for everything. For dynamic poising to work correctly the watch has to been serviced recently. The balance wheel especially the pivots have to be in good condition. The balance wheel should be round, it's very easy on bimetallic balance wheels to get squeezed and the arms are not quite where there supposed to be. So basically dynamic poising is a final step for ultra precision timekeeping.

For instance at work I had a basket case Illinois bunn can't remember if the word special was there or not. The dial feet were cut off it not break off somebody cut them off and it does look like the right dial? The canon pinion was broken and I don't remember what else was not right. Needed a new balance staff plus a heck of a lot of truing to get the balance wheel back flat again. Then static poising grossly out of poise. That required a heck of a lot of removal from multiple screws of weight on one side. I got it to where I perceived it was reasonably close.

One of the nice things about work is the nice timing machine. It's a witschi machine with a full automatic microphone. I don't think the lift angle was quite right although it did seem to have a heck of a lot amplitude. So upon static poising up with the balance back in the watch put it on the machine because I want to know where I stood as far as timekeeping goes before proceeding. So for the image attached 7 1/2 seconds for positional air seems pretty good doesn't it? It's running a little fast several timing washers and moving the regulator little solve that problem. Now I just have to figure out how to put the dial on without dial feet.

The reason I'm showing this is not to impress you with what I did but. I have a lot of experience at work with 12 size Illinois watches and my boss was obsessed with dynamic poising and I ignored Hammond statically poise. As long as the balance wheel is nice and round the 12 size has really soft arms to be very very careful how you hold the balance wheel. But I found with static poising you can get really really close unless you're super obsessed and think you want to time your watch like a chronometer static poising's going to be good enough. I often find people think dynamic poising is the solution to all the problems.

#### Slowmojoe

##### Registered User
Thanks for the information John. Out of curiosity how are you using the beat error fluctuations to improve accuracy? I suppose the fluctuations are mostly due to the orientation of the collet as well as the position of the heavy spot relative to the escape wheel?

One of the things you have to be careful with your dynamic poising is not to assume it's the magical fix for everything. For dynamic poising to work correctly the watch has to been serviced recently. The balance wheel especially the pivots have to be in good condition. The balance wheel should be round, it's very easy on bimetallic balance wheels to get squeezed and the arms are not quite where there supposed to be. So basically dynamic poising is a final step for ultra precision timekeeping.

For instance at work I had a basket case Illinois bunn can't remember if the word special was there or not. The dial feet were cut off it not break off somebody cut them off and it does look like the right dial? The canon pinion was broken and I don't remember what else was not right. Needed a new balance staff plus a heck of a lot of truing to get the balance wheel back flat again. Then static poising grossly out of poise. That required a heck of a lot of removal from multiple screws of weight on one side. I got it to where I perceived it was reasonably close.

One of the nice things about work is the nice timing machine. It's a witschi machine with a full automatic microphone. I don't think the lift angle was quite right although it did seem to have a heck of a lot amplitude. So upon static poising up with the balance back in the watch put it on the machine because I want to know where I stood as far as timekeeping goes before proceeding. So for the image attached 7 1/2 seconds for positional air seems pretty good doesn't it? It's running a little fast several timing washers and moving the regulator little solve that problem. Now I just have to figure out how to put the dial on without dial feet.

The reason I'm showing this is not to impress you with what I did but. I have a lot of experience at work with 12 size Illinois watches and my boss was obsessed with dynamic poising and I ignored Hammond statically poise. As long as the balance wheel is nice and round the 12 size has really soft arms to be very very careful how you hold the balance wheel. But I found with static poising you can get really really close unless you're super obsessed and think you want to time your watch like a chronometer static poising's going to be good enough. I often find people think dynamic poising is the solution to all the problems.

View attachment 606699

#### John Runciman

NAWCC Fellow
NAWCC Member
Out of curiosity how are you using the beat error fluctuations to improve accuracy?
The timing machine I had the watch on does show that. Personally because I'm working on vintage watches other than being reasonably close I'm not really worried about beat errors. One of the reasons you see a beat error change is that the hairspring is pulled by gravity downward and that sagging of the hairspring rotates the balance wheel and gives you a change in the beat errorr.

#### Bila

##### Registered User
I'm not really worried about beat errors. One of the reasons you see a beat error change is that the hairspring is pulled by gravity downward and that sagging of the hairspring rotates the balance wheel and gives you a change in the beat errorr.

This I find interesting John.

#### praezis

##### Registered User
Another simple reason for beat error change (expressed as "ms"): the value is dependent on the amplitude value. Amplitude lower -> B-E. higher.

The B.E stays constant if expressed as degrees, but most timing machines show "ms" only.

