Heuer Quartz running fast

[LittleWatchShop]

20 minutes ago, JohnR725 said:

Just to be clear here mechanical issues with the electronic components or mechanical issues with the watch itself in other words the gears the stepping motor etc.

 

Mechanical issues with the train wheels and stepper motor, cannon pinion, and hands.

Crystal is fine. IC is fine.

[JohnR725]

4 hours ago, LittleWatchShop said:

Mechanical issues with the train wheels and stepper motor, cannon pinion, and hands.

Crystal is fine. IC is fine.

An easy way to disprove your theory would be if somebody would replace the circuit board. But I look forward to hearing your theories on the mechanical issues that cause the problem that's occurring.

[Flagstaff]

10 hours ago, Marc said:

Pedantry..... not pedantics 

Pedant! 

[AndyHull]

13 hours ago, LittleWatchShop said:

How you gonna mount an oscilloscope on his arm?  It works fine in a static position.

You just need to find someone who knows about making very small (and very basic) oscilloscopes.

https://github.com/pingumacpenguin/STM32-O-Scope/wiki

https://github.com/pingumacpenguin/STM32-O-Scope/wiki/Construction

 

 

[LittleWatchShop]

4 hours ago, AndyHull said:

You just need to find someone who knows about making very small (and very basic) oscilloscopes.

https://github.com/pingumacpenguin/STM32-O-Scope/wiki

https://github.com/pingumacpenguin/STM32-O-Scope/wiki/Construction

 

 

Yup.  These are my two largest scopes and my smallest.  I have a total of five.  As you see, the little one could be strapped to the body.

[KimBowers]

Thanks to everyone for your thoughts on this.

I have looked at the test point located by JohnR725 using a PC based PicoScope. It has a square wave at 512Hz when the watch is running normally.

As I posted earlier I have been able to get the watch to run erratically by warming it up in an airing cupboard.

So in order to determine if the cause of the issue is movement or heat I have done some further experimentation.

By putting the movement on a warm radiator (40-45C) I can get the erratic seconds hand stepping. Rechecking the test point also gives 512Hz. So it doesn't look like it's the crystal or the early stages of the divider chain.

I then put the probes across the coil and could see the alternating second pulses during normal running as shown on the lower trace in the attached picture. After warming up the watch the trace changes to the upper picture. As you can see the voltage is much reduced on the positive pulse. So it does look to be electronics/coil related.

I guess  my nest move is to get a replacement circuit board and coil as suggested.

Regards, Kim.

 

[LittleWatchShop]

9 hours ago, JohnR725 said:

An easy way to disprove your theory would be if somebody would replace the circuit board. But I look forward to hearing your theories on the mechanical issues that cause the problem that's occurring.

12 minutes ago, KimBowers said:

Thanks to everyone for your thoughts on this.

I have looked at the test point located by JohnR725 using a PC based PicoScope. It has a square wave at 512Hz when the watch is running normally.

As I posted earlier I have been able to get the watch to run erratically by warming it up in an airing cupboard.

So in order to determine if the cause of the issue is movement or heat I have done some further experimentation.

By putting the movement on a warm radiator (40-45C) I can get the erratic seconds hand stepping. Rechecking the test point also gives 512Hz. So it doesn't look like it's the crystal or the early stages of the divider chain.

I then put the probes across the coil and could see the alternating second pulses during normal running as shown on the lower trace in the attached picture. After warming up the watch the trace changes to the upper picture. As you can see the voltage is much reduced on the positive pulse. So it does look to be electronics/coil related.

I guess  my nest move is to get a replacement circuit board and coil as suggested.

Regards, Kim.

 

Excellent debugging.  It looks like the output driver in the IC has a problem.  OR, there is a cold solder joint along the path between the IC driver output and the coil.  Can you post a picture of the circuit board?

Until this morning, I had not watched your video.  Did not realize that static operation was also erratic.  So my theories were based on my false premise.

At any rate,

1) the crystal is good

2) the divider chain is good

3) the IC output driver could be bad

I am <not> NOW placing my bet on some resistance in series with the coil.

Yesterday, I fixed an electronic watch that turned out to be essentially an open circuit due to poor  (perhaps oxidized) connections.

