Timegrapher recommended?

Posts
216
Likes
169
Some tentative questions on timegraphers if I may?

As I seem to be getting deeper into the world of collecting vintage Omegas, and as a relative novice with a small but increasing collection, I am wondering about the merits of buying a timegrapher.

This is mainly with a view to ascertaining the health of a watch, if it may need a service, and possibly looking to attempt regulating a watch myself, where necessary.
Having read up on physical timegraphers, the android apps available and watched numerous videos I'm still unsure. Have of course been cautious of vendors reviews.....vested interests and all that!

Cutting to the chase I guess what I'm asking is; would it be worthwhile me buying a physical timegrapher or perhaps having a try with an app, or as a novice would I be better off just relying on the accuracy of a watch to make decisions.

Many thanks in avance for any advice offered.
 
Posts
21,184
Likes
48,311
I always had mixed results with various apps and eventually I bought one of the Chinese Weishi-1000 timegraphers. It should be in the range of $130-$150 shipped, maybe less if you find one on fire-sale. Don't pay more, since they are all the same.
 
Posts
582
Likes
756
Yes, I also have a Weishi and is good enough for my use as a collector and not a watchmaker
 
Posts
631
Likes
788
I haven't given in and bought one yet. I try to make do with an app (Watch Accuracy Meter) for accuracy and beat error. Amplitude is hard to measure without a timegrapher, though. I have to remove the case back and film the moment in super slow motion with my phone, then watch it slowed down to estimate the amplitude. At some point that extra effort and risk opening the case is more expensive than the cost of a cheap timegrapher.
 
Posts
148
Likes
80
I've tried pc and android based devices without a lot of success, but one of them - the Windows based Tg (free) is actually very good, and tells you more than a basic timegrapher.
As well as timing etc.it shows you the escapement locking/unlocking in a graphical form in real time. You probably need to be a professional to interpret the data properly, but it's fairly obvious if something is seriously out of whack. Running it concurrently with a timegrapher gives very close, but not identical results. I think the timegraphers have an element of smoothing in them when it comes to displaying lines, mainly because the granularity on a PC screen is much finer.

Two caveats, though. It relies on your PC clock being accurate, and anecdotally they aren't. It does have a calibration option, but that relies on you having a very accurate ticking quartz watch to do it, and you need to let calibration run for a while to get a reliable result due to the way most quartz watches work.

The second caveat is getting a good microphone, especially one that makes it possible to orientate the watch in all positions. I've managed to get a mic that's sensitive enough, but the positional aspect is tricky.

With a bespoke timegrapher you (should) get an accurate timebase, and you definitely get the proper microphone.

I've seen highly regarded watchmakers using the Chinese ones, so that's good enough for me (although they obviously wouldn't make it onto Omega's list of workshop requirements !)

If you do get one, don't get too obsessive. Even if machines show good accuracy and precision, your wearing habits will have an influence anyway.
 
Posts
9,763
Likes
15,361
The apps aren't up to much to be honest, I wouldn't even bother. The ~$100-150 ebay timegraphers work just fine on older stuff, I have one, but wont give fully accurate readings on Co-ax modern stuff, particularly for amplitude. For that you will need to spend a lot more.
 
Posts
2,439
Likes
3,895
My vintage L&R timegrapher is Kaput. It is the transistorized model. At one point I had the vacuum tube (valve) schematics, which I think went the way of the extra box of clock books. I did find a copy of those in German. There are quite a few threads and forum posts on these old tanks. My guess is that the electrolytic caps have dried out (Do the English still call these condensers?) Tracing through the old discreet amplifiers would become a major project, Which I already have too many of.

I stumbled across Tg and it looks promising if I could get it to run on one of my pi-zeros. Found some random thread where others have used the old pickup units such as the one included with this post. There is a possibility that the problem is in this unit. This also is a major project as I would need to figure out a pre-amp to get the audio into the pi. On the other hand with the pipe organ electronics I have all sorts of left over junk in which to do this. I am also versed in digital audio, SPDIF and toslink. (Once tried to get a job at Dolby.) Surprisingly the postscript language does a good job analyzing audio waves (which are the results of converting piano rolls to MIDI to something that can be listened to. Which was part of the talk I did in Bern in June.

I think I blew my eBay budget over the weekend. Otherwise I might look into one of those Weishi units. On the other hand they probably run something similar to Tg. Tg is open source, so it is also possible (and a major project) to port it to run on one of the STM32F4 processors.

-j


 
Posts
21,184
Likes
48,311
The apps aren't up to much to be honest, I wouldn't even bother. The ~$100-150 ebay timegraphers work just fine on older stuff, I have one, but wont give fully accurate readings on Co-ax modern stuff, particularly for amplitude. For that you will need to spend a lot more.

There is apparently an upgraded Weishi (for about $100 more) that claims to work with co-axial movements. Also has an enhanced display. Wasn't necessary for me.
 
Posts
148
Likes
80
You should be able to compile Tg for a variety of OS.
This is a screen display I did earlier. I couldn't get the in-line mic to connect properly for some reason, so just hung the watch over the PC microphone. Not the best way !
There is a calibration adjustment of -1.7 sec/day for this pc clock. Watch is a lowly 2824 movement, but running well on the wrist at about +1/day in normal use.