Frank

#### DeweyC

NAWCC Member
Several points. The first is pointed out explicitly in my article and is emphasized by John. Positional adjustment is done only after the watch is put in perfect mechanical position.

Secondly, isochronal rate is adjusted first, then horizontal then vertical positions. This is because wide isochronal rate differences will confound the vertical positional rates since vertical amplitude is always less than horizontal amplitude. Horizontal rates must be checked to ensure the watch is ready for vertical checks.

Beat error less than 1 ms will have no impact on timing. But beat error can have a significant impact on amplitude and timing.

Remember, beat error reports the difference in time for two consecutive balance vibrations to deliver impulse. One vibration is always shorter than the other. This is why visual observation of the pallet fork is only a first estimation of beat error and actual measurement is needed for final adjustment.

It is important because it plays a role in the impacts of "the natural escapement errors". For further info see Jendritski.

Because of its impact on natural escapement errors, beat error can have a significant impact on amplitude and timing.

In an isochronal system, rate is independent of amplitude. If a watch has a reasonable isochronal rate, it will keep the same rates at 270 and 200 degrees of amplitude. (full wind and 24 hours after). This differs from the ideal isochronism of physics. We do not measure the isochronal rate at 270 and 100 degrees amplitude.

Note the difference in terms of "isochronal" and "Isochronal rate". They are not the same. The former refers to the behavior of an ideal spring. The latter refers to the behavior of the entire oscillating system such as found in a watch.

Beat error changes between the limits that define isochronal rate are not due to changes in amplitude since, by definition, the rate differences between these limits (270 and 200 degrees) are essentially the same.

Once we get to lower amplitudes, we do see changes in beat error due to the system operating below the limit of concern for adjusting the isochronal rate.

As John points out, there are many things going on as the balance spring expands and contracts. The Breguet spring was an attempt to mitigate some of these. But given the ability to adjust a flat spring 2892 to within 4 seconds across positions, the benefit of the Breguet may have been overstated.

Last edited:
NoraE and Slowmojoe

#### praezis

##### Registered User
Beat error less than 1 ms will have no impact on timing. But beat error can have a significant impact on amplitude and timing.
...
It is important because it plays a role in the impacts of "the natural escapement errors".
...
Because of its impact on natural escapement errors, beat error can have a significant impact on amplitude and timing.
Is it helpful to frighten beginners with beat error? No, it is neglible and not worth too many words. Many authors even ignored it.

However there seem to be some misunderstandings.
There is no such limit as the often heard 1 ms. It is arbitrary and only means: the lower, the better. Zero is cosmetic and pleases owners of timing machines.
In fact there is a hard limit for beat error, but far far away from 1ms.

There is no "significant impact on amplitude and timing". The impact of one half swing compensates the opposite impact of the previous half swing! This includes the mentioned "natural escapement error".

For further info see Jendritski.
Very true, Dewey. I appreciate Hans Jendritzki and his books very much. In "Watch Repairer's Manual" he calls a small B.E. "negligible for timing".
In "Watch Adjustment" (I had the pleasure to contribute to the modern edition of this book) he wrote "even if the distance of the two (diagram-) lines is some millimeters, this has no influence on timing in practice and can be left".

Beat error changes between the limits that define isochronal rate are not due to changes in amplitude ... these limits (270 and 200 degrees)
Are you really sure?
In fact a beat error never changes unless you turn the collet. Unit of the real beat error is degrees. But what you call beat error (milliseconds) changes with changing amplitude and with changing bph.
It is easy to understand: Higher amplitude means the balance wheel has to run through a wider angle at same time. Only possible with higher angular speed - so the (error-)angle is passed in less time = lower "ms" value = .lower "beat error".
I can supply the equation that also shows this, if you like.

And - Jendritzki wrote exactly the same in "Watch Adjustment", directly after the above citation.

Frank

Slowmojoe

#### DeweyC

NAWCC Member
Is it helpful to frighten beginners with beat error? No, it is neglible and not worth too many words. Many authors even ignored it.

However there seem to be some misunderstandings.
There is no such limit as the often heard 1 ms. It is arbitrary and only means: the lower, the better. Zero is cosmetic and pleases owners of timing machines.
In fact there is a hard limit for beat error, but far far away from 1ms.

There is no "significant impact on amplitude and timing". The impact of one half swing compensates the opposite impact of the previous half swing! This includes the mentioned "natural escapement error".

Very true, Dewey. I appreciate Hans Jendritzki and his books very much. In "Watch Repairer's Manual" he calls a small B.E. "negligible for timing".
In "Watch Adjustment" (I had the pleasure to contribute to the modern edition of this book) he wrote "even if the distance of the two (diagram-) lines is some millimeters, this has no influence on timing in practice and can be left".