Edited February 3, 2021 by LittleWatchShop
corrected "not" to "now"

[LittleWatchShop]

5 minutes ago, LittleWatchShop said:

I am <not> NOW placing my bet on some resistance in series with the coil.

Is the coil attached with screws?  If so, loosen and then re-tighten.

[LittleWatchShop]

17 hours ago, LittleWatchShop said:

Yes, I am suggesting mechanical issues.

After watching the video...I now do not believe it is mechanical (in the sense of gear train or friction or some such).

[KimBowers]

Hi LittleWatchShop,

I have attached a picture of the circuit board. I have also looked at it under a microscope and can't see any cracked tracks or dry joints. I did try loosening and retightening the screws before removing the board, but the problem is still there.

To fully eliminate the possibility of movement when wearing the watch I took it for my daily covid constitutional on the outside of my jacket, and it kept good time. So it does look like temperature is the cause. But it does seem strange that such a small rise (20C to 30C or so) is enough to cause this problem.

[LittleWatchShop]

22 minutes ago, KimBowers said:

Hi LittleWatchShop,

I have attached a picture of the circuit board. I have also looked at it under a microscope and can't see any cracked tracks or dry joints. I did try loosening and retightening the screws before removing the board, but the problem is still there.

To fully eliminate the possibility of movement when wearing the watch I took it for my daily covid constitutional on the outside of my jacket, and it kept good time. So it does look like temperature is the cause. But it does seem strange that such a small rise (20C to 30C or so) is enough to cause this problem.

Temperature seems to be the culprit.

When temperature goes up, resistance goes up (generally). But there is no obvious place where resistance is in the way of your contacts are good.

Are the battery contacts clean and engaged well with the battery?

Is it a fresh battery?  I know this is a dumb question, because you seem very proficient with regards to electronics...but I have been fooled before.

 

[KimBowers]

Yes the battery is new. Contacts look okay, but I'll give everything a clean up before I re-assemble.

Not optimistic though. Looks like I'll have to find a replacement circuit board and coil assembly.

Regards, Kim

[JohnR725]

6 hours ago, LittleWatchShop said:

These are my two largest scopes

Other than an interesting shade of blue which tells us who made it you didn't give us the numbers?

 

2 hours ago, KimBowers said:

Looks like I'll have to find a replacement circuit board and coil assembly

When you get the circuit board make sure you get one that looks similar to yours. As far as I know all of the 963 whatever circuit boards should interchange. But when you look at the cross reference sites they don't indicate that. That's because there were variations. So originally this watch came with flexible circuit board into pieces. I have attached photographs showing this. Also needed a screw with insulating washer to connect the coil to the rest of the circuit. I would stay away from the flexible circuit boards.

The flexible circuit boards are an interesting design as I'm not entirely sure how they made them. I'm also attaching additional photographs of the women's version . You notice that the conductive material can bridge over gaps which are typically don't find in standard circuit boards. To hold the quartz crystal in place there's a part of the circuit board goes over to hold it in place. In the case of the smaller women's circuit board this did not work very well. You'll notice the connections to the quartz crystal itself bridge between the circuit board in the crystal. But in the case of the women's and one of the pictures shows this it would break. That would cause a variety of interesting problems. Fortunately the company is working for was ordering so many of these we just put a request in the Switzerland and a droppable epoxy on the other side prevented the crystal from hinging and breaking its contact. But it does show flexible boards are not a good idea and watches.

 

 

 

[LittleWatchShop]

20 minutes ago, JohnR725 said:

Other than an interesting shade of blue which tells us who made it you didn't give us the numbers?

Tektronix 465B, Tektronix 2440

Tektronix is the "gold standard" for oscilloscopes.

My other Tek (not shown) is a TDS 2020b

Interesting pictures of busted PCB connectors.

[LittleWatchShop]

2 minutes ago, LittleWatchShop said:

Tektronix 465B, Tektronix 2440

Tektronix is the "gold standard" for oscilloscopes.

My other Tek (not shown) is a TDS 2020b

Interesting pictures of busted PCB connectors.

The 2440 was given to me.  I never use it.  I plan to give it away when I find the right person.  I use the 465B as a curve tracer--nothing else.

