Problem with a timegrapher?

[FitOutPost 0]

Hi, my name is Ross. I am a rookie watch enthusiast and I am really puzzled here.

Could someone explain to me what kind of a problem am I facing with my timegrapher?
I do two sets of measurements with the same watch (1 day or 6 days apart) and receive vastly different results - to the point of being completely different from what I observe in real life. 
For example, my timegrapher shows that my watch is running fast (or ahead of time), while in real life I observe that it runs 7 seconds per day behind. I even recorded a video about it so you could see it for yourself: https://youtu.be/mhGzf6aLMlY

How should I interpret that? Am I doing anything wrong?

[m1ks 196]

There are numerous reasons for the differences.
First of all, ensure there's little to no ambient noise in your timing environment.
The watch itself-
How new/old is it?
Has it recently/ever been serviced?
What was the state of wind on the mainspring for each test?
What is your typical daily use and most common position of the watch throughout the day, (Most common are pendant down, pendant right and dial up)
If it's a non manual wind movement like the 7S26 etc then unless you wind directly on the barrel it's pot luck knowing what the state of wind is and the readings on some movements due to cost/manufacture/lubrication/dirt/general quality can differ dramatically in varying states of wind and even be affected by variances in temperature, (try regulating an old pin pallet 0 jewel escapement, they're fun).
Don't be drawn into thinking the machine gives you the ability to regulate any watch into a definite precise loss or gain in seconds, it's simply a very accurate device to help diagnose and set timing much more quickly and accurately than the old fashioned methods of timing against another watch or clock over several hours or more.
You always need to think about your typical wear and activity and always try to average results across your typical wear positions.
You'll notice if you ever time a very high quality movement there'll be very little positional error and temperature related error because they are made to much tighter tolerances hence the reason most movements that we wear or work on will be guaranteed to be within -/+ 26s a day or thereabouts whereas a COSC certified will be within -2+6 (if memory serves there).
For consistency in checking after a service I typically wind fully then unwind one full turn of the barrel using a screwdriver or the stem whilst releasing the click, this ensures good camparison readings across a set amount of time.

Edited October 26, 2019 by m1ks

[FitOutPost 0]

Thank you so much for your prompt and such a detailed reply. I really appreciate that.

A couple of answers to your questions - both watches are relatively new - Invicta is a 5 months old, while Seiko is virtually brand new (just a couple of weeks old). What puzzles me the most is the fact that I observe in real life that both of my watches run -5 to -7 seconds per day, while the timegrapher shows it to be plus almost a minute per day. Does that mean a timegrapher was a complete waste of money?

[FitOutPost 0]

As for "the state of wind on the mainspring for each test" if you watch my video you'll notice I performed each of my tests with a caseback on a watch and therefore I assume the wind should have no influence whatsoever. It is also very quiet in my room during the test. 

[m1ks 196]

Not at all no.
As per my reply, what the TG tells you is what it reads there at the time, the watch however is static and it can't estimate a reading based on use.
They are incredibly accurate but a mechanical watch being what it is, is prone to forces applied to it all day long in use, an auto wind, especially one without a manual wind function, is subject to your daily activity according to how well wound it is.
You say almost a minute a day but the figures I saw on the video show an average of a few seconds gain?
i.e. If you were to wear your seiko from a full wind whilst standing stock still with your hands hanging down for 24 hours it 'should' reflect the initial reading of +50s (this is discounting temperature differentials power fluctuation from the mainspring unwinding and movement).
Look at your daily activities, which positions do you wear your watch in, is it subject to shock that could occasionally halt the balance causing a loss over a day.
Is it subject to wide temperature variations, for example, you live somewhere hot and work somewhere cold or vice versa.
I pretty much guarantee that both could see improvement with a clean and re lube, (i've serviced brand new seiko 5 movements and found huge variances in the oiling of said movements) Invicta I'm not sure about but believe they use decent quality Chinese movements? so that will be a similar story.
One more important thing to bear in mind, there's always a 'settling in' period for a new or freshly cleaned and lubricated movement, I tend to work on a principal of a week of wear then check and regulate again by which time it should have settled into a more regular rythmn.
I suggest that you try getting a reading in the most common positions you wear yours throughout the day, (take into consideration the resting position if you remove your watch at night or when you get in from work etc), take an average and then regulate that average to 7s+ and it should get you close to the ideal 0.
If you want more guaranteed accuracy, there are three options.
1 Strip, inspect, adjust, clean, oil and rebuild but there's only so far that will take you depending on the finish of the original movement.
2 Get a higher quality mechanical movement, (which are costlier of course, it doesn't have to be a 'name', something with an ETA movement for example will give better results)
3 Get a quartz watch, which are and always will be more accurate, (but that's not why we love watches) 

Edit added, the state of wind is in reference to how much wind on the mainspring, if you can manually wind you can wind fully before each test, if your seiko movement is like the 7*** series you can't wind via the stem, only manually by the barrel arbor screw so the mainspring on one test can be fully wound, on the next, half unwound, on the next almost full due to an active day, on the next half again, on the next almost unwound due to a lazy day for example.
 

