Weishi vs. Vibrograf conflicting results

[LittleWatchShop]

Here are results from a Weishi 1900 and a Vibrograf B200.

The Weishi detects a beat rate of 21800 whereas the Vibrograf results are with the beat rate set to 19800.

This is a C11ks/1 which, according to Ranfft, should have a beat rate of 19800.

As you can see, they both indicate slow by different amounts, but the Vibrograf (which I inherited) has not been calibrated in at least 40 years. 

The Weishi indicates near perfect beat whereas the Vibrograf does not.

The disturbing difference is the beat rate.  Any thoughts??

My money is on the Vibrograf being correct.

[nickelsilver]

You can adjust the strike force on the B200, I think you are seeing ghost lines. The heavy line looks like the two beats, and that the watch is in beat to me. If you look inside the cover you'll see that it can detect some duplicate beats over different frequency buttons, this is down to how it chops the time signal and speed of the bar under the paper (with the spiral on it). In the pic you can see that on 19,800 it will also pick up 21,600. It's been a while since I tried but I seem to remember that the frequencies in parenthesis end up with some duplicate lines on the paper- which come to think of it could be what you're seeing.

 

If you want to know the true beat just count every other beat over 30 seconds and multiply by 4, then 60.

 

The little knob at bottom left adjust the force.

 

 

Edited November 10, 2021 by nickelsilver

[praezis]

I wonder why you did not select 21600 on the B200?

Frank

[JohnR725]

5 hours ago, LittleWatchShop said:

Weishi detects a beat rate of 21800

Actually the picture looks like 21,600? Then what happens if you change the setting for the Weishi to 19800

Then as far as calibration goes for the B200 some of it depends upon which oscillator board it has? Over the years they had different versions some running at insanely low frequencies with really huge crystals and others that much much higher frequencies divided down. So is probably quite a bit of variation of how well the different boards are going to work with time as far as needing calibration goes.

4 hours ago, nickelsilver said:

The little knob at bottom left adjust the force.

Then that may be a problem? The B200 is a real interesting timing machine made over a span of years and basically was the sole timing machine of anyone in watch repair the factories everyone use this machine. Then it came in variations they all visually more or less look the same on the outside. The US version does not have a ribbon like the machine in your picture because that's a European version. The Europeans could use cheaper paper with a ribbon we get stuck with expensive pressure sensitive paper. Then the various ways of changing the print density like I've attached an image.

The basically  a machine that's blue in color has a number B200 made over considerable quantity of years that has a whole bunch of variations. That other than the electrolytic capacitors drying up or leaking the machine will probably last forever.

 

 

[LittleWatchShop]

1 hour ago, JohnR725 said:

Actually the picture looks like 21,600? Then what happens if you change the setting for the Weishi to 19800

I ran the Weishi in both 19800 and 21600 and the only valid reading was for 21600.

I ran the B200 in both settings as well--this morning--and only 19800 gave a valid reading.

HOWEVER--this afternoon, I powered up the B200 again and ran it at 21600 and got a valid reading as shown here.

Hmmm....

I believe the striker is hitting too hard and will attempt to adjust that. 

4 hours ago, praezis said:

I wonder why you did not select 21600 on the B200?

Frank

That was the first thing I did, but I got a bad reading this morning...however, this afternoon things changed.

Edited November 10, 2021 by LittleWatchShop

[LittleWatchShop]

5 hours ago, nickelsilver said:

If you want to know the true beat just count every other beat over 30 seconds and multiply by 4, then 60.

I put it on the pickup I designed and measured the beat on my oscilloscope.  Yup...21600.  I did that after the original post.

But...the nagging thing is that all of the published data conflicts with this number.

[LittleWatchShop]

51 minutes ago, JohnR725 said:

Then the various ways of changing the print density like I've attached an image.

I played around with this method with not a lot of change.  Most change came from reducing the gain.  Not sure what that other knob is supposed to do but messing with it while running can stop the machine.

[LittleWatchShop]

I think the reason why the 21600 setting did not work before is perhaps because the gain was too high. 

[LittleWatchShop]

As a follow up.

I have four watches with a C11KXX movement.  Three are automatic and the other manual.

Two of the automatics are 19800 and the third automatic is 21600 as is the manual.