Are you really sure?
In fact a beat error never changes unless you turn the collet. Unit of the real beat error is degrees. But what you call beat error (milliseconds) changes with changing amplitude and with changing bph.
It is easy to understand: Higher amplitude means the balance wheel has to run through a wider angle at same time. Only possible with higher angular speed - so the (error-)angle is passed in less time = lower "ms" value = .lower "beat error".
I can supply the equation that also shows this, if you like.

And - Jendritzki wrote exactly the same in "Watch Adjustment", directly after the above citation.

Frank
Frank,

I find your post to be somewhat churlish. Not sure why you need to take such a tone.

You will simply have to accept that I value Tony Simonin's experience over yours. If he tells me the factories have shown empirically that beat error less than 1 ms has no impact on amplitude and timing, it is not arbitrary.

The only misunderstanding seems to be yours.

You also seem to misapprehend the meaning of isochronism. It means that the time to make a vibration is independent of the amplitude. It is definitional.

In a real oscillator, we put limits on the amplitude of concern; generally 200 and 270 degrees. We know that below 200 degrees that rate varies significantly because of all of the forces involved besides the isochronism of the spring itself.

Frank, if you wish to continue this discussion with me, you will need to stop with the attitude. I have no idea where it comes from and I see no reason for me to put up with it.

Last edited:

#### praezis

##### Registered User
Dewey,
I had to look up what "churlish" means, not part of my vocabulary. You see, I am able to translate to English, but am not sure what undertone is perceived. Additionally a culture difference: we are used to adress issues directly, without circling around. I am sorry if my words seemed to be rude, I tried to stick to the facts.

I contradict if claims are wrong in my opinion and I use to give reasons (not just hearsay if possible).
I contradict if the impression was raised, I did claim wrong facts or nonsense.

To the facts: there is nothing wrong with the advice "adjust to max. 1ms", a good rule of thumb for repair.
But it is no law: what happens if B.E. is 2 ms or 3 ms? Nothing, I say (and not only me).
And not to forget: "x ms" is a blurred value with limited meaning if not accompanied by the corresponding amplitude (I wonder if this is understood meanwhile?).

I do not know, why you refer to isochronism. It has nothing to do with the dependence of "ms" value on amplitude.

Regards, Frank

#### DeweyC

NAWCC Member
Dewey,
I had to look up what "churlish" means, not part of my vocabulary. You see, I am able to translate to English, but am not sure what undertone is perceived. Additionally a culture difference: we are used to adress issues directly, without circling around. I am sorry if my words seemed to be rude, I tried to stick to the facts.

I contradict if claims are wrong in my opinion and I use to give reasons (not just hearsay if possible).
I contradict if the impression was raised, I did claim wrong facts or nonsense.

To the facts: there is nothing wrong with the advice "adjust to max. 1ms", a good rule of thumb for repair.
But it is no law: what happens if B.E. is 2 ms or 3 ms? Nothing, I say (and not only me).
And not to forget: "x ms" is a blurred value with limited meaning if not accompanied by the corresponding amplitude (I wonder if this is understood meanwhile?).

I do not know, why you refer to isochronism. It has nothing to do with the dependence of "ms" value on amplitude.

Regards, Frank
Frank,

The quote below was gratuitous and unfounded. No one has tried to scare beginners (not even sure what that entails) and the information I provided was from authoritative sources. From your reply, I will presume it was just an unfortunate inclusion.

Is it helpful to frighten beginners with beat error? No, it is neglible and not worth too many words. Many authors even ignored it.
I know I have apologized for thoughtless sentences and will probably need to do so in the future. It is part of being human.

I readily admit my knowledge is not perfect which is why I learn from others as much as I can. You do not commit to a year at Neuchatel at 56 years of age unless you are a willing learner. (Of course, many of the younger kids think it is a casting call for a job. Saw some very interesting behavior).

As an ex Ph.D., I have no issues with having my understandings improved or corrected. Certainly, if you (or anyone) disagrees with what I write, speak up. It is the only way we all will learn. The only caveat I have is that I hold data free opinions in contempt (maybe too strong a word, but best I can do at the moment).

"In God we trust; everyone else must bring data".

Let's just move on. I am trying to emulate my old beagle who never held a grudge.

#### John Runciman

NAWCC Fellow
NAWCC Member
In fact a beat error never changes unless you turn the collet. Unit of the real beat error is degrees. But what you call beat error (milliseconds) changes with changing amplitude and with changing bph.
It is easy to understand: Higher amplitude means the balance wheel has to run through a wider angle at same time. Only possible with higher angular speed - so the (error-)angle is passed in less time = lower "ms" value = .lower "beat error".
I can supply the equation that also shows this, if you like.

