timegrapher problem?

[ArtNalika]

Hi 

I just buy a new timegrapher from ebay

I tesed with ETA2836 after overhaul but....

Why beat error not match with graph line ?

In picture beat error must be 0.0 or less but why beat error too high to 5.7ms.

Problem with my timegrapher or movemevt?

[JohnR725]

Visually looking at the watch is it in beat?

[khunter]

Try adjusting the lift angle setting. 52° isn't always the right value. You will also see the amplitude go up or down with this setting, as I suspect 240° is a bit low. Try a different movement if available also and compare the readings.

[JohnR725]

The correct lift angle for the 2836 is 53 degrees which isn't going to cause much of a change from the 52 it's currently set at.

Visually what does the watch look like it's doing? Usually when the numbers do not agree with the graphical display on a Chinese machine is because something is beyond the normal range. For instance if the beat is adjusted beyond the 9.9 ms it will rollover giving you weird numbers possibly like you're seeing here. If the amplitude is insanely low timing the machine will read the wrong part of the waveform give you a nice happy amplitude which is totally wrong. 

With watches like this that have a floating stud it's best to visually put it in beat first before putting it on the timing machine. Then you will be really close when you put on the timing machine otherwise if you're not careful it is so easy to go past where you're going and end up with a hopelessly horrible beat error.

For a list of lift angles PDF found that this link

https://www.witschi.com/assets/files/sheets/Scope Schlagzahlen def 2-3.pdf

[clockboy]

Adjusting the lift angle will not correct the error. Looking at the pic the watch is in beat. Send it back as faulty.

[jdm]

50 minutes ago, clockboy said:

Adjusting the lift angle will not correct the error. Looking at the pic the watch is in beat. Send it back as faulty.

No, this issue (lines overlap but beat error shows 5.7ms) has been discussed already, it happens because drawing of lines rolls over edges.

 

[clockboy]

1 hour ago, jdm said:

No, this issue (lines overlap but beat error shows 5.7ms) has been discussed already, it happens because drawing of lines rolls over edges.

 

Thats a new one for me JD.  Have found if it is a mile out it just does not show a measurement.  A 5.7ms should still register but the two sides of the beat will just be a long way apart but should show. If you are measuring a pin lever movement because of unrelated noises you have to turn the sig value to it's minimum setting. Adjusting the lift angle shows a more accurate amplitude reading but IMO is of little use when regulating a watch. If there is a line cross over issue I have not encountered it.

[rodabod]

Try another watch or adjust the beat maybe. I thought mine overlapped at 5 or 10mS, but that's from memory; it may vary from unit to unit.

Edited February 24, 2018 by rodabod

[jdm]

6 hours ago, clockboy said:

Thats a new one for me JD.  Have found if it is a mile out it just does not show a measurement.  A 5.7ms should still register but the two sides of the beat will just be a long way apart but should show.

The 1000 timegrapher has a low resolution display, so it can happen. Easy to experiment and reproduce

[clockboy]

21 hours ago, jdm said:

The 1000 timegrapher has a low resolution display, so it can happen. Easy to experiment and reproduce

My timegrapher is the 1900 model. Perhaps this model does not suffer this error.

[jdm]

2 hours ago, clockboy said:

My timegrapher is the 1900 model. Perhaps this model does not suffer this error.

Which has a much better screen resolution, beside colors and some more features.

[JohnR725]

The Chinese timing machines are interesting as far as the internal programming goes. The earlier machines have limited ranges in just about everything. I think they only did 0 to 4 ms beat error. Then  the newer 1000 and the 1900 appear to have the same internal programming but the 1900 has a bigger graphical display and some minor enhancements. What would be interesting if someone With a 1900 machine and a watch with a floating stud to make this easy slowly push it all the way to that extreme and what do the numbers indicate and what does the graphical display do.

So the real problem is  when things are out of range what does the timing machine tell us? The Chinese machines try really hard to always show something obviously lacking error handling capabilities. The witschi machines usually just will not display anything at all.  Then for that problem you switch the machine to rate only Beat and amplitude are not displayed or calculated  and usually get a display.

[clockboy]

I can tell you now if it goes a mile out then it can not register and gives a blank reading. I think it has shown 9.0ms out but more than that then no readings


Sent from my iPhone using Tapatalk

[JohnR725]

The maximum beat error for the 1000, 1900 and witschi machines is 9.9 ms.

[jdm]

2 hours ago, JohnR725 said:

Then  the newer 1000 and the 1900 appear to have the same internal programming but the 1900 has a bigger graphical display and some minor enhancements.

Well, it's not just bigger, it has colors which are quite useful, not just to take better pictures, but to show which "side" the beat error is, as the yellow and blue lines will invert position. The other features include a 10x accuracy resolution mode, configurable measurement interval, and rate rage.  

Quote

So the real problem is  when things are out of range what does the timing machine tell us? The Chinese machines try really hard to always show something obviously lacking error handling capabilities.

Actually rate range is adjustable as mentioned above. I think the minimum "detected" amplitude is about 100 deg, but when things are really bad that doesn't really matter, as the pattern or just looking at the balance is enough.

[JohnR725]

You forgot there's one other feature that makes the 1900 interesting is the A and B display feature. One of the nice features of a paper tape machine is you can run long or infinite until you run out of paper timing plot liquid crystal displays can only go across the display typically. Some timing machines have a memory feature you can get one or two screens. The 1900 has another mode that lets you display twice as long of a display. Nice if you're looking for problems in the gear train binding issues.

Then the amplitude thing does come up. I'm not even sure If it go down to the 100 before it gives you false numbers. I was visiting a friend with the 1000 machine he was very excited with how well his pocket watch was doing. He was looking at the amplitude on the timing machine I was looking at the balance wheel and visually look very bad. It's why I typically ask people how does the watch look or even in this discussion you need to physically put the watch close to in beat before you put it on the timing machine.

[jdm]

4 hours ago, JohnR725 said:

Then the amplitude thing does come up. I'm not even sure If it go down to the 100 before it gives you false numbers. I was visiting a friend with the 1000 machine he was very excited with how well his pocket watch was doing. He was looking at the amplitude on the timing machine I was looking at the balance wheel and visually look very bad.

You mentioned this often but in my own experience (practically limited to current Seiko only) I've never had a problem with amplitude reading. When it was visually poor -unfortunately happes a lot with these watches - the instrument showed a consistent value, if any at all. But I have seen that sometime the caseback needs to be removed or the gain slightly adjusted for a clean reading.

We all agree that the machine is not a substitute for our eyes, and ears as well as discussed In other threads.

[JohnR725]

I had to go back and look at my previous example of the amplitude problem and I do agree they are extreme cases.specifically timing pocket watches with very poor running conditions.