Thermocompensated Citizen movements (Cal.0330G and Cal.E510G) tested on the Witschi QT6000

[ppaulusz]

While we know the thermocompensation scheme of the ETA movements we do not have as much info about the thermocompensated Japanese movements. Even tests carried out by fellow forum members on these Japanese movements did not reveal much about the actual thermocompensation (when and how the thermocompensation did take place).
Good thing about the Witschi QT6000 that it can check the accuracy of the quartz crystal (by the instrument's acoustic sensor) and the accuracy of the stepping motor (by the instrument's magnetic sensor).

Background info:
- if there would be no inhibition (digital count suppression) then the readings from the quartz oscillator and the readings from the stepper motor would be identical.
- modern quartz movements (including non-thermocompensated ones) do have inhibition.

Now, I happen to have two thermocompensated Citizen Exceed watches. One is fitted with Cal.0330G and the other one is fitted with Cal.E510G. And I also have a Witschi QT6000 so let's see what it can reveal about Citizen's thermocompensation technologies:
I started with the stepping motors. According to the magnetic sensor both movements have an inhibition period of 10 seconds! Now that is good news as far as testing is concerned because 10 seconds is a short period compared to the 4 or 8 or even 16 minutes(!) in case of the thermocompensated ETA movements.
Here are the results of the stepping motors' measurements (over the 10 seconds inhibition period):
- Cal.0330G: +0.01 s/d
- Cal.E510G: -0.00 s/d
(The above results were delivered at 22 degrees of Celsius - in other words at room temperature. Though this tests are not about the accuracy of the movements but rather about the thermocompensation schemes, as a side note I can add that because my Witschi is well calibrated the results of the measurements of the stepping motors are showing the actual accuracy of these movements at the given temperature.)
Next steps are the measurements of the quartz crystals by the acoustic sensor of the Witschi. The Witschi measures the rate of the quartz crystal once in every second. I could observe a 10 seconds continuous loop in both cases:
- Cal.0330G:
1st second: +3.46 s/d
2nd second: +2.80 s/d
3rd second: +2.80 s/d
4th second: +3.56 s/d
5th second: +4.17 s/d
6th second: +4.16 s/d
7th second: +4.16 s/d
8th second: +4.17 s/d
9th second: +4.16 s/d
10th second: +4.09 s/d
then it started again: the results might have changed a little bit here and there but the trend remained the same for the next lot of 10 seconds then the following lot of 10... and so on.

- Cal.E510G:
1st second: +4.43 s/d
2nd second: +5.65 s/d
3rd second: +5.70 s/d
4th second: +5.69 s/d
5th second: +5.70 s/d
6th second: +5.71 s/d
7th second: +5.70 s/d
8th second: +5.69 s/d
9th second: +5.70 s/d
10th second: +5.46 s/d
just as was the case with the other movement: the results might have changed a little bit here and there but the trend remained the same for the next lot of 10 seconds then the following lot of 10... and so on.

Conclusion: the 2 movements have identical thermocompensation schemes that feature:
- an inhibition period of 10 seconds
- continuous inhibition (digital count suppression) combined with the continuous adjustment of the frequency of the quartz oscillator(!)
The above described thermocompensation scheme is much harder to detect than the thermocompensation scheme of the thermocompensated ETA movements if one only measures the accuracy of the stepping motor of the movement.
Just to indicate the differences: I subjected my thermocompensated Bretling Aerospace (B75 = ETA E10.351) to the same tests on the Witschi and got these results (at the same 22 degrees of Celsius room temperature):
- inhibition period: 16 minutes (960 seconds)
- accuracy of the stepping motor (by the magnetic sensor): -0.05 s/d
- accuracy of the quartz oscillator (by the acoustic sensor): +9.65 s/d (steady, non-changing rate!)