And - Jendritzki wrote exactly the same in "Watch Adjustment", directly after the above citation.
I have questions and I don't mind looking stupid today? Your statement above that beat error never changes unless you turn the collet but what about when they hairspring is being pulled by gravity that rotates the balance wheel doesn't that simulates or cause the same effect as a rotating the collet? This is what I was told by someone who I assume was looking at one of those nice big models were they can show gravity pulling in the hairspring etc. so are you trying to tell us that the beat error changes only because of amplitude changes?

I'm attaching some images one of them is from witschi it explains what beat is supposedly. It looks like to me an absolute thing the timing of one side subtracted from the timing of the other produces an error in milliseconds and that's the beat it looks pretty absolute with zero reference to amplitude here.

Then there is the other one this is from a timing machine manual it shows that has amplitude changes the spacing a line changes for those not familiar with the graphical display and that would cause and numeric increase in beat. The other place where this is very common is what amplitude drops dramatically too low I've noticed everything gets magnified you'll see watches grossly out of beat and if you just get the amplitude back up to where it's supposed to be that problem with instantly go away.

Then I'm not a math whiz you can post your equation but is there some other way to explain this to people like me I'm not saying I'm not the brightest of the world but I do find a conflict with the two pictures interesting because witschi at least to me doesn't have any mention of amplitude is affecting beat yet we've all see it occur because of amplitude.

Than the noting of the reference from the book watch adjustment can you give me a clue as to what page I would find that on or what section as there's a lot of pages in that book and I'm not going to go through the entire book today not that I would like to but I have other things I'm supposed to be doing.

Slowmojoe

#### praezis

##### Registered User
Hello John,
you surely know, reason of beat error is: in rest position the roller jewel is off the line balance - pallet fork, the balance wheel is twisted by a certain angle (unit: degrees). Roughly this angle does not change without manipulating.

Of course you are right, there can be other reasons, too. A by gravity distorted hairspring or a balance wheel thrown out of static poise by dynamic poising (Dewey often wrote about it). But these are influences of lower order, I suppose.

Changing B.E. numbers with amplitude, I tried to explain without equations in a previous post (changing angular speed). All older watch repairers (like us!) who once used paper printing timing machines know the effect that the two lines get nearer with increasing amplitude. Any millimeter on the paper print represents a certain number of milliseconds (often = 1 ms).

Jendritzki: I refer to the German edition. I heard that English and French edition have very different content. My quotations were from the timing machine chapter, rather towards the end of the book.

Frank

#### DeweyC

NAWCC Member
Hello John,
you surely know, reason of beat error is: in rest position the roller jewel is off the line balance - pallet fork, the balance wheel is twisted by a certain angle (unit: degrees). Roughly this angle does not change without manipulating.

Of course you are right, there can be other reasons, too. A by gravity distorted hairspring or a balance wheel thrown out of static poise by dynamic poising (Dewey often wrote about it). But these are influences of lower order, I suppose.

Changing B.E. numbers with amplitude, I tried to explain without equations in a previous post (changing angular speed). All older watch repairers (like us!) who once used paper printing timing machines know the effect that the two lines get nearer with increasing amplitude. Any millimeter on the paper print represents a certain number of milliseconds (often = 1 ms).

Jendritzki: I refer to the German edition. I heard that English and French edition have very different content. My quotations were from the timing machine chapter, rather towards the end of the book.

Frank
What I pay attention to most (and only worry about Beat Error on initial setup) is amplitude. I do not know the standard, but if amplitude varies by more than 10 degrees in the vertical positions I look for issues. This is often a result of locks, dart or (shudder) the two-piece safety roller being slightly off.

Without consistent amplitude in the verticals, you cannot get precision timing results.

I have found that some watches, the 992B for example, are happy with a horizontal amplitude of 250 and then they go to 230 vertical. Dan Fenwick once asserted that amplitude is not everything, and I appreciate his point. It is really the difference in amplitude between vertical and horizontal and fully wound and 24 hours later.

This is precisely why 8 day chronometers fell out of favor and 8 day clocks behave the way they do.

I have been unable to ever get more than 230 degrees out of an Elgin A-11 aircraft clock and have talked with others who say the same. Yet, they appear to serve well enough.

One of the major factors in amplitude in bimetallic balances is centrifugal force. It forces the arms out at higher amplitudes. This is one reason the Guillaume chronometer balance is so effective; it is split in the middle instead of at the far ends. Hamilton upped the amplitude on the M21 because of the solid balance (forgot where I read this but it was a Hamilton doc.)

