Rolex 3135 Beat Error Adjustment

[clockboy]

Hello John; I knew some parts (may) get changed out. Obviously that hasn't been done in my situation. Question remains whether I have a timing issue?

The watch was "floating" in oil, the train-wheels were very dirty, a new spring is fitted, so I expected the watch would run faster. Before servicing it ran +3 sec/day, now + 8 - 9 s/d (?). For some reason I can't find the BE before servicing just now, but I remember in a certain position it was around 2.5ms

 

Another parameter I like to throw into the equation; How accurate is my Watch-O-graph? It's a computer based program (written by forum member Svorkoetter) and works on the sound-card crystal. The sound-card crystal isn't that accurate and all one can do is to fill in a correction number to adjust the daily-rate. I took a quartz watch, with a known daily rate, as benchmark.

 

Also, which I like to bounce around; I red somewhere the a BE of 0 - 0.5 ms was very good, from 0.5 - 1ms acceptable.

 

Given the fact that my balance staff wasn't changed out, the unknown accuracy of the Watch-O-graph, am I not "close enough"? Obviously I'll give it another try, but if this is the best I can get..........?

In my research before i serviced my 3135 the most common wear issue is the auto rota pinion. The 3135 is an extremely well engineered watch & I would have my doubts about any wear to the balance even if it have not been serviced. You have to disassemble/assemble a Rolex to really appreciate what high end watches they are & why they cost £5k to purchase..

[clockboy]

As a reminder timekeeping is not purely the length of the hairspring and the physical characteristics of the balance wheel there's something else that screws timekeeping up that's the escapement. The escapement supplies the energy to run the balance wheel but in doing so it influences the rate of the oscillation. Then there is the other things like external influences physically moving the watch temperature changes vibrations. But the biggest one is the escapement. So your word for the day is Isochronism. So a balance wheel is considered to be isochronous when the duration is independent of the amplitude.

 

So here's a thought on why beat at anything less than zero is bad. First what exactly is the hairspring for? We think about it as part of the oscillation system but it has another purpose which is it supplies the force for unlocking the escapement. So the strange way to think about this is by being out of beat it's almost like one side of the escapement is running at a different amplitude because it doesn't have as much energy to unlock the escapement. Everything we do with the escapement is bad for timekeeping if it's not adjusted correctly.

 

So I'm attaching timing requirements for the 2230-2235 Which are almost identical to the 3135. I don't actually have a beat error for the 3135 but it should be the same. So the first criteria for the 3135 is 15 seconds and the second criteria is -1 to +5 seconds. Then to understand how Rolex comes up with those numbers I have another image. 

Excellent post John & saved to my Rolex file.

[omgiv]

I went digging through my manuals and found some information on the 3135 you might find helpful.  It is very close to the same material that John posted but from the 3135 manual.  Hope that it is helpful.  I apologize for the bad scans.

 

 

 

[Endeavor]

Excellent material Omglv, thank you very much for taking the time digging through your material ;) It's stored already, together with John's material, in my file :)

 

I seem to have indeed some timing problems, but how much exactly is a bit hard to tell as I'm investigating the accuracy of the Watch-O-graph / computer. Today I'm going to try another computer and review my measuring procedures. More to come........

What seems pretty sure though, is that the daily rate of the watch hasn't increased, as previous Watch-O-graph numbers seemed to suggest. Perhaps on the contrary, it looks like the daily rate went slightly down from +8.5s/d to below somewhere between +7  +8 s/d. Bit early days to tell exactly.......

 

If the other computer reveals the same timing problems, then I may just make the maximum BE of 0.8ms as per Rolex manual. The previous numbers I have given seemed to be about the best and I guess I have to live with those. Currently it's all "off" and I have to try to find the "sweet-spot" again. Tightening the securing clamp on top of the balance-bridge seems also of some influence.

Hopefully I hit the sweet spot soon, with a stroke of luck, otherwise it may become a long day........ It's all a matter of nerve-control :)

Edited February 23, 2016 by Endeavor

[Endeavor]

Quick question; I'm trying to "calibrate" the other computer and the "benchmark" quartz watch shows a BE of 0.2 to 0.3ms. If a quartz watch is not supposed to have a beat-error, I have to throw my hands in the air..........

With this in mind, the Rolex shows on the other computer beat errors of DD: 0.8ms, DU: 0.8ms, CU: 0.9 and CD: 1.0ms

[Endeavor]

I'm stuck and have to either, find a very cheap solution, or leave it all up to the universe........ !?

I tried another computer to run the Watch-O-Graph and found that the readings were highly unreliable. I couldn't get two the same consecutive readings and the deviation between them was quite substantial.

