Kundo Want to replace coil driver in Kundo electromechanical clock

[debrota]

I have a small Kundo electromechanical (quartz) torsion pendulum clock. See photos. I am unable to find any information on this clock online, but perhaps I am looking in the wrong place. It appears to have a very simple coil driver circuit, which connects to the driver coil by just two wires. The circuit has just a quartz crystal, a fixed capacitor, a variable capacitor, and an 8-pin DIP integrated circuit, all driver by a single AA battery (1.5VDC). I'm unable to find info on either the crystal or the IC from their markings.

It appears that the coil driver circuit is fried, and I assume this is because the variable capacitor developed a green deposit within it. I tried to remove the deposit with 20% acetic acid, but apparently did not succeed. My oscilloscope does not detect any pulses coming from the circuit even when I find DC voltages of ~0.5VDC on some of the IC pins.

So I'd like to replace the coil driver circuit somehow. Any suggestions?

 

[debrota]

Oops. I confused two clocks I am working on. The clock pictured is NOT a torsion pendulum clock. It is simply a Kundo electromechanical (quartz) clock. I'd like to replace its coil driver circuit.

 

[Willie X]

There is some hope in finding these kind of parts in used condition, or in another like clock. But, the chances are slim that a replacement (if found) will be useable. Willie X

 

[Schatznut]

That variable capacitor is apparently a trimmer and its absence should not affect the oscillator's ability to run. It's going to be only a matter of a few pf at most, and the green junk probably has a much higher parasitic capacitance, which could swamp the circuit. What happens when you remove the variable cap completely and clean the circuit board carefully? It might not be terribly accurate but it should run. At 1.5VDC, it seems unlikely that anything in the IC would get cooked. The date code on the IC indicates it was made in 1982, long before the days of ASICs, and with a 1.5VDC, it's probably CMOS. If so, it will be static-sensitive, so careful handling is indicated. It would be interesting if you could trace out the PCB and get an idea of what the overall circuit is.

 

[shutterbug]

Check your coil also. When they go, they can be a real pain to replace or repair.

 

[Wimberleytech]

Lots of unknowns here. Since you have an oscilloscope, you most likely have the skill to remove the crystal and build an oscillator using an unbufferred CMOS inverter (e.g., 74HC04). Measure the frequency. You have the crystal frequency now.
For the IC? That is probably an asic. Yes, there were asics designed for specific customers in the early 80's. Often gate arrays patterned for the asic application.
My guess is that the crystal is 32,768 kHz and the asic divides that down to 1 Hz, (for example). Replacing this board with your own design would relatively trivial using modern chips.

 

[Raymond101]

Hi the circuit is just a simple oscillator. That outputs 1hz pulse to the coil . You have quite a few options. NE555 . Or use 2 generally purpose transistors as a astable . If you are good with electronics use a mini arduino
And use blink code set for 1000
And you will have exact time adjustment.
Ps never use ascetic acid on any electronics or clocks . It will eat and destroy brass copper etc.
Switch cleaner or IPA
The small cap was probably 2-20pf

 

[Wimberleytech]

Sure 555, arduino, etc will work electronically but size matters (and a 555 will be difficult to make accurate). If you want to get this into something close to the size available in the case, you need a compact solution. CD4060 and a CD4013 in SOIC to get you to 1Hz. Probably need a driver (maybe just a transistor) and a diode for the coil. These chips are rated at 3V, so iffy at 1.5. You might need to modify your battery system.

 

[Raymond101]

Sure 555, arduino, etc will work electronically but size matters (and a 555 will be difficult to make accurate). If you want to get this into something close to the size available in the case, you need a compact solution. CD4060 and a CD4013 in SOIC to get you to 1Hz. Probably need a driver (maybe just a transistor) and a diode for the coil. These chips are rated at 3V, so iffy at 1.5. You might need to modify your battery system.

I could not see the circuit on the board there will be a back emf from the coil . The ic were a divider to get the xtal to 1hz .
If you have an oscilloscope. Replace the cap with a new pf10 ceramic cap and make sure there is a diode on the coil .
The coil will be fine as it's getting a sine pulse so your 1.5v will be 1.5 x 1.8 apx @ the coil.
Just clean the board with IPA replace the trim cap and try it . The chip is probably ok .

 

[Schatznut]

Check out this little circuit - about three bucks apiece at Digi-Key. Would require a transistor to drive the coil (MOSFET?), but a really slick widget.