(This is a photo of the screen - didn't want to do a full screen copy)
 
Posts
216
Likes
169
My vintage L&R timegrapher is Kaput. It is the transistorized model. At one point I had the vacuum tube (valve) schematics, which I think went the way of the extra box of clock books. I did find a copy of those in German. There are quite a few threads and forum posts on these old tanks. My guess is that the electrolytic caps have dried out (Do the English still call these condensers?) Tracing through the old discreet amplifiers would become a major project, Which I already have too many of.

I stumbled across Tg and it looks promising if I could get it to run on one of my pi-zeros. Found some random thread where others have used the old pickup units such as the one included with this post. There is a possibility that the problem is in this unit. This also is a major project as I would need to figure out a pre-amp to get the audio into the pi. On the other hand with the pipe organ electronics I have all sorts of left over junk in which to do this. I am also versed in digital audio, SPDIF and toslink. (Once tried to get a job at Dolby.) Surprisingly the postscript language does a good job analyzing audio waves (which are the results of converting piano rolls to MIDI to something that can be listened to. Which was part of the talk I did in Bern in June.

I think I blew my eBay budget over the weekend. Otherwise I might look into one of those Weishi units. On the other hand they probably run something similar to Tg. Tg is open source, so it is also possible (and a major project) to port it to run on one of the STM32F4 processors.

-j



Thank you @sheepdoll, comprehensive and fascinating!
Unfortunately my ancient PC is of the steam-powered variety and although I have tinkered with a raspberry Pi in the past I think that level is beyond me now.
On a positive note your input has spurred me to do some more research on the Tg angle, with a view to using my tablet, and a look at a few more youtube clips to see how different people approach calibration.
Thanks again.
 
Posts
5,340
Likes
9,072
So...., summary: Don't spend money on watch makers tools. Rely on Apps. YouTube gets it done. And when your whatever internet based gadget is down = no fixing watches. And if the Internet is out, which it will in the next phase of international War fare , no more watch fixing. Correct ?
 
Posts
521
Likes
3,178
Here are my thoughts on the subject . . .

A timegrapher is an easy-to-learn-to-use invaluable tool when applied to assess the condition of a mechanical watch.

I use the results gathered from six positions to establish a benchmark when I first buy a watch. When I send a watch out for a full service, I expect to see a dramatic improvement upon return and only once did I discover a self-described "watchmaker" who adjusted the watch so that it appeared to run properly, but otherwise, never serviced or disassembled the watch.

To a person without a timegrapher, when one has a watch running 45 seconds fast or slow, one sends the watch out for service and gets the watch back running with an under 10 S/D gain or loss, one may be tempted to say something like: "My watchmaker is fantastic. My watch was running poorly and now it is right as rain! He only had the watch for about a month, so he offers a quick turnaround as well." Meanwhile, an unscrupulous person may have only blown a little air into the watch and adjusted it.

How do I know this happened to me just the one time? Apart from the fact that timegrapher results don't lie, after giving the so-called watchmaker a chance to service the watch two more times over the space of a year, I sent the watch out to a real watchmaker and it came back running nearly as new. I asked the real watchmaker what he had found and he told me the watch was dirty and needed to receive a good cleaning and a full service. Painful lesson learned.

If you own even one watch worth over $500 and you are about to have it serviced for a further $500 by a watchmaker lacking an established reputation in a place such as this, then why not spend $150 more to make sure you are receiving value for money spent?

For a collector of mechanical watches, I see a timegrapher as a practical necessity.
 
Posts
216
Likes
169
I've tried pc and android based devices without a lot of success, but one of them - the Windows based Tg (free) is actually very good, and tells you more than a basic timegrapher.
As well as timing etc.it shows you the escapement locking/unlocking in a graphical form in real time. You probably need to be a professional to interpret the data properly, but it's fairly obvious if something is seriously out of whack. Running it concurrently with a timegrapher gives very close, but not identical results. I think the timegraphers have an element of smoothing in them when it comes to displaying lines, mainly because the granularity on a PC screen is much finer.

Two caveats, though. It relies on your PC clock being accurate, and anecdotally they aren't. It does have a calibration option, but that relies on you having a very accurate ticking quartz watch to do it, and you need to let calibration run for a while to get a reliable result due to the way most quartz watches work.

The second caveat is getting a good microphone, especially one that makes it possible to orientate the watch in all positions. I've managed to get a mic that's sensitive enough, but the positional aspect is tricky.

With a bespoke timegrapher you (should) get an accurate timebase, and you definitely get the proper microphone.

I've seen highly regarded watchmakers using the Chinese ones, so that's good enough for me (although they obviously wouldn't make it onto Omega's list of workshop requirements !)