#### John Runciman

NAWCC Fellow
NAWCC Member
I have found that some watches, the 992B for example, are happy with a horizontal amplitude of 250 and then they go to 230 vertical. Dan Fenwick once asserted that amplitude is not everything, and I appreciate his point. It is really the difference in amplitude between vertical and horizontal and fully wound and 24 hours later.
As you mentioned Hamilton 992B I've attached an image. This is a watch I purchased a couple years ago at what seemed like a really nice price because the dial was considered defective. The red numerals at the five-minute mark had faded away to almost nonexistence but the movement was Listed as in whatever their terminology was basically nice condition. Other than a minor dial up and down issue it looks quite decent for a watch that I have no idea when it was last serviced. Then as expected the effect of 24 hours amplitude has dropped but the timekeeping seems quite, really hasn't changed very much has it as amplitude doesn't seem to be a major factor for at least this watch.

DeweyC

#### John Runciman

NAWCC Fellow
NAWCC Member
Hello John,
you surely know, reason of beat error is: in rest position the roller jewel is off the line balance - pallet fork, the balance wheel is twisted by a certain angle (unit: degrees). Roughly this angle does not change without manipulating.
Amazing I thought it was just a number on the timing machine it actually means something of real life?

#### Slowmojoe

##### Registered User
Hi all,

Thanks for the responses so far. Despite being a beginner I am always in favour of more information rather than less, because I am a firm believer of doing my own due diligence and checking the sources. No doubt the concepts are complex and well beyond my current ability, but I am no stranger to complexity. With time and some effort i'll eventually get there!

Just a question to Dewey and Frank: I understand that a high beat error may affect the ability of the movement to restart at the lower amplitudes. I've also read through the above debate but still don't quite understand: is there any discernable relationship between beat error, and amplitude and rate in a watch with reasonable isochronal rate (by Dewey's definition, viz. a system keeping similar time between 270 to 200 degrees), and how would one describe this relationship? Leaving aside the debate on the relevance of the 1ms benchmark, i'm asking in relation to an extreme example of a movement with 9.9ms beat error. Dewey cited Simonin (and the original factory results), but I'm struggling to understand what exactly the relationship is (e.g. greater variance/deviation in amplitude and/or rate due to high beat error? Or lower amplitude across all positions? Or causing the periodic time (as defined by Jendritzki) to increase in certain positions? Etc etc.).

On the original Illinois PW, this thing is confounding me. I've rechecked and recleaned all pivots and staffs/arbors a few times, lubricated and let it settle, and this is the current stats so far at full wind:

Amplitude seems to be okay across the board, but for the life of me I cannot understand what is causing the DD/DU to diverge given their similar amplitude. My best guess is something related to the interaction between the hairspring and regulator pins. Any thoughts on the above would be appreciated.

One final question- as Karl highlighted above, not all watches are capable of high accuracy. Given that this is an Illinois grad 604 (not RR approved), is there any point in tweaking this further? I am not sure how much more accurate this particular movement can get.

Thanks all for indulging in my questions above.

Jon

#### John Runciman

NAWCC Fellow
NAWCC Member
Karl highlighted above, not all watches are capable of high accuracy. Given that this is an Illinois grad 604
If you read Jendritzki Book on timekeeping you will find out something interesting. There is a heck of a lot of things that will screw up timekeeping. Some of those things you can change some you cannot. So as a generalization modern watches will typically keep better time than older watches because the knowledge to make a watch to keep time has improved with time.

In the case of an American pocket watch the theories and knowledge at the time the watch was made are probably not quite the same as they are today. Then I don't know the year your watch was made but a long time ago that means it's been the hands of a lot of people. Pocket watches have things that can be adjusted like banking pins mean time screws and all sorts of other things that will skip over if you'd like to sleep tonight. Then of course because they can be moved adjusted whatever conceivably they have been. But if you have time and patience you can hopefully fix all of the undesirable things that may or may not have happened to this watch and it should still keep reasonable time. But this doesn't necessarily mean you can take a pin lever Timex and have it keep the same time as a chronometer grade Rolex. But still pocket watch like this is capable keeping reasonable time.

Slowmojoe

#### gmorse

NAWCC Member
Hi Jon,

...but for the life of me I cannot understand what is causing the DD/DU to diverge given their similar amplitude. My best guess is something related to the interaction between the hairspring and regulator pins. Any thoughts on the above would be appreciated.
My initial reaction to these figures is that the states of the staff pivots and their jewels, including the endstones, require extremely close examination, because the pivots are running on their cylindrical portions in vertical positions, (or at least they should be), but on their tips when horizontal. It's to be expected that vertical positions will show some differences from horizontal ones, simply due to the mechanics of it, but reducing the variance is what it's all about, as I'm sure Dewey will confirm.

As a side question, have you tried checking the timing figures without any lubrication on the staff pivots?