Final thoughts:
- It is clear that Citizen's thermocompensation scheme is more complex than the ETA one however it does not necessary mean that one is better or more accurate than the other. Both can be very accurate if done well and both can be dissapointing if not done properly.
- The accuracy in case of both methods can be aided by digital calibration options.
- If one is only equipped with an instrument that only has a magnetic sensor then the ETA-method is easy to detect and the Citizen-method is harder to detect. However if one's instrument is a frequency counter or an acoustic sensor that can listen to the quartz oscillator then the ETA-method gets undetected while the Citizen-method is easily detectable in that scenario.
- In the next couple of weeks I'll try to test a 1998 model The Citizen (Cal.A610H) with my Witschi though I'm pretty sure that it has very similar if not identical thermocompensation scheme to the one featured with my two Citizen Exceeds.

[South Pender]

Absolutely great stuff, ppaulusz. This kind of thing is what many of us on this forum crave!

[Hans Moleman]

Wow!
That is a great breakthrough.

Great idea to use both sensor methods. Good thinking!
I would not have thought that they were periodically changing quartz frequency. But there you have it.

I agree there must be a count suppression at work. All the quartz frequency rates are high.

[dwjquest]

Nice work Mr. P. This would explain why I could not detect any compensation scheme. I only get one data point every 10 seconds. The compensation method remains hidden between the 10 second data intervals.

[ppaulusz]

Thanks, Gents!

Of course, I could not have done it without the Witschi (and my watches). The beauty of the Witschi is that it is very powerful still it's extremely easy to use.

Now, a little addition to the topic opener:
I've just checked my Omega Constellation PerpetualCalendar (Cal.1680 = ETA 252.511) on the Witschi. Here are the results at room temperature (22 degrees of Celsius):
- inhibition period: 8 minutes (480 seconds)
- accuracy of the stepping motor (by the magnetic sensor): -0.05 s/d
- accuracy of the quartz oscillator (by the acoustic sensor): between +4.10 s/d and +4.50 s/d (not as rock steady as the Aerospace's result though far from the extremes of the results of my Citizens)

Couple of extra notes:
- The on-wrist performance of the Swiss ones (Aerospace and Constellation) are +/-2 sec/year (23 hours/day on-wrist) as I calibrated them to be at their optimum when on-wrist.
- The on-wrist performance so far (after 50 days) of the Citizen Exceed Euros (Cal.E510G) is about +5 to +6 seconds per year or about 0,5 second (or a bit less) per month.
- The on-wrist performance of my other Exceed (Cal.0330G) is unknown as I've never put it on.

[Catalin]

However I believe what we see is the 'generic inhibition' that is most likely present in any modern 10$ watch with a Miyota movement - I would actually be very curious what you could see on something like that (Q&Q for instance).

Anyway that generic inhibition on a 32 kHz quartz only gets to about +/- 96 seconds/year, so there is a separate extra 'thing' which goes from that level to about 10 times better. And if that is still related to the design of the very old caliber 2700 I believe it will be almost impossible to measure it 'from the outside' ...

Oh, and I would be more tempted to believe that the acoustic sensor is not measuring the quartz raw vibrations but instead is following the mechanical 'tick' - but that should be very easy to check on E510 - when at the first click out the seconds-hand should no longer tick so any results in that position should indeed come from the quartz (or the 15-second tick of the minutes hand).

EDIT:
Disregard this - I am only now really waking-up and noticed the values in seconds/day - it seems that also the Citizen quartz is set very, very fast - a thing that we already knew about ETA TC calibers from previous measurements.

[ppaulusz]

...Oh, and I would be more tempted to believe that the acoustic sensor is not measuring the quartz raw vibrations but instead is following the mechanical 'tick' - but that should be very easy to check on E510 - when at the first click out the seconds-hand should no longer tick so any results in that position should indeed come from the quartz (or the 15-second tick of the minutes hand).

The acoustic sensor does in fact measure the quartz raw vibration! That - by the way - should have been obvious not because the Witschi user's manual says so (though that is a strong indication, in my opinion) but rather because I measured the Aerospace that has no second-hand and its minute-hand moves just once in every 30 seconds and the Witschi picks up the signal once in every second when the acoustic sensor is selected (in other words the duration of the measurement is 1 second and it is repeated as long as I want to - the sensor's LED indication is lit on and the LCD displaying the results blinks with each reading - once in every second).
Having said that, I followed your suggestion and covered the dial of my Exceed (Cal.E510G) during the measurement (via the accoustic sensor) and the results were identical to the earlier uncovered ones.