If you do get one, don't get too obsessive. Even if machines show good accuracy and precision, your wearing habits will have an influence anyway.
That sounds like pretty good advice, thank you.
I think you're saying that I could faff about with a virtual / app based solution but, although there would be a bigger initial outlay, it would be easier and probably produce comparable results to go for a cheap chinese (Weishi?) timegrapher.
I'm thinking that, based on research and advice given on the thread, that's definitely the way for me to go.
Edited:
 
Posts
27,702
Likes
70,404
It's certainly useful if used correctly, and you understand it's limitations.

It can tell you when a watch needs servicing, but it can't tell you that the watch is okay inside. That requires a visual inspection, under magnification, and knowing what to look for.
 
Posts
216
Likes
169
It's certainly useful if used correctly, and you understand it's limitations.

It can tell you when a watch needs servicing, but it can't tell you that the watch is okay inside. That requires a visual inspection, under magnification, and knowing what to look for.

I realise that I'm way out of my depth here but if, using the timegrapher, I find that the amplitude is OK (280-320) and beat error <1ms but accuracy is out by say 90 s/d, if fine tuning then achieves < +/- 10s/d could there still be signiicant underlying problems requiring a service?
I'm assuming here an inpection, not on the level you mention but a novice's uninformed inspection, probably on a cleanliness and 'glaringly obvious' level.
Please excuse the simplicity of the question......that's the stage I'm at.....never too old to learn!
😀
 
Posts
521
Likes
3,178
Archer - Al - feel free to contradict, clarify or amplify.

I also look for symmetry of results with the timegrapher with respect to inverse positions.

For example. Dial up and dial down should be about the same amplitude and no more than a couple of tenths apart for beat error.
Similarly, I expect crown up and crown down to be similar as well as crown left and crown right. A lack of symmetry in the result would indicate a bearing surface issue.

When the amplitude is as expected.
When the beat error range is average below 1 msec and the differential between high and low msec is +/- 0.5 msec.
When the watch is keeping time consistently per each position, even when in need of adjustment.
Added: When the plot on the timegrapher is clean and stable.

I pronounce the watch as lower in priority for full service than other watches.

Here are examples of my testing before and after service for one watch:

Below shows my testing of an Enicar Sherpa Graph "Mark Id" (Valjoux 72 - fixed stud carrier) shortly after receipt:



The watch appeared on the surface to be running OK in that, it was only losing less a second a day on average.

Amplitude is on the low side.
Beat error is well above 1.0 msecs
The difference between the lowest S/D (-12.8) and the highest S/D (17.3) is 30.1.

Same watch, still not serviced, main spring half wound:



Amplitude is on the low side.
Beat error is well above 1.0 msecs
The difference between the lowest S/D (-25.3) and the highest S/D (14.3) is 39.6 .

Yet, to a person without a timegrapher, the watch was working pretty well and with all the chrono functions OK.

Below is the same watch after return from my watchmaker:



The plots on the timegrapher are smoother or more stable.

Amplitude now peaks at over 320.
Beat error is no higher than 1.0 msecs.
The difference between slowest and fastest position is now 12 s/D.

The watch was running a bit fast, and I waited for it to run in before adjusting myself.

Here is is again 24 hours after a full wind and a year and a half after service by my watchmaker and later readjustment by me:



Cheers,

Joe
Edited:
 
Posts
27,702
Likes
70,404
Dug Dug
I realise that I'm way out of my depth here but if, using the timegrapher, I find that the amplitude is OK (280-320) and beat error <1ms but accuracy is out by say 90 s/d, if fine tuning then achieves < +/- 10s/d could there still be signiicant underlying problems requiring a service?
I'm assuming here an inpection, not on the level you mention but a novice's uninformed inspection, probably on a cleanliness and 'glaringly obvious' level.
Please excuse the simplicity of the question......that's the stage I'm at.....never too old to learn!
😀

Certainly if you see low amplitude, varying numbers, large beats error - the machine can tell you, or confirm there is a fault.

However, the lack of those things doesn't really mean the movement is okay. I get a fair number of movements in where the amplitudes look fine, yet the jewels lack oil - it has dried up.

This is sort of an artifact of modern oils. In the past, the natural oils would get very gummy when they failed, and in some cases would stop the watch entirely:



Modern synthetics don't fail in the same manner, so they basically dry up and leave little behind (in particular the lighter oils):



Dial up amplitude on the watch the above photo is from was in the mid 280's, just as an example, but the oil in the escape wheel jewel is completely gone.

Hope this helps you with the point I'm making.

Cheers, Al
 
Posts
27,702
Likes
70,404
For example. Dial up and dial down should be about the same amplitude and no more than a couple of tenths apart for beat error.
Similarly, I expect crown up and crown down to be similar as well as crown left and crown right. A lack of symmetry in the result would indicate a bearing surface issue.

When the amplitude is as expected.
When the beat error range is average below 1 msec and the differential between high and low msec is +/- 0.5 msec.
When the watch is keeping time consistently per each position, even when in need of adjustment.

I pronounce the watch as lower in priority for full service than other watches.

Generally good, but I'll just mention a couple things...

Beat error is an adjustment, and not really used as a sign of a movements health or not.

Same with differences in opposing vertical positions - these would be poise errors, and aren't really related to the condition of the movement.

For me, a visual inspection on top of all the things you mention is still required to make a determination.