Regards,

Graham

DeweyC

#### John Runciman

NAWCC Fellow
NAWCC Member
Jendritzki: I refer to the German edition. I heard that English and French edition have very different content. My quotations were from the timing machine chapter, rather towards the end of the book.
I'm attaching an image out of the book I think this is the image that is referenced up above in several of our responses.

One of the things I find interesting with the book is the cause and effect of various things. This means various subjects to be covered all the way across the book as to how and why they can affect timekeeping. Then we have the above image which does it really fall in the category of copious quantities of explaining cause-and-effect.

Quoting from the attached image there is the reference to beat and the spacing of the lines. it says" however, even if the distance between the lines is several millimeters the beat error will have no effect on the rate of the watch." This might seem like a problem in that we would need access to a paper tape machine and a modern digital machine to figure out what the line spacing is. Then there is the other problem what is the definition of several milliseconds? I am going to assume several is more than one let's side perhaps it's 2 milliseconds. In the early days of timing machines Greiner produce some really outstanding manuals for their machines. In that a sizable part of the first section made the assumption that the watchmaker acquiring a new timing machine didn't actually understand how a watch worked as far as timekeeping goes and it's a beautiful manual on understanding timekeeping of watches. Then there's an interesting bonus for us it even explains how to measure amplitude on a paper tape machine. I'm going to skip over the hot but in the back of the book is a transparent template that you would lay over your paper tape to measure the milliseconds of the lines. Technically there are three templates as the milliseconds spacing changes the various Rates ofthe machine. For the 18,000 beats per hour it works out really nice that is 1 mm/ms. I guess simplistically this is saying that 2 ms has zero effect on the rate.

He goes on to say that the beat error spacing lines is affected by amplitude without really explaining at all which of course conflicts with witschi. Witschi has zero reference to amplitude and just absolute numbers of this is what beat areas.

#### praezis

##### Registered User
Congrats, John, you found it, that was the text from where I quoted.

This might seem like a problem in that we would need access to a paper tape machine and a modern digital machine to figure out what the line spacing is.
Both work basically in the same way. Usually the witdh of the paper tape / screen display corresponds to 1/5 or 1/6 of the beat period (= 200 ms @ 18000).

He goes on to say that the beat error spacing lines is affected by amplitude without really explaining at all which of course conflicts with witschi. Witschi has zero reference to amplitude and just absolute numbers of this is what beat areas.
Witschi needs to sell, their publications have to support this goal. They will not point to any limitation of their products.

Frank

#### DeweyC

NAWCC Member
I'm attaching an image out of the book I think this is the image that is referenced up above in several of our responses.

One of the things I find interesting with the book is the cause and effect of various things. This means various subjects to be covered all the way across the book as to how and why they can affect timekeeping. Then we have the above image which does it really fall in the category of copious quantities of explaining cause-and-effect.

Quoting from the attached image there is the reference to beat and the spacing of the lines. it says" however, even if the distance between the lines is several millimeters the beat error will have no effect on the rate of the watch." This might seem like a problem in that we would need access to a paper tape machine and a modern digital machine to figure out what the line spacing is. Then there is the other problem what is the definition of several milliseconds? I am going to assume several is more than one let's side perhaps it's 2 milliseconds. In the early days of timing machines Greiner produce some really outstanding manuals for their machines. In that a sizable part of the first section made the assumption that the watchmaker acquiring a new timing machine didn't actually understand how a watch worked as far as timekeeping goes and it's a beautiful manual on understanding timekeeping of watches. Then there's an interesting bonus for us it even explains how to measure amplitude on a paper tape machine. I'm going to skip over the hot but in the back of the book is a transparent template that you would lay over your paper tape to measure the milliseconds of the lines. Technically there are three templates as the milliseconds spacing changes the various Rates ofthe machine. For the 18,000 beats per hour it works out really nice that is 1 mm/ms. I guess simplistically this is saying that 2 ms has zero effect on the rate.

He goes on to say that the beat error spacing lines is affected by amplitude without really explaining at all which of course conflicts with witschi. Witschi has zero reference to amplitude and just absolute numbers of this is what beat areas.

View attachment 607867
Congrats, John, you found it, that was the text from where I quoted.

Both work basically in the same way. Usually the witdh of the paper tape / screen display corresponds to 1/5 or 1/6 of the beat period (= 200 ms @ 18000).

Witschi needs to sell, their publications have to support this goal. They will not point to any limitation of their products.

Frank
John and Frank,

The discrpancyis easy to explain withou resorting to conspiracy theories.

Jendritski empahsizes theory. The manuals (Greiner, Withschi) empahsize real world osicllators.