 

[Wimberleytech]

Yup, that is a cute circuit. It will have to be programmed via I2C in order to get it in a useful state. Working voltage is great...down to 1.1 V.

 

[Raymond101]

Check out this little circuit - about three bucks apiece at Digi-Key. Would require a transistor to drive the coil (MOSFET?), but a really slick widget.

Yup that is a cool chip , & as you said with a mosfet, or maybe an isolated gate would be better.
. You would get a very accurate time peice. 1ppm . What's that about a second ever 10 years .
Go for it . Put a photo of the working device.

 

[Wimberleytech]

Another idea...remove the guts from a quartz movement...one that is built to drive long hands. See if it can drive this coil. Simple enough to test and very cheap.

 

[Raymond101]

Check out this little circuit - about three bucks apiece at Digi-Key. Would require a transistor to drive the coil (MOSFET?), but a really slick widget.

Yup that is a cool chip , & as you said with a mosfet, or maybe an isolated gate would be better.
. You would get a very accurate time peice. 1ppm . What's that about a second ever 10 years .
Go for it . Put a photo of the working device.

 

[Wimberleytech]

a mosfet, or maybe an isolated gate

A MOSFET has an isolated gate by its very nature.

 

[Raymond101]

A MOSFET has an isolated gate by its very nature.

Yes I was thinking of IGBT.
There is different you need to choose a mosfet with a low voltage gate . A IGBT has a lower requirement to give your coil a hard switching power.
Maybe better than a mosfet.
You will have to do some research. As I don't know any details about the coil . Or the movement. Is it a single pulse & gravity back swing. Or a reverse pulse ? .

 

[Raymond101]

A MOSFET has an isolated gate by its very nature.

Have a look at this article.

Kundo Electronique Repair

ESP Clock Information - Kundo Electronique motor repair.

sound-au.com

 

[Wimberleytech]

Have a look at this article.

Kundo Electronique Repair

ESP Clock Information - Kundo Electronique motor repair.

sound-au.com

The Kundo described in this article is not the same as the one shown in the OP.

 

[Schatznut]

The Kundo described in this article is not the same as the one shown in the OP.

Agreed - it's a totally different animal.

 

[Wimberleytech]

Looks like debrota has left the building so we are now left to our own devices (pun intended!!).
I wish I had his clock...I would make it work.

 

[debrota]

Here's an update... sorry to have been AWOL for a while.

First, I believe the coil is OK. I took a cheap 1.5VDC quartz clock module apart, and put its circuit on my oscilloscope... it basically puts out a positive pulse, then a negative pulse, then a positive pulse, then a negative pulse, etc. I connected the output of this circuit to the coil in my Kundo clock. The result was that the second hand twitched back and forth. That's to be expected, since the drive gear magnet in the field of the coil seems intended to be driven in just one direction. There is no escapement or other means to turn a back-and-forth rotation of the drive gear magnet into continuous or interrupted forward rotation for the gear train leading to the clock hands.

I believe the circuit in the Kundo clock to simply enable a self-starting synchronous motor. If it was working, I think the clock's second hand would just sweep smoothly around at one revolution per 60 seconds, and there would be no "ticking".

I've not been able to find a signal from the Kundo clock circuit on my oscilloscope. I don't see either pulses or any sort of repeating pattern, as I would expect. I agree that a single AA battery has not likely killed the IC, but I wonder if the gunked-up variable cap is shorted, bringing the output of the circuit to just zero or some DC.

I've not yet tried to get the circuit board off the battery holder, since I suspect doing this will make it difficult to put the holder back on. I suppose I will, however, since I'd like to trace the simple circuit on the board, and I can't do this without taking the battery holder off. I could also replace components one by one.

I've been unable to find any data sheets for either the IC or the crystal in the Kundo circuit. If anyone has any ideas where to find these...

I'm contemplating programming an Arduino to drive the magnet with the coil. I assume only a very small amount of current will be needed, likely within the capabilities of an Arduino output pin. I can easily play with frequency of pulses, etc. until I get continuous forward rotation from the magnet. If I get the clock running with an Arduino drive circuit, I'll then know what sort of simple circuit might accomplish the same end.

 

[Schatznut]

Debrota, you are going to have to get that circuit separated from the battery holder sooner or later, and taking that variable cap off and removing any residual corrosion from the board is the logical next step before resorting to working up an Arduino circuit.

