Jump to content

Citizen 8945 circuit board repair


gkmaia

Recommended Posts

I've go a couple faulty wingmans from ebay. 

And they both have no life. 

Did fix some of the tracks and tested continuity in each track, front/back.. all good. 

Hooked them on my oscilloscope and got absolutely no signal, nor noise. Just stable positive voltage across the rails and some current usage. No shorts between positive and negative rails, in fact impedance is very high.

As per manual the IC drives everything. It has a built-in oscillator to drive the crystal but I get no signal on the crystal and coil. 

My assumption is the IC is fried... 

Any hope? ideas?

Screen Shot 2020-09-25 at 10.01.35 PM.png

8943.pdf

Link to comment
Share on other sites

Nice of you to include the manual. Did you notice in the manual that it seems be really lacking on oscilloscope use for instance? This is because in watch repair other than the basic tests in the manual like supplying power measuring current consumption. Or if you have a timing machine that can pick up the Quartz frequency that's about it. Basically it's either perceived to work or not work and you just replace the Circuit board as a module.

Link to comment
Share on other sites

Well... that is sad. I was hoping to get these guys back to life.

Yes, I noticed the lack of component level testing. It would make sense as the ASIC does everything anything that fails will lead to the IC. Unless a trace, etc... not the case here.

I've seen boards with electrolyte damage and very rarely I see IC damaged by electrolyte shorts. Specially at such low voltage and current. Also when these batteries start leaking they have pretty much no in charge so I would worry just with corrosion. I am amazed by how fragile these ICs are. Or could be that I did something wrong or missed something. 

I cant seem to find these boards anywhere. Just expensive working wingmans on ebay. Some cheap ones like the ones I bought but now knowing I am not sure I would take the chance.

 

 

Link to comment
Share on other sites

One of the unfortunate problems with watch repair is the availability of spare parts. Especially if the spare parts are older spare parts than the chance of getting them drops considerably. Just keep an eye on eBay sooner or later one will show up. Because unfortunately like the link below shows there no longer available.

 

 

 

http://www.julesborel.com/s.nl/it.A/id.181120/.f

Link to comment
Share on other sites

Judging from the contents of your posts so far, you seem to be an electronics guy.

Many things about quartz watches have puzzled me for several years.

Like... how is it possible to burn the coils with the current output of a silver oxide battery?

IC's seldom fail in other devices. It's usually other components like capacitors and power devices. Why is it that many repairers blame it on a bad IC?

Many watch pcb's have through hole connections. Is it more likely that these have failed?

How about the quartz crystal itself? Doesn't it fail?

Link to comment
Share on other sites

The crystal is quite a common failure point, and I have had some luck replacing them. If you have no oscillator, change the crystal and possibly, if it still wont go, change the load capacitors. These are the two capacitors associated with the crystal, typically these are a few pF in value, exact value will depend on the replacement crystal.

The coil burning out could also be due to crystal failure. My hypothesis is that the coil transistor in normal operation is only on for a short time, but if the oscillator fails the transistor may stay on permanently. This might cause the coil to heat up and ultimately fail. However corrosion and physical damage are more likely causes of coil failure. 

The IC failing is the least likely problem, however corrosion may eat the bond wires that connect the chip to the lead frame, or eat the lead frame itself, or eat the PCB traces. In other words the chip may be fine, but the connections to it may have failed. 

EDIT: If the watch has an LCD, you should clean the zebra strip and the contacts on the PCB. Use isopropanol or similar for this. You can also clean the contacts on the glass with isopropanol.

Edited by AndyHull
Link to comment
Share on other sites

Yes, I am keen on electronics and always curious about these small devices.

By looking at the circuit blocks there is not much I could blame but the IC.

It is quite unlikely that one of these caps is shorted. Could look at the buzzer driver transistor but I don't think it will have any affect on the oscillator. 

I never come across a faulty crystal, but I will try removing it and placing it on an oscillator.

