Hi all, Please be gentle as electronics and I are far from good friends I have acquired a Kundo battery/mechanical clock that I would like to repair. It is a latter model clock around 1960 vintage. It appears the coil is not functioning and from what little I could find on the internet, it seems the TF65 transistor is most likely the problem as they are prone to fail. Can someone tell me how to test the TF65 (I have a multi meter) and what I should replace it with if it is faulty? [plain]http://i10.photobucket.com/albums/a142/leeinv66/KundoATO.jpg[/plain]
The KundO circuit diagrams were provided by Burkhard Rasch and posted in a series of threads of text and diagrams all in the German language which is not difficult for the technically inclined and a translation dictionary. A search using Kundo in Electric Clocks will reveal the thread. Or for starters, here's a URL of a PDF file of the diagram. https://mb.nawcc.org/attachment.php?attachmentid=4139&d=1184154917 The transistor is a small germanium (not silicon) device that fits into the electromagnet spool along with the shunt resistor. Transistors with the matching form factor and of Germanium junction are difficult to locate. Too, with the diagram, you should "ohm-out" the two common windings. A break in the fine wire is problematic. This due to the scramble-winding plus environmental heat-cool stress on the fine wire. IMHO, these KundO battery clocks are very disappointing when found. Either have come to end-of-life electronically or are badly disfigured owing to dry cell chemical corrosion. Any maintenance parts except the suspension spring have to come from salvage.
Thank you for the information Les! Can you please explain why the transistor cannot be modern silicon type? If the size and shape is a problem I don't see why something couldn't be wired remotely from the coil.
The replacement transistor of choice is one with a Germanium junction because these have a lower turn-on voltage than the Silicon junction transistors. The 1.5 dry cell voltage source is the limiting factor. Some repairers have re-worked these KundO battery clocks with an external circuit using a two Silicon transistor circuit. Additionally, certain factory versions of the KundO battery clocks were manufactured with a one-winding solenoid and an external two transistor circuit.
Thanks Les, I can get my head around the different turn-on characteristics of the two types of transistors.
Peter Germanium devices usually leak from emitter to base. With your meter, one lead on base, other on emitter should give you from 300 ohms - 1k ohms one way, 10k - infinity the other way. Put leads on base and collector, readings will be the same. From emitter to collector, either way should be more than 10K. Both PNP and NPN germanium devices are still about (in UK anyway) but have a look at this first.
Thanks Mike! Now I should be able to test the transistor. Just to make sure I have all this electrickery stuff right. PNP = positive/negative/positive. One positive is the emitter and the other is the collector. And finally, negative = base. Have I got that bit sorted?
You've got it, Peter. That's why, if you look at the symbols for PNP & NPN, the emitter and base symbols appear as a diode symbol, and EB and BC tests with a meter test like diodes. So the "P" bit is like a diode cathode and "N" like its anode; current will flow from +ve (P) to -ve (N).
Thanks Mike! Now I know where I am at. The coils are ok and the TF65 is toast. Now to see if I can track down a replacement down here.
Hi It isn't so much the voltage of the battery since it isn't what turns on the transistor. It is the tickler winding in the coil doesn't have enough turns. Also, note that most cheap analog meters reverse the positive and negative leads when used for ohms. Meters like Simson 260's have the correct polarity as do all digital meters. Red is positive and black is negative. If not using one of these, check with a diode to assure the correct polarity. Tinker Dwight
True, Dwight, but as well as that, there has to be sufficient current after you add turns, and also it will have thinner wire because of the available space. Many analogue meters do that, not necessarily cheap ones; all the Avos and Taylors did and they sure weren't cheap! Just my opinion, but I'd say it's completely irrelevant; if you are testing a PN junction, as long as you get a much higher reading in one direction, it's OK. You will never get a PN junction that reverses itself. At least, I've never come across one; it'll either be OK, O/C or S/C.
I thought I would post an update on repairing this clock. The TF65 transistor proved impossible to find down here and the only online source I could find had a minimum order of 50 bucks. I decided I didn't want to spend that much without knowing for sure it would get the clock running. Any way, I ended up buying another clock that was working but in poorer condition than this one. And now the clock looks great and is running perfectly. Now being fond of experimenting, I couldn't just have a non working clock lying around and put my mind to finding an alternative fix for the original movement. The pictures and short video that follow show how you can make one of these work with a six dollar quartz pendulum movement, without doing anything that effects or changes the originality of the clock. It's not pretty, but it works View attachment 4870 View attachment 4871 View attachment 4872 View attachment 4873 View attachment 4874 View attachment 4875 Click on picture to watch it work View attachment 4876
I could have posted you a transistor, Peter, but too late now. Though if the original one was OK and the fault elsewhere, then maybe not. Another member is sending me a similar non-working clock to sort, just for interest.