 

[Raymond101]

Here's an update... sorry to have been AWOL for a while.

First, I believe the coil is OK. I took a cheap 1.5VDC quartz clock module apart, and put its circuit on my oscilloscope... it basically puts out a positive pulse, then a negative pulse, then a positive pulse, then a negative pulse, etc. I connected the output of this circuit to the coil in my Kundo clock. The result was that the second hand twitched back and forth. That's to be expected, since the drive gear magnet in the field of the coil seems intended to be driven in just one direction. There is no escapement or other means to turn a back-and-forth rotation of the drive gear magnet into continuous or interrupted forward rotation for the gear train leading to the clock hands.

I believe the circuit in the Kundo clock to simply enable a self-starting synchronous motor. If it was working, I think the clock's second hand would just sweep smoothly around at one revolution per 60 seconds, and there would be no "ticking".

I've not been able to find a signal from the Kundo clock circuit on my oscilloscope. I don't see either pulses or any sort of repeating pattern, as I would expect. I agree that a single AA battery has not likely killed the IC, but I wonder if the gunked-up variable cap is shorted, bringing the output of the circuit to just zero or some DC.

I've not yet tried to get the circuit board off the battery holder, since I suspect doing this will make it difficult to put the holder back on. I suppose I will, however, since I'd like to trace the simple circuit on the board, and I can't do this without taking the battery holder off. I could also replace components one by one.

I've been unable to find any data sheets for either the IC or the crystal in the Kundo circuit. If anyone has any ideas where to find these...

I'm contemplating programming an Arduino to drive the magnet with the coil. I assume only a very small amount of current will be needed, likely within the capabilities of an Arduino output pin. I can easily play with frequency of pulses, etc. until I get continuous forward rotation from the magnet. If I get the clock running with an Arduino drive circuit, I'll then know what sort of simple circuit might accomplish the same end.

without being rude I already told you that it was reversal pulse in my previous reply . ether sine or square.
The chips were customized for the each model . On some of the Kundo clocks it is not uncommon that there is a small magnet bedded in the arm .
You will have to reverse engineer the board which should not be hard . Assuming that the Acetic acid you used to clean the trim cap has not removed all the copper tracing WD40 would have been less destructive . If you know electronics . Switch cleaner non oil version should have been your first choice.
..
If that was my clock I would have removed that PCB ASP . taken a high res photo of the complete underside PCB tracking. The UTC xtal look on the top end or the other side .

 

[debrota]

I applied acetic acid only to the trimmer capacitor, with a Q-tip. It did manage to remove a good deal, but not all of the green corrosion from the trimmer. I have the entire circuit board with battery holder soaking in 100% isopropyl alcohol at the moment. I plan to try powering on the board again soon, to see if the soaking has helped, but I'm not optimistic.

My current assumption is that if the board was working, it would generate only unidirectional pulses. Unless perhaps the reverse polarity pulses are designed to keep the motor from spinning too fast?

 

[Schatznut]

Why do you seem resistant to the idea of taking it apart and removing the trimmer capacitor? All the residual crap and the acid, which is polar and probably conducts current, is probably adding enough stray capacitance to swamp the oscillator circuit.

 

[Raymond101]

I applied acetic acid only to the trimmer capacitor, with a Q-tip. It did manage to remove a good deal, but not all of the green corrosion from the trimmer. I have the entire circuit board with battery holder soaking in 100% isopropyl alcohol at the moment. I plan to try powering on the board again soon, to see if the soaking has helped, but I'm not optimistic.

My current assumption is that if the board was working, it would generate only unidirectional pulses. Unless perhaps the reverse polarity pulses are designed to keep the motor from spinning too fast?

Your assumptions are painfully
Weird. . The timing cap plate were gold plated . An new one is less than a $ .
The circuit will run with out it .
. Also the chip maybe a hybrid.
Cleaning with IPA . Does not meaning drop it in a tank soaking in the stuff .
You are repairing a delicate clock with the equivalent of using a 20kg sledgehammer.

 

[Wimberleytech]

A better solution to clean the PCB would be 50/50 IPA, acetone. That is what I use after assembling a board.
Removing the variable capacitor is a good idea. The oscillator should still work, but in the rare case, drop a 20pF cap in its place.