We assume the reference balacne in the vibrating tool is ideal. It is fully isochrhonal (and not influenced by an escapement). When a balacne assembly is vibrated against it, the standard I was taught was 1 minute in sycnchronization for a precision timepiece.

This is difficult because the test assembly will come out of snyc eventually. Why? Because it is not as iscohronal as the reference balance in the tool! And this assumes our reference is perfect; yet there is not an equivalent of the Invar meter locked up at the international institute of standards against which out "perfect" reference is compared.

So in reality, we know that as the balance assemblies in our watches fall in amplitude, they change the behavior of their vibrations (remember, one vibration wants to to be longer than the other); hence beat error varies at low amplitudes.

I see no discrepancy and no consipacy to sell machines.

John, BTW, I have that Greiner manual and I agree it is an excellent reference. I really did like the Micromat.

#### DeweyC

NAWCC Member
Hi all,

Thanks for the responses so far. Despite being a beginner I am always in favour of more information rather than less, because I am a firm believer of doing my own due diligence and checking the sources. No doubt the concepts are complex and well beyond my current ability, but I am no stranger to complexity. With time and some effort i'll eventually get there!

Just a question to Dewey and Frank: I understand that a high beat error may affect the ability of the movement to restart at the lower amplitudes. I've also read through the above debate but still don't quite understand: is there any discernable relationship between beat error, and amplitude and rate in a watch with reasonable isochronal rate (by Dewey's definition, viz. a system keeping similar time between 270 to 200 degrees), and how would one describe this relationship? Leaving aside the debate on the relevance of the 1ms benchmark, i'm asking in relation to an extreme example of a movement with 9.9ms beat error. Dewey cited Simonin (and the original factory results), but I'm struggling to understand what exactly the relationship is (e.g. greater variance/deviation in amplitude and/or rate due to high beat error? Or lower amplitude across all positions? Or causing the periodic time (as defined by Jendritzki) to increase in certain positions? Etc etc.).

On the original Illinois PW, this thing is confounding me. I've rechecked and recleaned all pivots and staffs/arbors a few times, lubricated and let it settle, and this is the current stats so far at full wind:

View attachment 607863

Amplitude seems to be okay across the board, but for the life of me I cannot understand what is causing the DD/DU to diverge given their similar amplitude. My best guess is something related to the interaction between the hairspring and regulator pins. Any thoughts on the above would be appreciated.

One final question- as Karl highlighted above, not all watches are capable of high accuracy. Given that this is an Illinois grad 604 (not RR approved), is there any point in tweaking this further? I am not sure how much more accurate this particular movement can get.

Thanks all for indulging in my questions above.

Jon
Jon,

Very nice. I agree that DU/DD difference is annoying. I am inclined to think it is at the pins as well. If it was endshakes changing the locks, I would expect a change in amplitude. Same with pivot ends.

What happens to the relative rates if you move the regulator 20 seconds?

Slowmojoe

#### praezis

##### Registered User
Dewey, I surely did not assume any conspiracy. But if my product shows B.E. as "ms", I would not emphasize on the existence of the unit "degree".
Also agreed that most users don't care.

Like you, I also assumed that the balance wheel of the vibrating tool is isochronous if not perfect. Until I tested it with optical sensor and timing machine! You would be astonished how far from isochrone it is - like in the picture, tested by others. But I should have known, remebering theory.

Frank

#### DeweyC

NAWCC Member
Dewey, I surely did not assume any conspiracy. But if my product shows B.E. as "ms", I would not emphasize on the existence of the unit "degree".
Also agreed that most users don't care.

Like you, I also assumed that the balance wheel of the vibrating tool is isochronous if not perfect. Until I tested it with optical sensor and timing machine! You would be astonished how far from isochrone it is - like in the picture, tested by others. But I should have known, remebering theory.

View attachment 607874

Frank
Frank,

I now understand.

#### John Runciman

NAWCC Fellow
NAWCC Member
I agree that DU/DD difference is annoying. I am inclined to think it is at the pins as well. If it was endshakes changing the locks, I would expect a change in amplitude.
I thought I would point out that currently I'm being distracted even though it's something relevant for this conversation.. I have an interesting boss at work he is able to service a Rolex regulated to a second a day although that's not actually how Rolex does the regulation. Then he likes to show off with his superior knowledge of just enough knowledge to be dangerous. As He's the boss I can either fight him or Try to ignore him. But his theory on hairsprings have to breathe between the regulator pins is conflicting with me trying to get a watch to keep time. I am Going to conveniently skip over the particular watch which Was really really was irritating me.. This is why I went through the entire book casually last night Post-it notes on every single reference to regulator pins.. Then reread the entire third section on adjustment.