Ranfft does not give evidence of 21600 which caused me to misdiagnose the issue this morning.

[LittleWatchShop]

Ugh, one more thing.  That KNOB shown in @nickelsilverpost is a fuse.

[JohnR725]

1 hour ago, LittleWatchShop said:

I have four watches with a C11KXX movement.  Three are automatic and the other manual.

Two of the automatics are 19800 and the third automatic is 21600 as is the manual.

Here's the problem with databases what makes you think they're right? Yes I went back to the website I looked at all of the variations of C11 and they're all 19,800. But what if we look somewhere else will that agree with us? The answer is it does not agree how weird. The reason why make the reference of how weird is that I know of other databases out there that are wrong in horology.

 

52 minutes ago, LittleWatchShop said:

That KNOB shown in @nickelsilverpost is a fuse.

That's not the knob he was talking about? Yes everyone should know that was the fuse. I snipped out your image and crossed out the fuse I believe he's talking about the other Knob like thing. But I've seen another variation of the printer where the knob is on the right-hand side.

[LittleWatchShop]

52 minutes ago, JohnR725 said:

That's not the knob he was talking about?

That is why I did not quote you...I quoted @nickelsilver!  I was quite aware of the knob you referenced with the image.

[LittleWatchShop]

55 minutes ago, JohnR725 said:

But what if we look somewhere else will that agree with us?

This is somewhere else...a document I got off of this forum from years back.

But fundamentally I agree with the point you are making.  There are lots of data on the web but one must "trust but verify."

[JohnR725]

3 hours ago, LittleWatchShop said:

There are lots of data on the web but one must "trust but verify."

Problems with stuff online? It's not necessarily online but the problem with horological documentation is it was valid at the time was printed. If variations were done after was printed that may or may not show up. If the watch company never did an update document then you don't get an update.

Noticing in your image basically this is the 11 K and back to the material house website notice that a few of them are a different frequency. It looks like they must a done an updated version that runs faster.

Then the print density? Somewhere on the machine is a way of adjusting the print density and that is not done by adjusting the volume knob. I know you can adjust it from basically oh fade away to nothing and if you go the other direction you will poke holes in the paper. So somewhere on the machine is a way of adjusting.

[LittleWatchShop]

So...how does this thing work?  Well, the motor spins at a precise speed.  The motor drives a roller that his a raised spiral.  When the watch ticks, it causes a hammer to impact the roller.  If the impact hammer hits the roller at the same angular position each time, a vertical line will be marked on the paper.  If the impact hammer hits the roller at a different angular position, the line will walk to the left or to the right.  This operation in signal processing terms is similar to mixing.  For a vertical line the hammer must hit at integer multiples of the motor rotation.  In order for the lines marked to be meaningful, the motor must spin at a precise, known speed.  Using the numbers from the B200 manual and comparing with the image @nickelsilverposted, I figured out what speed the motor must be running for the different beat settings.  With a little math, I came up with these numbers showing how many rotations of the motor per hammer hit for each watch beat.  Notice that the green entries correspond to the case where you will get a double line printed.

I have not figured out how the motor precision is achieved.  Pondering that.

 

 

Edited November 11, 2021 by LittleWatchShop

[praezis]

27 minutes ago, LittleWatchShop said:

I have not figured out how the motor precision is achieved.  Pondering that.

Motor is an AC synchronous motor. Here he does not follow the mains frequency but the divided quartz frequency (range about 30 ... 60 Hz).

Frank 

edit: B200 motor frequencies are 32, 40, 44, 48 Hz 

Edited November 11, 2021 by praezis

[LittleWatchShop]

21 minutes ago, praezis said:

Motor is an AC synchronous motor. Here he does not follow the mains frequency but the divided quartz frequency (range about 30 ... 60 Hz).

Frank 

Yes, the master oscillator is 10,560 Hz.  For the four ranges, the dividers necessary are 440. 352, 320, and 293 1/3.  The logic board is comprised of two 2493s (4-bit counter) and a 7472  (JK flipflop), so a maximum divide capability is 512, so that is consistent.  The last divide is non-integer...I will give this some thought.