Back to the first computer and now the question is; how reliable is that one? I can get similar consecutive numbers, but I have no way of telling whether the beat-error shown is anywhere accurate. Like any (precision) measure instrument, it needs (yearly) calibration against a highly-accurate, calibrated source.

If I were to buy a Timegrapher, the same problem arises; it needs (yearly) calibration. OCXO (Oven controlled crystal oscillators) and fairly accurate over time, but I doubt if they are build in a Timegrapher either?

 

I do suspect that my watch has a BE problem. It seems that if I adjust the BE with DD to say 0.2ms, then in an another position the BE gets up to 1.8ms. After many attempts and adjustments, it seems that I can get reasonable equal BE numbers in all position with a BE of around 1ms. The last adjustment yielded DD: 0.5ms, DU: 0.8ms, CD: 1.1ms and CU: 0.8ms, but as soon as I tightened the clamp on top of the balance-bridge, it became DD: 1.0ms, DU:0.8, CD:1.1 and CU: 0.8ms. This seems to be the best I can get, but if the numbers are any accurate ? ....... no idea ! Only a known and calibrated source can tell.

I guess I have to except that the watch has a slight problem, after all the last service was in 2003. I also seriously doubt whether the 2003 service was done by an official Rolex (approved) watchmaker. There were problems with the watch just after I got it back and I now noticed the "excessive" oiling and it wasn't a "ghost" watchmaker; he surely left his marks behind.....

It may well be that the balance staff is still the original one, now 24 years old.

It seems that I got to the limitations of what I can/could achieve. I'm not prepared to spend any more money, which now most likely leaves me with the enjoyment of my freshly self-serviced Rolex, for what it is....... :)

Regards: Roland.

Edited February 23, 2016 by Endeavor

[jdm]

Quick question; I'm trying to "calibrate" the other computer and the "benchmark" quartz watch shows a BE of 0.2 to 0.3ms. If a quartz watch is not supposed to have a beat-error, I have to throw my hands in the air..........

 

How a quartz watch can have a beat at all? Let alone a beat error.

Edited February 23, 2016 by jdm

[clockboy]

I'm stuck and have to either, find a very cheap solution, or leave it all up to the universe........ !?

I tried another computer to run the Watch-O-Graph and found that the readings were highly unreliable. I couldn't get two the same consecutive readings and the deviation between them was quite substantial.

Back to the first computer and now the question is; how reliable is that one? I can get similar consecutive numbers, but I have no way of telling whether the beat-error shown is anywhere accurate. Like any (precision) measure instrument, it needs (yearly) calibration against a highly-accurate, calibrated source.

If I were to buy a Timegrapher, the same problem arises; it needs (yearly) calibration. OCXO (Oven controlled crystal oscillators) and fairly accurate over time, but I doubt if they are build in a Timegrapher either?

 

I do suspect that my watch has a BE problem. It seems that if I adjust the BE with DD to say 0.2ms, then in an another position the BE gets up to 1.8ms. After many attempts and adjustments, it seems that I can get reasonable equal BE numbers in all position with a BE of around 1ms. The last adjustment yielded DD: 0.5ms, DU: 0.8ms, CD: 1.1ms and CU: 0.8ms, but as soon as I tightened the clamp on top of the balance-bridge, it became DD: 1.0ms, DU:0.8, CD:1.1 and CU: 0.8ms. This seems to be the best I can get, but if the numbers are any accurate ? ....... no idea ! Only a known and calibrated source can tell.

I guess I have to except that the watch has a slight problem, after all the last service was in 2003. I also seriously doubt whether the 2003 service was done by an official Rolex (approved) watchmaker. There were problems with the watch just after I got it back and I now noticed the "excessive" oiling and it wasn't a "ghost" watchmaker; he surely left his marks behind.....

It may well be that the balance staff is still the original one, now 24 years old.

It seems that I got to the limitations of what I can/could achieve. I'm not prepared to spend any more money, which now most likely leaves me with the enjoyment of my freshly self-serviced Rolex, for what it is....... :)

Regards: Roland.

Without a reliable readings then it is best leave alone & see how it performs on your wrist.

[clockboy]

How a quartz watch can have a beat at all? Let alone a beat error.

Quartz watch movements do not register on a timographer.

[jdm]

Quartz watch movements do not register on a timographer.

 

My point exactly.

[Geo]

My head's starting to hurt!

[svorkoetter]

Endeavor PM'd me, so I'd like to address some of the points brought up above, in no particular order.

 

Quartz watch movements do not register on a timographer.

 

Of course they do. Analog quartz movements tick. If they tick, they register. However, there are some caveats to timing a quartz movement, or using a quartz watch to calibrate Watch-O-Scope (not Graph).