Will post the results here shortly.

Link to comment
Share on other sites

8 minutes ago, AndyHull said:

The crystal is quite a common failure point, and I have had some luck replacing them. If you have no oscillator, change the crystal and possibly, if it still wont go, change the load capacitors. These are the two capacitors associated with the crystal, typically these are a few pF in value, exact value will depend on the replacement crystal.

The coil burning out could also be due to crystal failure. My hypothesis is that the coil transistor in normal operation is only on for a short time, but if the oscillator fails the transistor may stay on permanently. This might cause the coil to heat up and ultimately fail. However corrosion and physical damage are more likely causes of coil failure. 

The IC failing is the least likely problem, however corrosion may eat the bond wires that connect the chip to the lead frame, or eat the lead frame itself, or eat the PCB traces. In other words the chip may be fine, but the connections to it may have failed. 

EDIT: If the watch has an LCD, you should clean the zebra strip and the contacts on the PCB. Use isopropanol or similar for this. You can also clean the contacts on the glass with isopropanol.

ok, that is quite a good hint. Will look at it.

Link to comment
Share on other sites

3 hours ago, HectorLooi said:

Like... how is it possible to burn the coils with the current output of a silver oxide battery?

What makes you think that the coil burned out? I think you'll find the most common reason why quartz watches lose their coils is somebody changes the battery. The problem with getting a battery changed is the workers are always super skilled. All it takes is one touch a lot of times with some sort of metal tool and the coil is now open. The wires are really very very thin. Over the years I've seen an unbelievable quantity of coils damaged by external influence of other people.

Then there is another possible problem which is the circuitry energizes the coil you have a nice magnetic field and then the field collapses the collapsing field generates a voltage which can be considerably higher than the input voltage. Over the years that can't possibly be good on the driving integrated circuit.

 

Link to comment
Share on other sites

33 minutes ago, JohnR725 said:

Then there is another possible problem which is the circuitry energizes the coil you have a nice magnetic field and then the field collapses the collapsing field generates a voltage which can be considerably higher than the input voltage. Over the years that can't possibly be good on the driving integrated circuit.

 

Back EMF can easily be shorted out with a diode. Unless the chip maker deliberately left it out.... but we have no access to the IC schematics.... so we'll never know. ?

Link to comment
Share on other sites

6 hours ago, HectorLooi said:

but we have no access to the IC schematics

Who needs schematics when you have nitric acid?

https://zeptobars.com/en/read/Luch-quartz-wristwatch-IC-at-sc-tuning-fork

Generally they do have a clamping diode or modified transistor junction across the coil, otherwise the chip would fry. 

 

Edited by AndyHull
  • Like 1
Link to comment
Share on other sites

Toke the board mostly apart. transistor, res and caps are all good. 

The crystal is acting weird. It should be responding to a 32khz signal, but it is not. It does a flash peak when near the right frequency but nothing else.   

I imagine the signal should be a sine with between -1.5 and +1.5? 3vpkpk.

I ordered a pack of equivalent crystals. Will test those in and out to see if I get the same results. 

That sounds like a good lead. No clock, no processing. IC seems dead but may be not. Lets see.

Also, coil resistance seems way off at Mohms. Will inspect to see if the external wiring is cut. I fixed a cut coil today with some solderpaste. I lost 100k from 1.5k but it is still functional. These 100k must be a full layer. 

Do anyone know the gauge of these coil wires? The thinnest gauge I have here is 42.    

Edited by gkmaia
Link to comment
Share on other sites

My money is on the crystal. I've replaced a few in various watches. I suspect they get shocked or moisture gets in and kills them.

Here is the seller I used -> https://www.ebay.co.uk/itm/20Pcs-32-768KHz-32768HZ-Crystal-Oscillator-2-x-6-mm-TOP/193460646910?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

I have no affiliation with them, but they were a good price, and the batch of crystals I got work fine, although I didn't test them for accuracy or drift, but I suspect that would be fine too.

