I have a Current (2018) Speedmaster and a Seamaster, as well as some other mechanical watches, Seiko, etc. Been considering purchasing a Timegrapher, more out of curiosity than any real use or need. My question is, what's the difference between the various models on Amazon? Seeing various brands, all with the model # 1000, also some similar units with 1900, 1600, etc with prices from $130-$300. Reading the "information" which seems obviously translated from Chinese, is mostly bad syntax/gibberish. My Seamaster is a coaxial, some models mention coaxial, but nothing further.
So, any experience here on what to buy? What has worked for you? Is Amazon the place I should be looking? Also looking at an inexpensive ultrasonic jewelry cleaner for bracelets, looks like a $30 or so unit would work. Appreciate any insight here. Thanks

 

This ultrasonic cleanerPurchases made through these links may earn this site a commission from the eBay Partner Network on eBay has worked for me for several years, as has this timegrapher.Purchases made through these links may earn this site a commission from the eBay Partner Network All the 1000 model timegraphers are pretty much the same, and the word "coaxial" in the title means nothing---it's just keyword spamming.

 

Mine is a basic "1000" version, although there seem to be several sub-versions of that too. It has a monochrome screen, and some people find a colour screen better to separate the lines more easily.
Some have a gain control, others are automatic. Mine is automatic and it does a good job of picking up every watch I have apart from an old Certina which has a slight "ting" noise that confuses it. After a few attempts it will pick it up though.
Amazon or ebay is likely the best source, and since they all seem pretty close in quality I'd buy on price/location.
It's good enough for me - and I think more than a few watchmakers who don't have to gear up to the level some manufacturers demand, also use them.
They aren't state of the art though, and you can't calibrate them if they start to drift.
Good for - general diagnosis of how well or otherwise your watch is running - differentiate acceptable/good performance from the bad.
Not good for - the in depth analysis needed to fix some watch problems, when a much more detailed examination of the movements sounds is needed. But you're not going to venture into those realms anyway.
Mine tells me if I think I should have a watch serviced, and it also tells me generally whether that service I paid for has delivered results acceptable for that particular watch.
If you have the confidence to do so, and are prepared to take the small risk of damage, regulating your own watches is made feasible with one of these. I wouldn't attempt it on a free sprung balance though
They will time co-axials OK, but the amplitude reading will be meaningless, as they can't pick up the sounds needed to calculate it correctly.
Don't get one if you're even slightly inclined to be obsessive, because you just might find out that watches you thought were fine and accurate, aren't quite as good as you think.
A £20 ultrasonic from Lidl works fine for what I need, but I've never used a real professional one, and I don't let bracelets get too grungy anyway.

 

Thank you for this input, reinforced what I had concluded, I'd be surprised if all of the "1000" Timegraphers didn't come from the same factory. Freely admit to being mildly obsessive, you tend to get that way working on & calibrating medical equipment for nearly 25 years, kinda (hopefully) goes with the job. The comments pertaining to verifying if service is needed or was performed properly rings with me, that and seeing what effect different positions have on accuracy. I'll probably order off Amazon, primarily because of the ease of return if there's a problem.
Again, appreciate the depth of keepitsimple's response, this forum has added greatly to understanding why I really prefer Omegas.

 

Thanks for posting. I was looking into this too

 

All 1000 timegraphers do come from the same source. Common names are Weishi and MTG although Ofrei sells their own branded version. I highly recommend the 1900 as it can be had for a little more and has a much better color display. The two traces appear in yellow and light blue to discriminate entrance and exit pings for easier beat error adjustment.


(Sorry I'm showing on my Milgauss in violation of OF terms of service)

It also reportedly has a digital filter and can function properly on an Omega Co-Axial movement. I can tell you that I have put both my 9300 based watches on it and get satisfactory results, though even with the lift angle set to 38 degrees, the amplitude comes out low.

There are a couple help pages, one shows that the crystal is temperature compensated, accurate to one part in a million and calibrated against an atomic source.

Unless you are going to part with $1.5K for a Witschi, I highly recommend this model.