Edited October 26, 2019 by m1ks

[TheFixer 61]

29 minutes ago, m1ks said:

Not at all no.
As per my reply, what the TG tells you is what it reads there at the time, the watch however is static and it can't estimate a reading based on use.
They are incredibly accurate but a mechanical watch being what it is, is prone to forces applied to it all day long in use, an auto wind, especially one without a manual wind function, is subject to your daily activity according to how well wound it is.
You say almost a minute a day but the figures I saw on the video show an average of a few seconds gain?
i.e. If you were to wear your seiko from a full wind whilst standing stock still with your hands hanging down for 24 hours it 'should' reflect the initial reading of +50s (this is discounting temperature differentials power fluctuation from the mainspring unwinding and movement).
Look at your daily activities, which positions do you wear your watch in, is it subject to shock that could occasionally halt the balance causing a loss over a day.
Is it subject to wide temperature variations, for example, you live somewhere hot and work somewhere cold or vice versa.
I pretty much guarantee that both could see improvement with a clean and re lube, (i've serviced brand new seiko 5 movements and found huge variances in the oiling of said movements) Invicta I'm not sure about but believe they use decent quality Chinese movements? so that will be a similar story.
One more important thing to bear in mind, there's always a 'settling in' period for a new or freshly cleaned and lubricated movement, I tend to work on a principal of a week of wear then check and regulate again by which time it should have settled into a more regular rythmn.
I suggest that you try getting a reading in the most common positions you wear yours throughout the day, (take into consideration the resting position if you remove your watch at night or when you get in from work etc), take an average and then regulate that average to 7s+ and it should get you close to the ideal 0.
If you want more guaranteed accuracy, there are three options.
1 Strip, inspect, adjust, clean, oil and rebuild but there's only so far that will take you depending on the finish of the original movement.
2 Get a higher quality mechanical movement, (which are costlier of course, it doesn't have to be a 'name', something with an ETA movement for example will give better results)
3 Get a quartz watch, which are and always will be more accurate, (but that's not why we love watches) 

Edit added, the state of wind is in reference to how much wind on the mainspring, if you can manually wind you can wind fully before each test, if your seiko movement is like the 7*** series you can't wind via the stem, only manually by the barrel arbor screw so the mainspring on one test can be fully wound, on the next, half unwound, on the next almost full due to an active day, on the next half again, on the next almost unwound due to a lazy day for example.
 

An excellent reply.

[FitOutPost 0]

Thanks again. Please take a look at my last measurements of Seiko SRPA53K1. They show huge numbers in rate gain - from +36 to +58 seconds per day, while in real life the watch loses 7 seconds per day. It is an automatic watch (movement 4R36-05MO, caliber 880095) with a manual wind and hacking features, but I newer used a manual wind as I was wearing it 24 hours per day. And since I work at my desk most of the time, the watch is mostly rested face up, but have a look at a face up position in measurements - it shows +36 spd, which is far from -7 spd in real life. How should I interpret that?

[jdm 2518]

There are a couple things you can do. 

First, regulate the watch on the timegrapher so that it keeps a good time in normal use for one entire week, a good time to do that is at the end of day. I understand that you just need 7 secs faster than current, that is a minuscule adjustment on a Seiko. Once you have it running good, the machine has been put to a good use. 

Then bring your watch to a local watchmaker and ask for a quick timegrapher check, most would that for free. Its rate reading will let you know how much difference, if any, there is to your 1000.

Edited October 26, 2019 by jdm

[FitOutPost 0]

Thank you, jdm, but you have misunderstood the current state of things. My watch actually loses 7 seconds per day very consistently, but when I test it on a timegrapher, it shows a gain of +36 to +58 seconds per day. That makes no sense to me whatsoever.

[jdm 2518]

12 minutes ago, FitOutPost said:

Thank you, jdm, but you have misunderstood the current state of things. My watch actually loses 7 seconds per day very consistently, but when I test it on a timegrapher, it shows a gain of +36 to +58 seconds per day.That makes no sense to me whatsoever.

I understand that perfectly, now read again the part that i placed in bold in my answer above, once you do that you will know more about what's going on.

Nobody here can tell you why you machine doesn't read same as you daily use, but we can tell you how you can know if the machine is accurate.