There are a few different variations, with different characteristics, and calling for different load capacitor values, but by and large the majority of watch crystals are of the type in that listing. You would be unlucky to encounter one of the more obscure types. I've even substituted one of these in an ancient LED watch with a big bulky crystal and it worked, so worth a try for most situations and relatively easy to replace.

Edited by AndyHull
Link to comment
Share on other sites

11 minutes ago, AndyHull said:

My money is on the crystal. I've replaced a few in various watches. I suspect they get shocked or moisture gets in and kills them.

Here is the seller I used -> https://www.ebay.co.uk/itm/20Pcs-32-768KHz-32768HZ-Crystal-Oscillator-2-x-6-mm-TOP/193460646910?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

I have no affiliation with them, but they were a good price, and the batch of crystals I got work fine, although I didn't test them for accuracy or drift, but I suspect that would be fine too.

There are a few different variations, with different characteristics, and calling for different load capacitor values, but by and large the majority of watch crystals are of the type in that listing. You would be unlucky to encounter one of the more obscure types. I've even substituted one of these in an ancient LED watch with a big bulky crystal and it worked, so worth a try for most situations and relatively easy to replace.

Nice, those are the exact same as the ones I got. 

https://nz.rs-online.com/web/p/crystal-units/5476979/?sra=pstk

They come on a pack of 10 so I will have some fun experimenting with them.

Edited by gkmaia
Link to comment
Share on other sites

Tubular crystal units are hermetically sealed using glass, and the crystal itself is sealed under vacuum, so they are prone to damage. Furthermore any corrosion due to a leaked battery may creep up the legs and cause the vacuum to fail and moisture to get in.

Be a little careful when bending the legs, to avoid stress cracking the glass. Don't for example bend them too close to the body. This also applies to other types of crystal, and old germanium transistors and diodes for example. Also be careful not to overheat them when soldering, as this too may stress crack the glass.

The data sheet on the RS site goes into more detail about this.

https://docs.rs-online.com/54bc/0900766b80c643b2.pdf

Crystals are not particularly static sensitive, but electronic crystal oscillators (those square canned 'crystals') may be, depending on the technology as they often have CMOS driver circuits. 

Edited by AndyHull
Link to comment
Share on other sites

3 hours ago, gkmaia said:

the circuit seems to be able to operate just as a stand alone PCB. Is that correct?

I'm confused? I see a wires attached normally there is not wires when it's in the watch so are you fixing this for a specific watch or for some project?

Link to comment
Share on other sites

4 hours ago, gkmaia said:

 

the circuit seems to be able to operate just as a stand alone PCB. Is that correct?

If I understand your question, then yes, the circuit should operate fine without the display. The LCD plays a passive role in the circuit, you would be able to remove and re-fit it without any noticeable effect on the controller (assuming you didn't short out anything in the process).

To answer @JohnR725's question,  I presume the two wires are simply to allow the board to be from a suitable power supply while testing it. This makes life a little easier while trying to probe very small stuff for signals. 

@gkmaia You have access to an oscilloscope, good.  Be aware that you may be loading down the crystal while testing it, due to stray capacitance, so an oscilloscope with a shielded probe might give you a better idea of what is actually going on than a multimeter. Having said that, I have used the Hz range on my multimeter with some success, so I may be over thinking the problem. 
 

Edited by AndyHull
Link to comment
Share on other sites

1 hour ago, JohnR725 said:

I'm confused? I see a wires attached normally there is not wires when it's in the watch so are you fixing this for a specific watch or for some project?

I hacked the power supply of it's pcb, that is why you see the wires.

Interesting as the positive is the chassis. Never seen that. There must be a reason, something to do with the charge of the human body as it is in contact with it permanently. 

I am fixing this watch, yes. All the components have been tested individually, apart for the IC as I cannot test it. 