 

First, amplitude is completely meaningless on a quartz movement, as the ticks don't have the three sub-parts that a mechnical watch has.

 

Second, beat error is also meaningless, because a quartz movement doesn't have a mechanism that makes alternate ticks on opposing swings of a balance. All the ticks are timed by a quartz oscillator and frequency divider, so there's no concept of odd and even ticks like there is on a mechanical movement.

 

Finally, you have to measure a quartz watch over a longer period of time (I'd recommend 8 minutes, as that's one screenful on Watch-O-Scope). Most quartz movements have what's called an inhibition period, which is a period of time over which the electronics of the movement will make one correction to the rate of the watch. That correction consists of making the next tick slightly late or slightly early, so that the overall average tick rate works out to one tick per second. If you measure the rate over a length of time less than the inhibition period, it will be wrong, because the watch generally runs slightly slow or fast between corrections. If you know what the inhibition period is for your particular movement, you should test it over a period of time that is a multiple of that period. If you don't know what it is, then 8 minutes (480 seconds) is a good choice because that is a multiple of a lot of different possibilities.

 

Now, having said all that, the crystal in your sound card will not be as good as the one in an expensive timing machine (I'm less sure about the $200 eBay timegraphers). It's probably worth re-checking every month or so using the quartz watch (whose rate you should determine by checking it's drift compared to atomic time over a period of a week or so).

 

(Paraphrased) How accurate/stable is Watch-O-Scope?

 

It really depends on the accuracy and stability of your sound card crystal. If the crystal is stable, then you can correct for inaccuracy using the rate correction setting (using a known accuracy quartz watch for calibration). On the other hand, if the rate changes constantly, it will be useless for Watch-O-Scope use. However, I've never encountered a sound card that isn't stable (once the computer is warmed up). I get extremely repeatable results using Watch-O-Scope on a  variety of computers (a high-powered desktop at work, an old Pentium 4 at home, an even older Pentium 2 at home, my Atom-powered laptop, and my new Celeron desktop at home).

 

If you're seeing repeatability issues, I'd strongly suggest the problem is the watch. If in doubt, retest the quartz watch and see if you're seeing issues with that as well.

 

One area where you may see repeatability issues is with amplitude, and possibly beat error. Calculating amplitude requires that Watch-O-Scope (or any timegrapher) be able to differentiate the different parts of each tick. Calculating beat error requires that it determine the start of each tick at the same point in the tick. Both of these can be foiled if you're not getting a clean signal, since noise can confuse it. This depends a lot on the microphone you're using, how well it's in contact with the watch, and other sources of noise in the environment.

[jdm]

Of course they do. Analog quartz movements tick. If they tick, they register. However, there are some caveats to timing a quartz movement, or using a quartz watch to calibrate Watch-O-Scope (not Graph).

 

Not on the machine which I and most people here use. These are exclusively for mechanical watches.

[svorkoetter]

Yes, they're meant exclusively for mechanical watches, but I'd be really surprised if such a machine went out of its way to ignore a ticking quartz watch. It would have to stick its fingers in its ears and hum ... "I can't heeeeear you".

 

You may of course have to set the bph manually (you do for Watch-O-Scope too) since 3600bph is outside the range for most (but not all) mechanical watches.

 

Try it. I bet it will work.

Edited February 23, 2016 by svorkoetter

[jdm]

Try it. I bet it will work.

 

Not sure what are you trying to demonstrate. 

 

For quartz testing you need a quartz testing machine, or an integrated one, which is very expensive.

Not a common timegrapher. 

[svorkoetter]

I'm not trying to demonstrate anything. I'm merely refuting the assertion (made out of the blue) that a standard timegrapher won't detect a quartz watch. The original discussion was about using a quartz watch to calibrate Watch-O-Scope (and you could use it to check the accuracy of any timegrapher). You have your timegrapher set to the wrong bph in the photo above. It should be 3600, not 36000.

 

In short, no, you can't use a standard timegrapher to test and adjust a quartz watch. But you can use a known-to-be-good quartz watch to test your timegrapher. Which is what Endeavor was doing to calibrate Watch-O-Scope in order to test his Rolex.

Edited February 24, 2016 by svorkoetter

[Endeavor]

I do believe that any timegrapher picks up, via the microphone, the beat. In case of a quartz-watch 1 beat per second, or 3600 beats / hr. What it does with that beat is up to the programmed software. That a quartz-watch hasn't got a beat-error, I do understand now. Maybe I digress from the topic, but still very important to it. Any time-grapher, be it stand-alone or computer-based, needs, in order to function properly & accurately, periodically calibration.