[ppaulusz]

However I believe what we see is the 'generic inhibition' that is most likely present in any modern 10$ watch with a Miyota movement - I would actually be very curious what you could see on something like that (Q&Q for instance).

Anyway that generic inhibition on a 32 kHz quartz only gets to about +/- 96 seconds/year, so there is a separate extra 'thing' which goes from that level to about 10 times better. And if that is still related to the design of the very old caliber 2700 I believe it will be almost impossible to measure it 'from the outside' ...

Sorry, I do not have a Q&Q but I do happen to have a nice non-thermocompensated Citizen Pepetual Calendar Eco-Drive watch that is fitted with a Cal.E710 movement. I've just tested that watch (at room temperature that is 23 degrees of Celsius right now) and here are the results:
- inhibition period: 10 seconds
- accuracy of the stepping motor (by the magnetic sensor): +0.11 s/d
- accuracy of the quartz oscillator (by the acoustic sensor): +5.13 (steady, non-changing rate!)

As we both have pointed out earlier, modern quartz movements (even non-thermocompensated ones) do have "generic" inhibition so the Cal.E710 is no exeption either.

Thermocompensated quartz movements do apply a more sophisticated inhibition scheme and a thermo-sensor (thermister) to achieve high-accuracy. Some manufacturers like Citizen might complicate the above by adding periodical frequency adjustment to the scheme.
In my opinion, the Witschi QT6000 is perfectly capable to prove the aboves and it just did according to this thread.

[Catalin]

The acoustic sensor does in fact measure the quartz raw vibration! That - by the way - should have been obvious not because the Witschi user's manual says so (though that is a strong indication, in my opinion) but rather because I measured the Aerospace that has no second-hand and its minute-hand moves just once in every 30 seconds and the Witschi picks up the signal once in every second when the acoustic sensor is selected (in other words the duration of the measurement is 1 second and it is repeated as long as I want to - the sensor's LED indication is lit on and the LCD displaying the results blinks with each reading - once in every second).
Having said that, I followed your suggestion and covered the dial of my Exceed (Cal.E510G) during the measurement (via the accoustic sensor) and the results were identical to the earlier uncovered ones.

Interesting - it sounds like an ultrasonic sensor or something - does it need to be in direct mechanical contact with the case/dial ? I am also not quite understanding the part with covering the dial - my idea was for a measurement with the crown out on the first click.

That also reminds me - have you ever tested a Seiko twin quartz ? at least on the 9923 I believe there is a somehow strange mention that the timing should be done with the crown out (I don't remember but probably on the first click) - and I guess it would be very interesting to see results from both crown-in and that special position ...

[ppaulusz]

...EDIT:
Disregard this - I am only now really waking-up and noticed the values in seconds/day - it seems that also the Citizen quartz is set very, very fast - a thing that we already knew about ETA TC calibers from previous measurements.

OK, explanation is accepted!

[Catalin]

Interesting - it sounds like an ultrasonic sensor or something - does it need to be in direct mechanical contact with the case/dial ? I am also not quite understanding the part with covering the dial - my idea was for a measurement with the crown out on the first click.

That also reminds me - have you ever tested a Seiko twin quartz ? at least on the 9923 I believe there is a somehow strange mention that the timing should be done with the crown out (I don't remember but probably on the first click) - and I guess it would be very interesting to see results from both crown-in and that special position ...

Hmm, this also gives me another idea - have you ever tried a SpringDrive on the Witschi ???

[ppaulusz]

Interesting - it sounds like an ultrasonic sensor or something - does it need to be in direct mechanical contact with the case/dial ? I am also not quite understanding the part with covering the dial - my idea was for a measurement with the crown out on the first click.

That also reminds me - have you ever tested a Seiko twin quartz ? at least on the 9923 I believe there is a somehow strange mention that the timing should be done with the crown out (I don't remember but probably on the first click) - and I guess it would be very interesting to see results from both crown-in and that special position ...

