Hi Guys
I started this other thread about my Seamaster 120 project:
http://omegaforums.net/threads/seamaster-120-project.15504/

The subject of timing machines came up and interpretation of the results to check your movement work. I had one of these cheap Timegrapher 1000 machines on order and it arrived yesterday. It's easy to use but, presents a lot of data if the watch has possible faults so, probably best not to buy one unless you are interested in the movements. This is new to me so, would welcome any comments, advice or observations. I am putting down what I think.

I wanted to start with a watch I am happy with and I have seen on another timing machine although that was five months ago after I worked on it so, here it is with my machine. It's a 14393 bought from one of our members and was declared as 'needing a service'. It was not happy being manually wound but fixing that was easy. Anyway, it got a clean, oil, new mainspring etc. I like this style and it is:
Cal 561, chronometer spec from 1960 (have just realised this is a birth year watch )

I have the Omega specs for a 561 although am not sure if the chronometer spec would have closer requirements (I don't have the 1960 chronometer rating). The basic specs are:
19800 A/hr, Lift angle=49º
3 testing positions: Dial up, Crown down, Dial down.
Stabilize 30 secs before measurement and use 30 sec period on the machine (for averaging).

Here is the machine set so, it is pretty obvious I took this after I completed the tests as you can see my notes. The watch time is not correct as I just set it away from the date change. Now to see if my work has lasted 5 months and 1000 hours of use (not a long time by any standard)....


0H first test:
Let mainspring down, full wind and leave to rest for 30 minutes. Results:
Dial up: +4 secs/day, Amplitude =278º, Beat error=0.0 ms
Crown down: +3s, 255º, 0.1ms
Dial down: +8s, 269º,0.2 ms
My notes just say "smooth line".
I completely forgot to take photos of this but the pictures are basically the same as the next test.

24h second test:
Left watch for 24h at rest and repeated the test:
Dial up: +2s, 244º, 0.0 ms
Crown down: -5s, 222º, 0.0 ms
Dial down: +1s, 237º, 0.3ms
And the pictures although I now realise I should stop the test before taking the pictures as the screen shows slightly different to my notes.






The results can be compared to the specs and this is how I read them:
0h to 24h average should be -1 to +16 secs:
Results show Dial up = 3 secs, Crown down=-1 secs, Dial down = +4.5 secs.

Delta 0h should be within 25 secs and is 5 secs.

Delta 24h should be within 35 secs and is 7 secs.

Beat error should be <0.8 ms and is a maximum of 0.3 ms.

Amplitude should be >160º and exceeds that.

I seem to have set it a little slow as the target rate is +8 secs and mine is generally a little lower. The amplitude drops very regularly 33 to 34º between the 0h and 24h test and I would expect something like this. I'm not sure that the variation in amplitude between the three positions is what would be expected but am here to learn. Any help is much appreciated.

Thanks in advance for any comments.

Chris

 

If the watch is a chronometer, then it needs to be checked over 5 positions, not just 3. And yes the numbers for Delta would certainly be tighter - max of 12 seconds at full wind over 5 positions, and 15 at full wind -24 hours over 5 positions. Crown right (12H) is the only position not checked for COSC. Every COSC rated movement Omega makes has these same tolerances, so right from the older 1120 and 7750 based watches, right to the newest of the co-axial calibers.

The numbers on that watch are quite good. These often have low amplitudes, and yours are very good for this movement in my experience. Checks over 5 positions would tell me more, but with the information presented I would be happy with that. Don't get hung up on the target rate - I don't really pay attention to that to be honest as it's not a mandatory item.

Note - I check all movements over 6 positions, regardless of what the rating of the watch is, because I want as much information as possible.

Cheers, Al

 

Thanks Al. Your response is much appreciated and I will take your points on board.

I will run the full test this weekend and put it in a table for legibility and post it here for completeness. If you happen to see it, and have any comments, they would also be appreciated. This seems like a nice movement but the 120 Seamaster that started this does not look so good. I will run a full test on that and see how it performs but it seems to have far lower amplitude (you've put my mind at rest there) and a significant beat error. Let's see if I can improve it...

Cheers, Chris.