The beat switches control what appear to be relatively large precision capacitors (five) that connect to the motor in some fashion.  Pondering this as well.

[LittleWatchShop]

10 minutes ago, LittleWatchShop said:

Yes, the master oscillator is 10,560 Hz

Some units (perhaps later ones) use a 5 MHz oscillator.  The dividers do not work out so nicely in this case.

[JohnR725]

40 minutes ago, LittleWatchShop said:

I have not figured out how the motor precision is achieved.  Pondering that.

Have you thought of microscopic elves who feverishly push the motor at a controlled rate?

What's interesting with timing machines are some other companies use to publish nice complete schematics. Even have a few waveforms for troubleshooting. Or the vacuum tube version of this machine has way more technical information available. Then I really need to organize my PDFs are all in one place because looking at the various B200 PDFs there seems to be variations with additional bonus material?

While I'm looking at one of the other PDFs I found another way to adjust the printing pressure. Also there is a warning or a comments apparently there is no schematics for the individual boards because they discovered that inexperienced people basically our incompetence that's my interpretation of their wording. But they do have motherboard schematic this one is barely usable I might have a better one somewhere conveniently though it does not mention the motor drive frequencies. But I have seen that in one of the other manuals for some other machine.

Oh and even though there is the drawing were all the boards to go the boards are all keyed so they usually only go in one place so in case you disassemble your machine and panic over where they go they only go in one place.

 

5 minutes ago, LittleWatchShop said:

ome units (perhaps later ones) use a 5 MHz oscillator.

Are you sure that was the frequency or not some other frequency? Yes the old ones have the really nice glass tube with the visible crystal. Some of the later generation ones have integrated circuit dividers with a much higher frequency crystal.

 

[LittleWatchShop]

9 minutes ago, JohnR725 said:

Are you sure that was the frequency or not some other frequency?

As you pointed out earlier in this thread, one cannot be sure of anything, but here is where I got the 5 MHz number.

 

 

[JohnR725]

Then I knew that somewhere out there I had seen something on I suspect that this timing machine. I was right unfortunately it's not in English but it shouldn't be a problem to convert it does something that you can interpret. Lots of pretty pictures though of the inside of the machine in case you don't want to take yours apart.

https://www.pascalchour.fr/mesures/vibrograf_b200.htm

Just in case French isn't your thing this should be the English translation

https://www-pascalchour-fr.translate.goog/mesures/vibrograf_b200.htm?_x_tr_sl=fr&_x_tr_tl=en&_x_tr_hl=en-US&_x_tr_pto=nui

Found another reference with pictures and a different oscillator board.

https://jestineyong.com/vibrograf-b200-mechanical-watch-calibrator-repair/

Unfortunately in French again but some interesting information including I think references to frequencies.

http://forum.horlogerie-suisse.com/viewtopic.php?t=24156

 

[LittleWatchShop]

18 minutes ago, JohnR725 said:

But they do have motherboard schematic this one is barely usable

I have this one along with the letter of warning you mentioned.

Here is a teardown of sorts.

[LittleWatchShop]

It dawned on my while looking at these beat numbers that I could test the accuracy of my B200 by simply using an accurate quartz watch!  While you do not get the 5 (e.g.) beats per second, you do get one.  It will generate a vertical line on all settings.

I did this and it appears that my B200 is dead on.

[JohnR725]

20 hours ago, LittleWatchShop said:

Some units (perhaps later ones) use a 5 MHz oscillator.  The dividers do not work out so nicely in this case.

Maybe the reason the divider math doesn't come out right is the description of the quartz frequency doesn't exactly agree with reality? I had a suspicion I had pictures of old versus new and you'll notice that the oscillator crystal has a decimal point and a few additional numbers.

 

[praezis]

2 hours ago, JohnR725 said:

Maybe the reason the divider math doesn't come out right is the description of the quartz frequency doesn't exactly agree with reality?

Exactly. The average user was not interested in the decimal numbers but in the fact of a much higher (much better!) frequency.

 

23 hours ago, LittleWatchShop said:

The beat switches control what appear to be relatively large precision capacitors (five) that connect to the motor in some fashion.  Pondering this as well.

Still pondering?
They filter the (square) frequency,  imo  they form a LC resonator with the motor inductance.

Frank