Witschi just calls it "acoustic sensor". This sensor is on the first "pad" from left (on the attached photo) and the magnetic sensor is on the "pad" to its right.
Yes, the sensor needs to be in contact with the caseback or the sapphire in front of the dial to receive the signal from the quartz oscillator.
In case of measuring the stepping motor the magnetic sensor needs to be in contact with the caseback or the sapphire in front of the dial for best reception.

I forgot that by pulling the crown to the first click won't stop the time measuring function of the Cal.E510G therefore I just covered its dial for the measurement to force "power safe 1 function" (that is when the second-hand stops at the 12 position and stays there till lights hit the dial).
Anyhow since that I've just carried out the same measurements with the crown pulled out to the first click. The results are the same as earlier so the acoustic sensor does receive the signal from the quartz oscillator.

I no longer have a Seiko Twin-Quartz so I can't test one on my Witschi. However bear in mind that the Twin-Quartz's technical manual was written in the late 1970s and the Witschi is the design of the early 1990s - signal receiving and filtering technology must have advanced in that 12-15 years period.

[ppaulusz]

Hmm, this also gives me another idea - have you ever tried a SpringDrive on the Witschi ???

I bought the Witschi a couple of weeks ago and it spent 3 weeks in a workshop waiting for a small repair and it came back to me 2 days ago so these are my first serious measurements with this instrument. So far I'm extremely pleased with it.
Unfortunately I do not have a SpringDrive but the person who has the 1998 model The Citizen (Cal.A660H) also has a Seiko SpringDrive and he lives in the same city as me, by the way he is known as "Mechanikus" in our forum so I will try to organise a get together with him and subject some of his fine watches to the "Witschi-test".

[Hans Moleman]

Now to the explanation of it all!
Were you expecting a different quartz frequency distribution during the 10 seconds at another temperature?

Not that I want to give you some homework.
Fiddling with temperature is quite a hassle.

[ppaulusz]

Now to the explanation of it all!
Were you expecting a different quartz frequency distribution during the 10 seconds at another temperature?

Not that I want to give you some homework.
Fiddling with temperature is quite a hassle.

Oh, Hans, I was afraid of this...

Nevertheless, your suggestion of testing the movements at a different temperature is a valid one. I'll think about the proper execution of that test.
I'd expect a different quartz frequency distribution (different results by the acoustic sensor) in case of the non-thermocompensated movements and the thermocompensated ETA movements (though the results of the acoustic sensor mesurements of the Constellation - a 5 years earlier design than the Aerospace - makes me wonder a bit - I expected a non-changing Aerospace-like rock solid frequency reading...).
In case of the thermocompensated Citizen movements I'd expect that frequency adjustment would be activated by the scheme and therefore we would not experience much differences in the measurements by the acoustic sensor.
The big question is this:
What is the more dominant factor in the Citizen's scheme: inhibition or frequency adjustment? Or is it a case of fifty-fifty?
Citizen do admit automatic frequency adjustments in case of the Cal.A660H and do not mention inhibition at all.
We'll see it sooner or later.

[Eeeb]

George, this is a very yummy thread! I must re-read to get it all - I like that. I had not realized the vibrations of the quartz crystal could be read directly... good info!

Now all I have to do is find a Witschi I can afford.

[ronalddheld]

What he said!

[South Pender]

Now all I have to do is find a Witschi I can afford.

Good luck with that. I've been looking. Make sure that whatever model you find has the fine resolution of the QT6000 (discontinued, but occasionally some used ones pop up), as some of the cheaper ones have coarser resolution.

[ppaulusz]

George, this is a very yummy thread! I must re-read to get it all - I like that. I had not realized the vibrations of the quartz crystal could be read directly... good info!

Now all I have to do is find a Witschi I can afford.

Well, it meant to be yummy, Jim.

Somehow the words affordable and Witschi just do not connect...I'm afraid.

[ronalddheld]

George is almost certainly correct with respect to the Witschi. We need a massive HEQ discount on them.