 

Beat error is easy to correct - these movements have an adjustable stud carrier so it's just a matter of moving it until the error is within spec. Actually the closer to zero the better, but you will see drift in different positions - side shake of the staff can cause errors, and if they get too large the staff should be changed - not sure if you are up to something like that yet.

Adjusting one that does not have a stud carrier that can move (like on a Cal. 321) is a bit more tricky for the uninitiated.

One thing I didn't catch before - for non COSC watches that only require checking in 3 positions, they are not the 3 that you are checking. Those 3 should be dial up (CH), crown down (9H), and crown left (6H). And of course the crown left and crown right positions are with the dial facing you, so if you have the watch in the holder with the train wheel side up as you start (dial down) you have to flip the positions around in your head. Sounds more complex than it is - it becomes second nature after a while.

Cheers, Al

 

Your results are in line with the ones that could be expected from a reasonably preserved caliber 561.
The graph appears only slightly irregular (this may be due to a lot of reasons), but perfectly acceptable.

For your convenience, here are the results from a very well preserved 561 specimen that I personally overhauled a few months ago.
You will notice that a good balance shaft tends to yield a very low drop in amplitude in the horizontal positions over 24 hours. On the contrary, the vertical ones are normally subject to a drop in amplitude of about 25°.

Of course the results would be very different for a non-chronometer movement from tha same period (say a caliber 321).

 

Hi Al

Yes, I had previously moved the stud carrier, back in April, to bring the beat error down from 0.7-0.8 and you're correct that this is easy. It really was just a 'nudge' as the movement was so small. I also know that some will need the balance collet moving (I have a 354 which has a fixed stud carrier, as I recollect) but have not advanced that far yet. No, I am not up to a balance shaft change at the moment, a little too advanced, but I have some practice movements to try things on (they are still running so have not broken anything yet!).

I made a mistake with the positions and thanks for pointing that out. I now have a template sheet and included the pictures of all to avoid that error. Have just run all 6 positions on the cal 561 at 0h so should post a picture tomorrow evening with all values in a far easier to read format.

Thanks again, Chris.

 

Thanks (I need a name here!)

Part of the irregularity is the cheap LCD screen which has a very limited number of pixels so, unless the line is horizontal, it always has steps. There are some other irregularities though. You can't beat a CRT for real definition. Unfortunately, the machine has no output to enable me to hook up to my scope but for the price, I can't complain.

So, I see what you mean about the amplitude (thanks for the pictures). When I have the full test complete, I will get a better feel. I definitely have an amplitude drop for the horizontal one already made. I can't claim much for this movement as I only cleaned and oiled it after fixing a small click spring problem.

Cheers, Chris.

 

Hi Chris, my name is Maurice.

I do look forward to see the complete test results. Did you take a macro of the hairspring?
Cheers.

 

Hi Guys

Here are the full results of the test on the Constellation.


Unfortunately, I do not have a macro of the hairspring. I am using TM 9-1575 to try and interpret the results. Is there a better source? Anyway, ignoring the 12H non COSC position, I am just in the chronometer spec but have the rate set about 8 secs slow. As Al has implied this may not be the be all and end all as long as the spec is fine and is an easy tweak. Actually, in use, it runs very close to zero as it is mainly in the 9H position. I am going to review 9-1575 to see what these results mean but, am inclined to accept this one is fine at this stage in my knowledge and give you some more interesting results from my 120.....

Anyway, comments are welcome and I will try and explain what I see. Thanks.

Cheers, Chris.

 

I'm persevering with this thread although it is not too interesting for most people. Anyway, perhaps one or two are seeing something from it...

The first movement I built was for a Watchco style SM300 so, a cal 565. I have a few of these and didn't chose the one that had the best aesthetics but, it came up pretty well. It's been used for the last six months and so, it was worth a test. I did tweak the beat error yesterday as Al advised and ran the full test. It's non chronometer so only three positions are really applicable and I ignored the others to see if it is in spec. To me, this doesn't look too bad and after running all of this, I tweaked the rate a little as noted.



Here is a full wind plus a few hours picture after rate adjustment. Not easy to see but, it is 6H. All the curves look pretty much the same.