Edited October 26, 2019 by jdm

[jdm 2518]

--dup sorry 

Edited October 26, 2019 by jdm

[FitOutPost 0]

Oops ... SORRY! My bad! I hear you now, sir, but I still don't understand why my timegrapher is so inaccurate. In any case I will follow your advice, sir, and do what you said - "bring your watch to a local watchmaker and ask for a quick timegrapher check" to check how much their reading will be different from mine. Thank you for being patient with me.

[jdm 2518]

3 minutes ago, FitOutPost said:

Oops ... SORRY! My bad! I hear you now, sir, but I still don't understand why my timegrapher is so inaccurate.

Second part of my second answer:

Nobody here can tell you why your machine doesn't read same as you daily use, but we can tell you how you can know if the machine is accurate.

[FitOutPost 0]

I saw that. Thank you.

[AndyHull 1537]

Does the 1000 have a calibration routine? If so, you may need to calibrate it.

I use a piece of open source software called tg-timer, and it allows calibration against a quartz watch. Having said that, even without calibration it is accurate to +/- 1/2 sec per day on my laptop, not the +60sec per day or so that you are seeing.

All of the above advice from other forums members is sound, and pertinent to the problem.

External noise, for example, can throw things out with tg-timer dramatically.
Setting the lift angle incorrectly can give the wrong impression of the watches performance.

Wearing a watch involves much more agitation, swings in temperature and humidity, and normal daily wear generally favours one position rather than another throughout the day, so the time graphing machine or software can only give an estimate of the watches likely performance, rather than an absolute guarantee of performance. 

If the time grapher is giving answers that are at least in the correct ballpark, then you can probably refine the results based on experience. 

You might also like to watch some of Mark's videos regarding watch regulation.

For example here.

 

 

Edited October 26, 2019 by AndyHull

[jdm 2518]

24 minutes ago, AndyHull said:

Does the 1000 have a calibration routine? If so, you may need to calibrate it.

Not for the user.. The manual state that the machine is calibrated for life at the factory, and has a temperature compensated quartz. Both the calibration input (via USB) and trimmer pots are undocumented, which doesn't mean that are difficult do figure out.

However as mentioned above it's easy to have an idea of its accuracy using any watch as reference to another machine.

Edited October 26, 2019 by jdm

[JohnR725 1709]

The problem with timing machines are there not magical crystal balls that make predictions of what your watch is going to do 24 hours a day. There only a device that shows you what the watch is doing at the exact moment it's on the machine in that exact position. Or simplistically timing machines are instantaneous the watch is averaging over time.

Out of curiosity did you just turn your machine on and put the watch on it or have you changed any of the settings? For instance you can change the averaging time which will help 20 seconds works fine.

Then I notice it looks like a diving style case? The problem with Seiko cases like this is their heavy and they use plastic movement rings to hold the movement in. It does not make good for transmitting sounds out for the timing machine and you can get bad timing because of that. Which means sometimes the best way to time a Seiko movement is out of the case.

So if you want to see more realistic numbers from the timing machine you need to time in more than one position.The first link explains multiposition timekeeping and how to do the calculations start with just the rate. Set your machine to 20 second averaging and I would time at least 30 seconds per position. Allow 30 seconds between changing positions for the watch to settle down.

Then the other two links have technical specifications. Your 4R36 is equivalent to the OEM time module NH36. So the last link is the tech sheet time module usually does a better job of explaining the timing specifications than the equivalent Seiko tech sheet.

https://www.witschi.com/assets/files/sheets/X-D-DVH-Di-Im-N_EN.pdf'

http://calibercorner.com/seiko-caliber-4r36/

https://www.timemodule.com/upload/PDF/NH36_TG.pdf

 

 

[FitOutPost 0]

Thank you so much, John. I really appreciate your advice(s). 

To answer some of your questions:

I did change a few settings - the lift angle (obviously) and also the averaging time (20 sec).

I also timed my watch in 4 different positions (as could be seen from the table above) - Face Up and Down, Crown Up and Down.

I will also check those links you so generously provided.

Great help, sir! Thank you again.

[jdm 2518]

14 hours ago, JohnR725 said:

Then I notice it looks like a diving style case? The problem with Seiko cases like this is their heavy and they use plastic movement rings to hold the movement in. It does not make good for transmitting sounds out for the timing machine and you can get bad timing because of that. Which means sometimes the best way to time a Seiko movement is out of the case.

I routinely test Seiko watches on Weishi machines, going from the light SNK807 to chunky mono-case MarineMaster, either with a plastic or metal mov't ring, and never had any problem getting a correct base reading with the case closed.
That us I have never experienced any difference in timing with the mov't in or out of the case,  but in the first situation displayed pattern may be a more or less affected, with dots being drawn not where they should.
The machine has a gain control, but I have never found it much useful.

I think that no matter how one twists it, explain and goes around what the OP is saying is simple and sound: significant and constant difference between timegrapher reading and on-wrist performance.