For time, there are ample accurate sources available, like the atomic clock. As for calibrating the frequency of the crystal not so, and if the beat of a "known" quartz watch can be used, that would be a fantastic and an ample available source. It is my lack of understanding how an accurate 1 second beat gets translated back to the frequency of say a 18 MHz crystal. Once the frequency drift of the crystal, or what it has drifted from it's nominal value, is know, a correction factor can be used and the software takes care that the correct numbers are produced.

But as I said, I digress and maybe svorkoetter can enlighten me in a PM if I was close with my measurements on the Watch-O-Scope, or we could open a new topic which, I my view, could be very educational?

My apologies to svorkoetter that I used the wrong name throughout this thread for his brilliant piece of software and hardware: Watch-O-Scope. It is my lack of understanding which causes the confusion about accuracy.

Edited February 24, 2016 by Endeavor

[Geo]

I digress and maybe svorkoetter can enlighten me in a PM if I was close with my measurements on the Watch-O-Scope, or we could open a new topic which, I my view, could be very educational?

My apologies to svorkoetter that I used the wrong name throughout this thread for his brilliant piece of software and hardware: Watch-O-Scope. It is my lack of understanding which causes the confusion about accuracy.

I would recommend you open a new thread as you suggest, it could be a very interesting discussion.

[clockboy]

I think a new thread would be the correct way. I recently purchased the Henry B. Fried  book "Repairing Quartz Watches" & the conclusion is this is anther specialist area of horology that can get real complex but very interesting.

[Endeavor]

As for the accuracy and calibrating issues, I went ahead and opened a new thread in "Watch repair tools & Equipment", with the title: "Calibration of Electronic Watch Timers"; http://www.watchrepairtalk.com/topic/3713-calibration-of-electronic-watch-timers/

 

I hope it will become very interesting and educational.....

 

Roland.

[jdm]

I'm not trying to demonstrate anything. I'm merely refuting the assertion (made out of the blue) that a standard timegrapher won't detect a quartz watch. The original discussion was about using a quartz watch to calibrate Watch-O-Scope (and you could use it to check the accuracy of any timegrapher). You have your timegrapher set to the wrong bph in the photo above. It should be 3600, not 36000.

 

In short, no, you can't use a standard timegrapher to test and adjust a quartz watch. But you can use a known-to-be-good quartz watch to test your timegrapher. Which is what Endeavor was doing to calibrate Watch-O-Scope in order to test his Rolex.

 

You're wrong. As mentioned, common timegraphers do not work with quartz watches, no matter the beat rate. 

My assertion  was not made "out of the blue", rather to correct your mistake above. which can be misleading for others.

 

Edited February 24, 2016 by jdm

[jdm]

After sitting a long time my machine actually picked a signal from an HA 8F35, just to loose it after a while"

 

 

 

Instrument calibration is done with an input on the back, not with a quartz watch.

 

Edited February 24, 2016 by jdm

[JohnR725]

The topic of discussion has brought up an interesting problem. How does any of us know if our timing machine is really on time? Then there is the other problem with the timing machine which is it tells us why watches doing at the exact instant it's on the machine does not tell us long-term how the watches going to run.

 

Then I consulted with another individual familiar with the 3135. Somewhere in the manual there is a troubleshooting guide and if they're having timing problems they replace the balanced staff. So I now have two individuals who work on these watches on a regular basis telling me that timing problems can be caused by the balance staff. The only problem is both individuals have access to the official Rolex tools and changing balance staffs in this situation is easy and quick. For everyone else is not going to be easy quick and getting a decent staff is probably going to be an issue.

 

The other solution to the problem is set the watch to a known time standard. Then as a reminder whatever the time standard is actually has to be able to keep time. Where the watch for a couple weeks see what it really does.

 

[Endeavor]

Thanks JohnR725......

As said, I adjusted the Beat Error to the best I could, with the equipment on hand (I do get a feel that the accuracy may not be that far off, but I could be wrong.......awaiting feedback on the new topic) and hope that the watch will run fine. The balance staff could be 24 years old and a change out is not in the budget...... So far the watch seems to be happy, with a constant +7 sec/day.....compared to the for sure accurate reference: the atomic clock. Perhaps I should give them a call too :)

[Endeavor]

I'm getting to the last phase of my Rolex project, and that's adjusting the daily rate. A microstella tool is on its way. Doubting a lot what has happened to the watch during the so called "official Rolex service" in 2003, but what if my balance-wheel is out of poise? Who says any adjustment to the microstella screws was done 100% correctly in 2003?

Couldn't that explain, or perhaps, next to a balance staff problem, be a part of my beat-error problem(s)?

Is there anyway to check without specific tools and with the balance still mounted in the watch?

Edited February 28, 2016 by Endeavor