[JohnR725]

7 minutes ago, LittleWatchShop said:

Tektronix 465B

I thought it looked familiar which was why asked. I have one of those haven't used it in a long time though.

[LittleWatchShop]

9 minutes ago, JohnR725 said:

I thought it looked familiar which was why asked. I have one of those haven't used it in a long time though.

I love the feel of analog scopes, but appreciate the utility of digital ones.  I keep thinking I will get rid of it too, but it would break my heart.

[JohnR725]

Then in case anyone's curious about the technical specifications of a 963 movement on a witschi timing machine. I took a new old stock movement to work and hear are the results.

[watchweasol]

If I may stick my 5 pence worth in,  Taking a step back to the mechanical watch which has a fixed gear/wheel ratio and is driven via the mainspring. The gear train once the spring is wound will run down fast to start with and diminish as the spring weakens and eventually stop as it runs out of power. To control the power rush we introduce the escapement with a controlled oscillator (the balance). Now the power is controlled and the watch runs down in a controlled manner, at a rate set via the balance spring eg 28000 bpm. If we now remove the spring an fit a battery for power and a circuit to control the power output, a crystal that oscillates at 37.678 Hz. and the circuit controls the out put passing it to the stator,  the stepper motor at 1 Sec pulses,  the gear train of the watch will step once per second controlled by the out put of the oscillator.  Oscillators can and do stray from the given frequency ( I have changed a few)  causing timing issues. Forgive the diatribe just trying to get my head round it. My best bet is the circuit and or the oscillator.

[JohnD]

As the watch has been left in a drawer for a number of years prior to a new battery being fitted, my money is on nothing more complicated than dried up lube causing the problems, shown up when the watch is warmed up by wearing it. I have about 50 pre 1980 Seiko quartz analogue watches and this is a very common problem with watches that have been stored for a number of years.......

[jdm]

18 minutes ago, JohnD said:

As the watch has been left in a drawer for a number of years prior to a new battery being fitted, my money is on nothing more complicated than dried up lube causing the problems,

Dried lube can make a quartz watch run fast? How? That's the problem the OP is having. 

[JohnD]

7 minutes ago, jdm said:

Dried lube can make a quartz watch run fast? That's the problem the OP is having. 

I missed the degree of error (10's of seconds a minute) in the original post but 'double stepping' of the rotor does sound more mechanical related than the possibility of the oscillator running fast, as surely that would still single step, just at a faster rate than once very second?

[LittleWatchShop]

24 minutes ago, watchweasol said:

If I may stick my 5 pence worth in,  Taking a step back to the mechanical watch which has a fixed gear/wheel ratio and is driven via the mainspring. The gear train once the spring is wound will run down fast to start with and diminish as the spring weakens and eventually stop as it runs out of power. To control the power rush we introduce the escapement with a controlled oscillator (the balance). Now the power is controlled and the watch runs down in a controlled manner, at a rate set via the balance spring eg 28000 bpm. If we now remove the spring an fit a battery for power and a circuit to control the power output, a crystal that oscillates at 37.678 Hz. and the circuit controls the out put passing it to the stator,  the stepper motor at 1 Sec pulses,  the gear train of the watch will step once per second controlled by the out put of the oscillator.  Oscillators can and do stray from the given frequency ( I have changed a few)  causing timing issues. Forgive the diatribe just trying to get my head round it. My best bet is the circuit and or the oscillator.

This is intellectual discourse among very knowledgeable people seeking an answer to this puzzle--a fun exercise.

For the quartz watch of this vintage, I think the circuit is pretty simple.  Oscillator, divider chain, driver.  There are probably some control functions as well, but the "drive train" consists of these.

The oscillator comprises 1) a quartz crystal, 2) an amplifier, and 3) two capacitors (can be parasitic elements). In order for the oscillator to oscillate, the phase shift from the input of the amplifier to the output of the crystal/capacitor must be 360 degrees at an amplitude greater than but close to unity.  This requirement for oscillation is the Barkhausen criterion.  The crystal has a very high Q where the Barkhausen criterion can only be satisfied at a VERY narrow range of frequencies. Typically for a crystal of the type here, the range is +/-100ppm.