 

[debrota]

About ready to completely give up on the Kundo circuit board. IPA and Deoxit have removed corrosion. Still no apparent output from the IC other than 60 Hz AC hum (when I touch the circuit board). No pulses, anyway.

I mentioned in an earlier post that there is no escapement. If I manually rotate any gear in either direction, the drive magnet with its directly attached gear turns accordingly. I believe the clock basically contains an AC synchronous motor driven by the circuit from an single AA battery. I wish I knew what AC frequency the circuit was trying to generate.

 

[Wimberleytech]

I hope you are not giving up on the clock!!!!

 

[Wimberleytech]

I wish I knew what AC frequency the circuit was trying to generate.

Count the teeth on the gears and pinions and calculate the required frequency of the "motor" knowing, of course, the minute hand makes one full revolution in an hour.

 

[Schatznut]

Somebody once said, "You can lead a horse to water but if you can make him float on his back, you've got something." This horse don't float. I'm all done here.

 

[Raymond101]

It's a shame that kundo designs such beautifully designed clocks
For a 100 years that work & keep perfect time so Neanderthal man didn't have to rely on the sundial. .
I guess debrota would have trouble fixing that as well .
One dead horse.
Put it on Ebay 4 parts only ..

 

[debrota]

I have sketched the simple circuit on the Kundo clock board... see attached. All traces seem free of corrosion after cleaning. The positive terminal of the AA battery is connected only to pin #1 on the 8-pin DIP integrated circuit. Only two wires connect the IC to the coil.

 

[debrota]

I have now tried connecting the Kundo clock coil to the coil output of two different 1.5VDC cheap quartz clock circuit boards, with the same result. Attached is a scope photo of the output of one of the quartz clock circuit boards I have tried. When this alternating sequence of positive and negative pulses is applied to the Kundo coil, the drive gear twitches forward, stops, then twitches backward, then stops, then this pattern repeats. As I have previously described, I don't get any pulses or sinusoidal waveform from the Kundo circuit board.

If anyone has a quartz clock circuit board which produces a different sort of output, I'd like to see a scope photo. I plan to try driving my Kundo coil with an AC voltage (in the frequency range of 1-10 Hz, with amplitude ~1 V P-P, sine, triangle, square, etc.) when my function generator is working again. I'm also going to try applying a pulse train from a device intended to use pulse-width modulation to control motor speed... I found a cheap device of this sort and have ordered it.

Any other suggestions for how to drive the coil? As far as I can tell, the gear train from the drive gear (magnet sitting in the coil's field) to the hands on the clock is in perfect working order. There are multiple jewels in the clock. So if I could correctly get the drive gear turning in the forward direction, I think the clock would be operational.

 

[Wimberleytech]

I have now tried connecting the Kundo clock coil to the coil output of two different 1.5VDC cheap quartz clock circuit boards, with the same result. Attached is a scope photo of the output of one of the quartz clock circuit boards I have tried. When this alternating sequence of positive and negative pulses is applied to the Kundo coil, the drive gear twitches forward, stops, then twitches backward, then stops, then this pattern repeats. As I have previously described, I don't get any pulses or sinusoidal waveform from the Kundo circuit board.

As suspected, the Kundo board is just dead, so time to move beyond that.
It appears to me that the pulses from the quartz clock board are just too narrow and thus not enough energy is delivered to the coil to make the pendulum move. I doubt the original design output a sinewave, but could well have been a square pulse with something less than 50% duty cycle..

If anyone has a quartz clock circuit board which produces a different sort of output, I'd like to see a scope photo.

I will dig one out today and take some pictures. I expect to see the same thing though. For these small movements, not much energy is required to move the hands.

I plan to try driving my Kundo coil with an AC voltage (in the frequency range of 1-10 Hz, with amplitude ~1 V P-P, sine, triangle, square, etc.) when my function generator is working again.

This is a very good idea. It is what I would do. I have a Siglent ARB that is perfect for this kind of testing

I'm also going to try applying a pulse train from a device intended to use pulse-width modulation to control motor speed... I found a cheap device of this sort and have ordered it.

I am not sure there is much value here. Do the other experiments first.

Any other suggestions for how to drive the coil? As far as I can tell, the gear train from the drive gear (magnet sitting in the coil's field) to the hands on the clock is in perfect working order. There are multiple jewels in the clock. So if I could correctly get the drive gear turning in the forward direction, I think the clock would be operational.