 

I think you'll find that while the timing results are correct, none of these cheaper devices capture the correct co-axial sounds needed to calculate amplitude properly, so what you're seeing isn't true amplitude.
I know mine doesn't.
The Witschi devices have a special mode to do that I think.
If I were buying one again, I think I'd probably go for the 2 colour version you have, and it would be helpful adjusting beat error, but I don't find it that difficult to do using the one I have (provided there's an adjustable stud carrier of course).

 

It will tell you the timing and beat error properly, but not the balance amplitude. This requires special programming that none of these low priced units have.

 

Hi Al, if you have time, could you maybe explain a little bit why the coaxial movement requires special software? I haven't had any luck finding a trace for a coaxial. Looking at drawings (including in Daniels' 'Watchmaking' book, where I keep adding to my question/errata list almost every time I look, but that's another story...) all I can conclude is that there should still be a sequence of unlocking - impulse transfer - locking sounds, and that the two impulse transfers might sound quite different, since one is to the balance and the other via the lever. But the relevant time for the computation of the amplitude should still be the delay between the first and the last (third) sound, so the difference in impulse sounds shouldn't matter, should it? Seeing a trace would definitely help to understand this.

 

A timing machine looks at the specific sequence of sounds to determine the various measurements. If that sequence of sounds changes significantly, it requires the programming to be changed in order for the machine to know which sound is which in that sequence.

 

I've never found anywhere that shows what a co-axial should look like either, but just out of interest, I've recently taken traces of 2 of my coaxials - a 2403 movement (2500 without date) in a Railmaster, and an 8500 movement in a seamaster AT.

The 2403 is about 12-ish years old, was serviced about 18 months ago when they finally (after 2 returns to the company that did it) replaced the co-axial wheel. I t's been rarely worn since. The 8500 is abbout 3 years old from new and very little worn.

A few caveats ! I'm just using some free software on my PC, and a cheap contact microphone. The resulting signal is a bit noisy as it uses only the PC's sound card. It's certainly not professional stuff so I can't give any assurances about how well it is picking things up, especially through the fairly hefty Omega cases. However, on conventional swiss lever watches (see the first one below), it does show the kind of 3-impulse trace you would expect.

First one - just as a reference - an old Certina 28-10 - swiss lever - it's actuallykeeping very good time for an old watch :-



Second - Omega 2403 co-axial



Third - Omega 8500 coaxial


Both the Omega watches have much heavier cases than the Certina, and I think the software probably picks up the Certina more cleanly.

I don't know what the co-axial traces represent, especially as they are different, but then they are different co-axial versions anyway.

If anybody can shed any light on it and I apologise if it confuses rather than informs.

 

^^ I've read in the Witschi training manual that the first little spike is ignored, then the time between the next two spikes is used to calculate amplitude (along with the lift angle setting).

I can see how the two-plane co-ax (2403) would be completely unreadable. I presume that the noise from each impulse rings on enough to mask the next impulse. I've only used my Weishi on 9300s.

What I'm really curious on these traces is the difference between the left and right patterns on the 8500. Anyone venture an opinion as to why?

I can see why the left version would fool a timegrapher not expecting it, but the one on the right looks similar enough that the machine could resolve the two measurement pings. Perhaps the amplitude difference in patterns is fooling the cheap timegrapher.

I've always thought that the published 38 deg. lift angle on a 9300 means something different to the timegrapher than the same on a Swiss lever movement. Perhaps a setting other than 38 deg would result in the correct amplitude.

 

Thank you very much, keepitsimple, for posting the traces! My microphone (part of a headset) is not good enough to disentangle any details of the tic and toc of my Omega 9300.

Expiredwatchdog, I would expect the acoustic signals from the clockwise and counterclockwise half-oscillations ( 'tic' and 'toc', or left and right in your post) of the 8500 movement to be different since the impulse transfer is via different paths: I assume the left/first one with the four events/spikes is the indirect transfer via the lever, so (all guessing on my part, maybe your Witschi manual has specifics?) the first small pulse is the roller hitting the lever immediately followed by unlocking, the second is the tooth of the impulse wheel hitting the impulse pallet of the lever, the third right after is the lever hitting the roller (completing the impulse transfer to the balance), and the fourth is the locking and the lever hitting the bank (not sure about the order there, probably the lever hits the bank before the escape wheel tooth locks on the pallet).