The crystal drives a divider chain consisting of fifteen flip flops configured as toggle flops.  They divide the oscillator output frequency to 2^15 yielding a 1 Hz output frequency.

The output of the divider chain then goes to a coil driver circuit which is configured in a fully differential fashion.

What can go wrong?

The oscillator will oscillate at 32768 Hz +/- 100ppm or it will not oscillate at all.

The divider chain conceivably could (due to a circuit failure) skip one or more toggle flops.  Generally these circuits have a feature built in for test purposes allowing for the portions of the divider chain to be bypassed during test so that the test can be accomplished faster than the one-second output would otherwise require.  I think this would be rare, but possible.  If the divider chain is broken, no output would be provided to the driver.  So, it is possible that under some conditions, the chain is shortened and the 1 Hz output becomes 2 Hz but not 1.2 Hz.  If the shorting occurs intermittently (say once every five seconds) then the average output of the chain could be a non-integer value.

Next up!  The driver.  The driver's purpose in life is to provide the energy to the coil at the rate dictated by the oscillator and divider chain.  If the driver is weak in one phase--not sure how the stepper motor would respond...this is an electro-mechanical component.  Maybe a weak positive pulse could cause the stepper to make a partial rotation and then the next negative pulses cause it to jump two steps.  I cannot visualize how this particular watch is configured, so am at a bit of a loss.

If this watch were running slow, I think there would be multiple candidates for the problem cause.

[LittleWatchShop]

Just now, LittleWatchShop said:

This is intellectual discourse among very knowledgeable people seeking an answer to this puzzle--a fun exercise.

For the quartz watch of this vintage, I think the circuit is pretty simple.  Oscillator, divider chain, driver.  There are probably some control functions as well, but the "drive train" consists of these.

The oscillator comprises 1) a quartz crystal, 2) an amplifier, and 3) two capacitors (can be parasitic elements). In order for the oscillator to oscillate, the phase shift from the input of the amplifier to the output of the crystal/capacitor must be 360 degrees at an amplitude greater than but close to unity.  This requirement for oscillation is the Barkhausen criterion.  The crystal has a very high Q where the Barkhausen criterion can only be satisfied at a VERY narrow range of frequencies. Typically for a crystal of the type here, the range is +/-100ppm.

The crystal drives a divider chain consisting of fifteen flip flops configured as toggle flops.  They divide the oscillator output frequency to 2^15 yielding a 1 Hz output frequency.

The output of the divider chain then goes to a coil driver circuit which is configured in a fully differential fashion.

What can go wrong?

The oscillator will oscillate at 32768 Hz +/- 100ppm or it will not oscillate at all.

The divider chain conceivably could (due to a circuit failure) skip one or more toggle flops.  Generally these circuits have a feature built in for test purposes allowing for the portions of the divider chain to be bypassed during test so that the test can be accomplished faster than the one-second output would otherwise require.  I think this would be rare, but possible.  If the divider chain is broken, no output would be provided to the driver.  So, it is possible that under some conditions, the chain is shortened and the 1 Hz output becomes 2 Hz but not 1.2 Hz.  If the shorting occurs intermittently (say once every five seconds) then the average output of the chain could be a non-integer value.

Next up!  The driver.  The driver's purpose in life is to provide the energy to the coil at the rate dictated by the oscillator and divider chain.  If the driver is weak in one phase--not sure how the stepper motor would respond...this is an electro-mechanical component.  Maybe a weak positive pulse could cause the stepper to make a partial rotation and then the next negative pulses cause it to jump two steps.  I cannot visualize how this particular watch is configured, so am at a bit of a loss.

If this watch were running slow, I think there would be multiple candidates for the problem cause.

Here is the driver stage

[LittleWatchShop]

Can someone provide an illustration or photo of the actual stepper of either this or similar vintage watch?

When the rotor sees a magnetic impulse, it will turn.  What stops it?  Well, if the magnetic energy is applied long enough for the rotor to settle to resting state, then magnetic energy itself stops it.  However, if the rotor only receives a magnetic impulse, then there must be a mechanical feature (like the keeper on a calendar ring) impeding the rotor from running past the point where it should stop.  THIS WOULD speed up the watch if it is not impeding properly.

Ideas?