Can you make a model of the gear train--counting teeth so that you know what the frequency of the pendulum must be to run properly?

 

[Raymond101]

That chip is custom. I believe it has a divider by 15 . It also has a small inductive coil pick up . That picks up the pulse of the first trigger and then the reverse pulse . So each is set to the swing . Ie tick space tock .
You will have to count the teeth of the cog train to calculate how many beats are required.
If you have a signal gen & scope. Start with about 60 - 120
Square wave so u have 1, 50% pulse positive space 50% negative.
On your second Chanel make a small ferrite coil with about 6 to 10 turns of thin wire about 32awg .place the pickup as close to the pendulum escapement as possible.
The escapement has a small magnetic inserted into it .
Pulse 1 pulse pos or negative @ 50% and measure the return pulse on the second Channel.
Set probes @×10
An analog scope is best.
If you have pass ,,fail on your scope may help if you know how to use it .
.

 

[Wimberleytech]

That chip is custom. I believe it has a divider by 15 . It also has a small inductive coil pick up .

Where did you find this?
Nothing in the description so far indicates any feedback. The timing is controlled by the crystal--it is open loop.

 

[Raymond101]

Where did you find this?
Nothing in the description so far indicates any feedback. The timing is controlled by the crystal--it is open loop.

Even the very early kundo clocks used a feed back . The electronic circuit is emulating a mechanical pendulum. The new Chinese cheapo battery quartz use just a timed servo with a dot controller ic. His clock was made in the 80s . And all the very early kundo clocks just used reed relay and a magnet
Which is embedded in the coil assembly.
to simulate the gravity center of swing . The newer ones used transistors and a sence coil other wise the swinging servo or bar depending on his escapement will not have a stable center point .
The controller were built by uts
Germany .
All these clocks work on the same principle. Only technology has changed.
The first electromagnet pendulum did not use transistors. Pre 1900 ..

As he has not supplied a decent photo of the exact escapement servo .

Like any clock it's the pendulum has to be true. The manufacturer will change the gears etc to suit the timing of the pendulum or electromagnet pendulum. .
He has several options count the cog train as already suggested. To find the beat ..He has a scope so he could check the xtal frequency or confirm its 32.765.
It's a UTS xtal unclassified.
That chip is built for that clock model only .
Also did he remove the chip from the pcb and check if there were link wires.?

So far I not even sure that it's not a replica kundo.
.

 

[Raymond101]

Where did you find this?
Nothing in the description so far indicates any feedback. The timing is controlled by the crystal--it is open loop.

Even the very early kundo clocks used a feed back . The electronic circuit is emulating a mechanical pendulum. The new Chinese cheapo battery quartz use just a timed servo with a dot controller ic. His clock was made in the 80s . And all the very early kundo clocks just used reed relay and a magnet
Which is embedded in the coil assembly.
to simulate the gravity center of swing . The newer ones used transistors and a sence coil other wise the swinging servo or bar depending on his escapement will not have a stable center point .
The controls were built by uts
Germany .
All these clocks work on the same principle. Only technology has changed.
The first electromagnet pendulum did not use transistors. Pre 1900
As he has not supplied a decent photo of the exact escapement servo .
Like any clock it's the pendulum has to be true. The manufacturer will change the gears etc to suit the timing of the pendulum or electromagnet pendulum. .
He has several options count the cog train as already suggested. To find the beat ..He has a scope so he could check the xtal frequency or confirm its 32.765.
It's a UTS xtal unclassified.
That chip is built for that clock model only .
Also did he remove the chip from the pcb and check if there were

 

[debrota]

Update!

I connected the coil to a function generator, set to produce a 1 Hz wave. I was then able to get the clock gears turning in the forward direction with either a sine wave or a square wave. I've not confirmed that it is keeping correct time, but the second hand is "ticking" forward at what seems to be a proper speed. Interestingly, I needed to boost the wave voltage higher than 1 V P-P to get the clock to keep running continuously. I wonder if the custom IC on the Kundo circuit board is designed to switch the 1.5VDC (single AA battery) input to it back and forth so as to produce an AC driving voltage of 1.5-2.0 V P-P. for the coil.

I am unable to tell if there is a reed switch somehow embedded in the coil, so as to emulate a pendulum, but I suppose it is possible. The effect would be to use positive-going portions of the coil driving voltage to rotate the drive gear magnet forward, and to ignore the negative-going portions, right?