Link to comment
Share on other sites


Can you measure the capacitance of that little trimmer capacitor? If that has been cranked down so hard that it has shorted out, then it will kill the oscillator.

The positive chassis may be an attempt to limit the effects of corrosion, but it may simply be a convenience when laying things out.

Link to comment
Share on other sites

8 minutes ago, AndyHull said:

If I understand your question, then yes, the circuit should operate fine without the display. The LCD plays a passive role in the circuit, you would be able to remove and re-fit it without any noticeable effect on the controller (assuming you didn't short out anything in the process).

To answer @JohnR725's question,  I presume the two wires are simply to allow the board to be from a suitable power supply while testing it. This makes life a little easier while trying to probe very small stuff for signals. 

@gkmaia Do you have access to an oscilloscope? Be aware that you may be loading down the crystal while testing it, due to stray capacitance, so an oscilloscope with a shielded probe might give you a better idea of what is actually going on. having said that, I have used the Hz range on my multimeter with some success, so I may be over thinking the problem. 

Oh yeah, Scopes, I have 15 of them. Signal generators, etc.... all I need. 

I tried to excite the crystal with 32khz from my generator, hoping to see a spike in amplitude, but nothing. Then I hooked the crystal to my Oscilator, and nothing. Did try with some crystals I am sure work and they worked. So my gear and methods are ok.

Not only the display  is off, but the coil. 

I am 100% sure it should work without the LCD, but not sure if it should work without the coil. I guess it should work without the coil, as i have seen anadigi watches with faulty analog but woking digital.

I am almost spending $200 on ebay to get a working one so I can see this signals.

I have an old Tag with an ETA I am working on that sends signals with no coil. I could not see anything similar on the wing man

 

Edited by gkmaia
Link to comment
Share on other sites

4 minutes ago, AndyHull said:


Can you measure the capacitance of that little trimmer capacitor? If that has been cranked down so hard that it has shorted out, then it will kill the oscillator.

The positive chassis may be an attempt to limit the effects of corrosion, but it may simply be a convenience when laying things out.

I can try that tomorrow. If it is more than 20p. Less than that I do not have gear. The other two were 140nf each. 

Link to comment
Share on other sites

I feel it is pretty unlikely it needs the coil to operate. The analog part of the circuit is likely to consist of a simple driver transistor for the coil, fed by pulses from the divider from the crystal oscillator. No feedback or anything like that.

image.thumb.png.2acd2b2cbae8dbb39d87f65612bec498.png

I don't have an exact datasheet for the crystals, but they are likely to be something like the above.
The load capacitors are a faff, and they may be the cause of the problem. If you don't have 6pf or 12pf capacitors to hand, try using the stray capacitance from your fingers, or two short pieces of insulated wire as "ghetto" capacitors to see if you can tickle it in to life. Often once you get it started a crystal oscillator will continue to oscillate, even without the load capacitors. Once you remove the power it will die again obviously. You could also try removing the capacitor and injecting a 1V or so sine wave in to the oscillator through a 100k resistor or something like that to see if the rest of the circuit wakes up.

  • Like 1
Link to comment
Share on other sites

24 minutes ago, gkmaia said:

If it is more than 20p. Less than that I do not have gear.

If you can measure 20pf then you will at least get an estimate of their value. They are likely to be less, but put a know value in parallel and see how much its value increases by when paired with the ones on the PCB. They are not that precise, so 6pf or 12pf +/- 50% might be close enough to get it to at least start to oscillate, all be it not at exactly 32768 Hz +/- 20ppm per the datasheet.

I'd take "running near to the correct frequency" over "not running at all" as a win, at least till we get the issue understood.

EDIT: On the dosimeter I fixed recently, which uses a 32768 Hz crystal I simply removed the load capacitors and replaced the crystal. It worked, and since I wasn't shooting for accuracy I simply left it like that.

Edited by AndyHull
Link to comment
Share on other sites




×
×
  • Create New...