Regarding fooling the timing machine, Omega specifies that for version A of the 2403 you should set your older Witschi machine to 30 degree lift angle, and for version B to 38 (according to http://www.phfactor.net/wtf/Omega/1831_Omega2403,2403A,2403B.doc.pdf), but for newer machines using 38 in both cases is correct.

Naively just taking the distance from the first to the last event (which agree for 'tic' and 'toc' in all three cases) in relation to the half-oscillation length, and using 51 degree lift angle for the Certina, and 38 for the Omegas, I obtain amplitudes of 216 (Certina), 290 (2403) and 207 (8500) degrees. Keepitsimple, would you happen to know the true amplitudes going with these traces?

In any case, seeing the traces is really great!

 

Sorry - the software didn't pick up the amplitude for any of the 3 - it normally does for conventional swiss lever movements, but for some reason didn't that time. On my timing machine, the certina comes out ~240 (19800 bph) but the test conditions may not be identical. As far as the co-axials are concerned, I think the software just ran off into the corner sobbing........

I used one of these clip on pickups:-

https://www.ebay.co.uk/itm/Common-U...-up-w-Cable-/202856819029?hash=item2f3b356955Purchases made through these links may earn this site a commission from the eBay Partner Network.

Best to remove the rubber pad to get better contact with the watchcase (the clip is plastic so won't scratch), but I still need to turn the microphone up to 100% plus boost to get it to pick up the signal. The result, listening through headset, is a faint tick and a lot of noise.Probably due to a poor quality soundcard linked to a low quality mic. (all of that sort of thing is a mystery to me) but it means that on some heavy case watches the S/N ratio isn't good enough to get clean traces. I'm sure someone who knows about this stuff could put together a much better setup.

I have a timing machine anyway, so this was just out of a bit of curiousity. One idea might be to get the timing m/c mic. connected to the pc, but a) it has a non standard plug (as far as I can see) and b) I don't know which pins are which anyway.

If anyone ever figures that out I'd be interested to know

 

Actually, I think on a swiss lever, the amplitude is calculated from the first and third impulses, not the second and third (happy to be corrected).

The graphs I published are the time compressed versions - the software also gives expanded versions, and in the case of the 2403 even that shows only a single second impulse for both tic and tac.

I'm not sure how much reliability it is safe to put on the 8500 traces, although they do look pretty clean, and the expanded version of the trace does show 4 clear impulses for the tic, and 3 for the tac.

My timing machine does come up with amplitude figures for both the co-axials, but I know they are wrong. However, even if correct ones were calculated I don't know what is good/bad for these movements anyway !

As a PS - I think it's a remarkable job the author of the software has done to get it to the stage it is at, especially considering how tricky it must be to identify the pulses contained in a lot of noise.
How reliable it would be to adjust/regulate a watch is a different matter. Even after doing a timing calibration, software running under Windows (in my case) has to share so many resources with everything else going on that it must be impacted in some ways, as opposed to a bespoke piece of hardware dedicated to just the single job.
But then, it's free, not £££££.

Anyone who wants more info about the software, there is a complete thread on https://forums.watchuseek.com/f6/open-source-timing-software-2542874.html

 

Have a look at pages 5 to 19 (ignore 12 and 13) of this document which explain well how these machines read a Swiss lever escapement. These explanations might also give you an idea of why the cheap machines won't read a coaxial completely.

https://www.witschi.com/assets/files/sheets/Witschi Training Course.pdf

Good luck, Chris

 

Thanks for the link.

Edit:-

Thanks again for the link, Chris. I have had a shorter version of that document downloaded in the past, and (even I !) understand how most of that works for a swiss lever movement. It doesn't shed any light on co-axials though, or illustrate which impulses should appear, and which of those are used for amplitude calculations. Perhaps Witschi, having managed to get it to work, don't want to be too public about how to do it

All of this is only of general interest to me, as I wouldn't mess with a co-axial anyway. Personally, I'd be more inclined towards buying new Omegas if they'd stuck to more conventional movements, but that's another issue altogether. As it is, I'll stick with just the 2 I have.

As an aside, if measured correctly, do you know if Omega publishes an acceptable amplitude range for these movements ?

 

No range, only a minimum required amplitude 24 hours after winding. For the co-axial calibers, that is generally 200 degrees, regardless of position.

 

Thanks for the information - so broadly in line with their other calibres.