Pleased with this one but, comments welcome.

Out of interest, in my 120 project thread, Al said "We have gone through interpreting timing machine results a couple of times on this forum, and indeed when you start to see detailed numbers, it puts a new perspective on your work, and not always one you want to be aware of." Well, I am getting a lot of scatter with the 120 and another watch. Like an idiot, I changed my pallet jewel oiling and think it has to be that so, will investigate.

Cheers, Chris

 

Now for one with problems so, any help would be appreciated. This is my Seamaster 120 cal 565 whuch I thought was finished but it is not happy. Have decided to strip it all down again but am wondefing in which areas I should be concentrating. I read this as a escape locking problem but... thoughts?

First FH - looks fine.


Next CH - not so good.


Then 9H - two pics to cover sixty seconds ( don't know what I did but the second pic is the last one in this post showing +1, 220, 0.4)



Then 6H - still bad



And finally, 3H

 

A few more observations. I have the Seamaster 120 running fine now but, am still not happy with the timegrapher. There are still some spurious points in some positions even though I have tried a completely different train (spares watch), the jewels look fine and I re-cleaned and re-oiled it. All parameters are OK and in use, over this weekend, it has been running within 2 seconds a day. Ah well, will return to it later.

I am still learning how to read these plots and wanted to share something I believe I have worked out. I tested another Connie I have worked on. It's a 561 (chronometer) and is running within spec (perhaps a bit fast) but has this feature on the timegraph:


Sorry about the poor pic and strange colours. The points indicated occur regularly in all positions about every five seconds. The escape wheel rotates 11 times a minute (it has 15 teeth and the watch is 2.75 Hz =19800 bph) which is 5.45 (recurring) seconds for one rotation. From which, paraphrasing Al, I have learnt something I may not have wanted to know....

Either one escape wheel tooth is dirty/badly oiled.
Or
One tooth is damaged and I did not see it.

Will open it up and hope for the former as it should improve my work but, I suspect it is the latter. At $60+ for a new escape wheel, I'd be interested to know if this sort if variation is within reasonable spec for a fifty year old watch.

One of the problems with these cheap machines is they scale the plot relative to the beat error. So, these odd points look worse because the beat error is so small. They are probably only 0.4ms instead of 0.2 so would be within the 0.8 spec.

Cheers, Chris.

 

love the post Chris! so interesting.

 

Hi Chris,
as you say, with these cheap machines it is difficult to state beyond any doubt that one escape wheel tooth is damaged. It might be the case, but the graph is certainly acceptable for a fifty year old watch.
However, the amplitude seems a bit low. Which oil has been used for lubricating the pallet fork?

Cheers, Maurice.

 

I am pleased you are persevering with this thread because there is a lot of things which can be learnt from the behaviour of old calibers.
Most people are not interested in these problems, nevertheless it is a fascinating topic.

Cheers, Maurice.

 

I think that we have to bear in mind that there were many good watchmakers who could and did service movements to a high standard, long before there were timegraphers or ultrasonic cleaning machines. Somehow they managed . . .

 

Chris - you just know where this is heading and it is not going to be cheap:

http://www.witschi.com/e/produkte/?sub=1&cat=&id=231

 

Hi Maurice

Yes, these machines are not the best but those clicks are regular at about 5.5 seconds (the screen shows about 30 seconds) so, I will pull the train and clean the escape wheel again just to be sure. Good news that the line is considered acceptable anyway

I am using Moebius 8000 for the fork, pallet stones and balance shaft. Elsewhere, typically D5 except the train on this was done with 8141 which, I understand is better for a chronometer. Would take advice here, thanks!

This movement does not exceed 240º even on full wind. It was +30 hours or so when I took the picture. I usually use Omega mainsprings but this one has a quality generic at 1/3 of the price. Could that be another cause?

By the way, 8000 is not far off the price of gold by weight! Unbelievable...

Cheers, Chris.

 

Hmmmmm. Good point. The timegraphers have been around a long time but, I suspect were out of the reach of many. Apparently, some people can hear a beat error of 0.5 ms by ear alone...

 

Ha ha!