I'm also unable to tell if my clock is a genuine Kundo product or a knock off product whose maker chose to label it as Kundo. See the phots I have posted if you have any other ideas regarding confirming its authenticity.

Once again, I apologize for my initial post in this thread, in which I erroneously described the clock as having a torsion pendulum. It does not have any mechanical pendulum or escapement that I can see.

I've greatly appreciated the comments and suggestions from this forum!

Dave

 

[Wimberleytech]

Update!

I connected the coil to a function generator, set to produce a 1 Hz wave. I was then able to get the clock gears turning in the forward direction with either a sine wave or a square wave. I've not confirmed that it is keeping correct time, but the second hand is "ticking" forward at what seems to be a proper speed. Interestingly, I needed to boost the wave voltage higher than 1 V P-P to get the clock to keep running continuously. I wonder if the custom IC on the Kundo circuit board is designed to switch the 1.5VDC (single AA battery) input to it back and forth so as to produce an AC driving voltage of 1.5-2.0 V P-P. for the coil.

I am unable to tell if there is a reed switch somehow embedded in the coil, so as to emulate a pendulum, but I suppose it is possible. The effect would be to use positive-going portions of the coil driving voltage to rotate the drive gear magnet forward, and to ignore the negative-going portions, right?

I'm also unable to tell if my clock is a genuine Kundo product or a knock off product whose maker chose to label it as Kundo. See the phots I have posted if you have any other ideas regarding confirming its authenticity.

Once again, I apologize for my initial post in this thread, in which I erroneously described the clock as having a torsion pendulum. It does not have any mechanical pendulum or escapement that I can see.

I've greatly appreciated the comments and suggestions from this forum!

Dave

Excellent work debrota! The coil is driven differentially, so the PP voltage with a fresh battery would be close to 3V PP. I would set your function generator to at least 2 V P-P.

Anxious to hear about the correct frequency required to make the clock keep time!!

 

[Raymond101]

Ok atleast you now have a forward motion on your second hand . Now all you have to do is time it against the time from your phone. If it's happy with 1hz
+_ record when it's exactly correct. Build a new circuit. .
I wouldn't not worry if it's a genuine kundo. As this will not help you at this moment .
From your original photos I already saw the ferrite on a white nylon spindle. Which is the uts design.
Just work on building an astable use jelly bean transistors .
The original circuit is very simple. Try referring to the PDF that I load up before and even one said it was a different animal. It will be good guidance.
& these circuits are fairly general to most of this type . Yes you will have to tweet it a bit.

 

[Wimberleytech]

All these clocks work on the same principle. Only technology has changed.

In general (not specific to Kundo), there are three ways to convert electrical energy to a time piece ( a watch or clock)
1) Using an energizing coil and a sense coil to drive a pendulum that is magnetic. The oscillating pendulum provides electrical feedback to the sense coil in this scheme. A variation on this would not use a sense coil, but instead, a reed switch that is activated by a magnetic field (from the pendulum). I would say that the Accutron watch falls in this group where instead of a pendulum a tuning fork is the resonant component and instead of gravity helping out, it is Young's modulus.
2) Use a precision electrical source to drive a magnetized rotor. This is essentially a very slow motor. It is how quartz watches and clocks work. The precision is achieved using (typically) a quartz crystal operating and 32,768 kHz.
3) Use a solenoid that is activated by a switch coupled to a conventional clock movement. The clock movement in this case is generally a classic balance escapement, but does not need to be. When the clock runs down, a switch is activated that winds it up. I have not seen one of these in 50 years, but I have not been looking

Is there another way? I guess you could make a homopolar motor and it would be the source of oscillation. But...I doubt it.
Oh yes, there is the idea I came up with as a kid. A ball sits in a track that has a slight tilt. When the ball rolls down the track, it hits a switch that energizes a solenoid and shoots the ball back up the track. This contraption is the oscillator. Then thee is an escapement that is tripped each time the ball pass it (mechanically flipping the escapement pallet fork back and forth).

I cannot think, offhand, of any other ways, but I am all ears.

 

[Raymond101]

You have covered the basics idea's. Today's electronics make it much simpler.
The very old pre transistors Era used hand wound relays that would switch as the pendulum moved over the center point . And pull the verge either left or right. With a steal bar though the center of each coil. Clumsy but worked.
The other with ball bearing running over a zigzag seesaw platform.
I love this method it's fascinating to watch.. my aunt when she was alive had one of these clock she payed about 10k for it . .
I am familiar with most of the electromagnet pendulum systems. Just a hobby and long interest I have in electro mechanics .