Relevant for this discussion as were aware regulator pins are a necessary evil. The spacing of the pins combined with amplitude is very bad. The spacing of the pins in the book is interesting as there's a difference between a flat spring and over coil spring. On a flat spring you want a small amount of spacing.. But I'm reasonably sure the pocket watch in this discussion has an over coil and with over coil the pins are supposed to be as tight as possible as long as the hairspring can slide through when regulating.. Something the original poster should check that.

Then quoting something that was convenient reference to expecting a change in amplitude that wasn't there. But how do we know that it wasn't there? Because what was there it would explain what were seeing. Modern timing machines are interesting devices but we do have to be careful with what were seeing. They do like to average thing's and if were not careful sometimes we will miss thing's. But even if you're paying attention you may still not the things. Witschi was interesting with one of our machines it's an ice machine automatic microphone you can do six position timekeeping.. They did a firmware update which I put in and it added in time plotting capability.. Time plotting is a different way to look at what the watches doing. I find at any time I have a watch that I see timing issues and it's hard to explain what I see but if I see some that doesn't look right I will run the time plot. It's not the first time I've done this long time ago in the past I was having a timing issue in time plot was the only way to see the problem.. If you're having your train fluctuations which will average out in time after tearing the watch even the timing machine will sometimes average those out. If you stare at your machine for a while though oftentimes see the numbers slowly go up and down. On a paper tape machine supposedly would a scene a sine wave and you could measure the length of that to figure out which wheel was the problem.. But a time plot makes it very dramatic for the most part there's a limit to screen size so really long term fluctuations don't always show up.. So purpose of this paragraph is maybe there is an amplitude change of the timing machine just isn't showing it in a dramatic fashion that we can see.

Someone in the group may have a better thought than I have on this.. And thinking an experiment one of the things that's what help is the amplitude right now was at 196° let all the power off wind the watch up to maybe 200°.. Leave the watch in dial up or dial down let it stabilize for maybe 15 minutes 30 minutes. Then stare at it for a while are the numbers going up and down or they rock solid. This is where you can see power fluctuations sometimes if the numbers go up and down. Then turn it over and do the other dial position and see which you see..

#### Al J

##### Registered User
My machine has several modes. The regular timing machine mode is called "Diagram", but there's one called Vario - this will tell you the average for both over the time you let the watch run, but also will tell you the minimum and maximum for each parameter over that same time. There are others also, one used for positional checks for example.

The one you are referring to is called "Trace" on mine. It plots the amplitude and rate over a longer time period, and by adjusting the time for the plot of each set of numbers, I can adjust the time frame it tests over from as short as 8 minutes, to as long as 64 hours.

It definitely helps finding periodic variations due to a wheel that is slightly out of round. Based on the period of the oscillation you can determine what wheel the issue is coming from.

#### Slowmojoe

##### Registered User
Jon,

Very nice. I agree that DU/DD difference is annoying. I am inclined to think it is at the pins as well. If it was endshakes changing the locks, I would expect a change in amplitude. Same with pivot ends.

What happens to the relative rates if you move the regulator 20 seconds?
This was very good advice Dewey, i moved the regulator forward and back, and noticed that moving the regulator faster made the timing go haywire (while the reverse did not). Upon close inspection, noticed that the regulator pins were very slightly bent. Correcting its straightness dealt with a whole host of issues.

John Runciman

#### karlmansson

##### Registered User
This was very good advice Dewey, i moved the regulator forward and back, and noticed that moving the regulator faster made the timing go haywire (while the reverse did not). Upon close inspection, noticed that the regulator pins were very slightly bent. Correcting its straightness dealt with a whole host of issues.
If the timing is affected in unexpected ways by your moving of the regulator I would suspect that the terminal curve of the hairspring is not shaped as to follow the sweep of the indicator. This would require correcting if you intend to get the watch back to original condition. Bear in mind though that this can very likely result in disaster if it is one of your first times trying such an adjustment. It is very fiddly and tiring work that requires a No. 7 tweezer and a very smooth seweing needle. You need good lighting and high magnification. The procedure itself is described elsewhere in this forum.

For the sake of leaving nothing out: the reason for your rate changing in the horizontal positions due to the regulator pins not being straight and parallel is because of the end shake of the balance and the weight of the spring causes the effective spacing of the pins to change depending on position. This could also be and indication that you need to check the end shake of the balance and hot securely the hairspring stud is being held in the cock. If the pins were not very bent the end shake should not amount to much change.

Regards
Karl

Slowmojoe

#### Allepunta

##### Registered User
also,before you start manipulating the regulating pins make 100% sure that the hairspring is not coming "sideways" from the stud, meaning your breguet curve has to be totally perpedicular to let's say the balance rim. I don't know if I have explained myself...

Slowmojoe
RULES & GUIDELINES