 

[Wimberleytech]

Like I said in an earlier post, get a movement that outputs a signal like this one I just measured and try it.
This movement is generating a sweep so it is running faster than you need, but I bet there are ones pulsing once per second with a similar waveform. I am guessing, of course.

 

[debrota]

Color me stupid. Or maybe just lazy.

It occurred to me that when I earlier connected the Kundo coil to the output of a cheap quartz clock circuit board (powered by a single fresh AA battery), the drive gear twitched but did not make forward progress. When I just now powered this same circuit board from my adjustable DC power supply set to 1.75VDC, its output was then sufficient to drive the Kundo coil forward, moving the second hand crisply once per second.

I believe I can power the cheap quartz clock circuit board from two AA batteries instead of one without destroying it, and then I'll have something to install in the Kundo clock base. Hopefully boosting the coil drive voltage will not destroy the Kundo coil, since the pulses have what look like <10% duty cycle.

 

[Wimberleytech]

Color me stupid. Or maybe just lazy.

It occurred to me that when I earlier connected the Kundo coil to the output of a cheap quartz clock circuit board (powered by a single fresh AA battery), the drive gear twitched but did not make forward progress. When I just now powered this same circuit board from my adjustable DC power supply set to 1.75VDC, its output was then sufficient to drive the Kundo coil forward, moving the second hand crisply once per second.

I believe I can power the cheap quartz clock circuit board from two AA batteries instead of one without destroying it, and then I'll have something to install in the Kundo clock base. Hopefully boosting the coil drive voltage will not destroy the Kundo coil, since the pulses have what look like <10% duty cycle.

I would be a little concerned powering it with 3V. I suggest you look for another movement that generates a longer pulse (similar to what I show). I am guessing that a high-torque movement might give you a better duty cycle.

 

[Raymond101]

I would be a little concerned powering it with 3V. I suggest you look for another movement that generates a longer pulse (similar to what I show). I am guessing that a high-torque movement might give you a better duty cycle.

I would take Wimberleytech advise. His pulse is very nice .
Also the Chinese clock use a different type of servo motor.
Don't use 2 two batteries or a higher voltage to your coil .
Or you will kill your clock coil.
You need a 50% pulse.
Don't be lazy. Now we have got it into your head and you see it works . Just build 1hz oscillator that can be trimmer to get you clock within 1 second a year.
There is plenty of space in the base of that clock ..
Do it right. Save it from landfill.

 

[debrota]

What actually is the difference between a "high torque" quartz clock movement and a standard one, when talking about the modern, square plastic beasts available from Asian factories? Don't they both use the same single AA battery? Does the high torque movement send more current through the coil, drain the battery sooner? If so, does it do so by using longer pulses, or pulses with higher peak voltages, or both?

 

[Wimberleytech]

Color me stupid. Or maybe just lazy.

It occurred to me that when I earlier connected the Kundo coil to the output of a cheap quartz clock circuit board (powered by a single fresh AA battery), the drive gear twitched but did not make forward progress. When I just now powered this same circuit board from my adjustable DC power supply set to 1.75VDC, its output was then sufficient to drive the Kundo coil forward, moving the second hand crisply once per second.

I believe I can power the cheap quartz clock circuit board from two AA batteries instead of one without destroying it, and then I'll have something to install in the Kundo clock base. Hopefully boosting the coil drive voltage will not destroy the Kundo coil, since the pulses have what look like <10% duty cycle.

I would be a little concerned powering it with 3V. I suggest you look for another movement that generates a longer pulse (similar to what I show). I am guessing that a high-torque movement might give you a better duty cycle.

What actually is the difference between a "high torque" quartz clock movement and a standard one, when talking about the modern, square plastic beasts available from Asian factories? Don't they both use the same single AA battery? Does the high torque movement send more current through the coil, drain the battery sooner? If so, does it do so by using longer pulses, or pulses with higher peak voltages, or both?

The high torque units are for driving big (long) hands. I have a quartz wall clock that has hands that are 12 to 14 inches long. It takes a stronger unit to drive the extra mass. Look on Timesavers or Clockworks or Amazon and you will see this specified.