Hi Mike I guess I reference meters to either a Triplett or Simson 360. Most digitals ( I think all ) don't reverse the leads though. When trying to determine a NPN from a PNP or electrolytic capacitor, it does help to know which it does. It seems that the movement from the quartz clock could have been made to drive the coil in the original with very little work. I would guess the original and the quartz one were quite similar. Tinker Dwight
You could be right Tinker, I had the same thought. But, being far from competent with electronics, I went with the parts I already knew worked.
Peter So, you're using the quartz to actually drive the coil? I'm interested. Shame you're on the opposite side of my planet! Dwight I've only heard of those meters but never seen them here; I have a military spec Avo 8 that I got as a prezzie when I left my first job in 1967 to move up north, and as it's been calibrated a few times, trust it implicitly. Lead reversal is my usual transistor test method. There's another Avo 8 that lives in the workshop outside, plus about five cheap £5 digitals that actually have a transistor/diode test function. If I need to go as far as hfE testing, I have a Taylor instrument from the 1950's when transistors cost a small fortune then. Apologies to moderators for wandering a bit O/T, but electronics is becoming nearer and nearer to horology; the former was my working life, the latter my hobby.
Hi I believe the quartz had a separate circuit to swing a pendulum that is independant from the quartz time keeping. It is just to make the pendulum effect. I've not traced one down but think it is similar to the original Kundo. Next time I see one with the pendulum, at a garage sale, I'll have to rip one apart and see if it makes a good drive for one of these clocks. I don't see why it wouldn't. Tinker Dwight
Tinker, Timesavers sells a unit that just swings a pendulum, meant to wrap around your typical square quartz movement. Takes a separate AA battery. Part # 23552. Not sure how it would keep in sync with a Kundo pendulum, but maybe it doesn't matter.
Hi Harold It looks like it works on the same principle. There is a magnet on the end of the hanger and a coil in the base. I would think it is basically the same. Dwight
wow I like that I think you are on to something here not sure if this will work on my lantern as the coil and movement are mounted on clear plexiglass. I think i have one of those quartz movements around here somewhere.I might give this a try...............thanks tom
hello mike did I understand correct that the original tf65 transistors are avalable in your country? if so are they costly and could you help us out over here ? I like peter's idea and I will try it if all else fails............thanks TOM
You have it reversed...N is the "cathode", and P is the "anode". To forward bias a PNP base-emitter junction, the negative lead goes to the BASE lead, and the positive lead goes to the EMITTER lead. A transistor must be forward biased to turn on. Something else you need to remember...and it is bloody important!...the current through a semiconductor junction MUST BE LIMITED!!! The coils have enough DC resistance to limit the current through the transistor junctions to a safe value, but if you connect the battery voltage DIRECTLY across the transistor junctions, it will blow them out! As an example: if the (-) terminal of the 1.5 volt battery is connected directly to the base lead of a PNP transistor (whether silicon OR germanium) AND if the (+) terminal of the same 1.5 volt battery is connected directly to the emitter lead, the transistor is toast!!! In electronics tech lingo, this is called "excessive forward bias", and the result is called a "catastropic failure". Most digital multimeters have a "diode" function built in. This may be used to safely test the transistor junctions.
Using the circuit from a quartz clock to drive the coils is an interesting thought- but a quartz clock ticks at a one-second rate, and I though I read somewhere that these Kundo clock tick at a THREE-SECOND rate? These two frequencies are incompatible.
Yeah, I figured that may be a problem, considering on these pendulum drivers, it doesn't matter what the beat per minute is, as the pendulum is only for decoration. However, since all that is needed is a pulse to maintain pendulum oscillation, I do wonder if one of these might be able to be tuned in somehow to the beat needed for a Kundo? Would an extra, unneeded two pulses be a problem, as long as there is a pulse there when needed? Anyone tried this since last summer?
And I might add...stay away from using the Rx1 or Rx10 scales on the multimeter. These scales will very likely allow too much current to flow through the transistor junction(s), causing damage to the transistor. Use the Rx100 scale.
Hi The Pendulum would have to be really close to the right speed to synch up with a quartz one second pulse. The "Q" of the pendulum would need to be poor if it was even a little bit off. On the unit I have, the pendulum is driven, independant of the quartz movement and just uses a sense coil similar to how the early transistorized movements were. As the magnet swings by, it gets a kick from the transistor. If it has a tuneable pendulum ( not the fixed plastic it has ), one could adjust the frequncy of the pendulum to match close enough to a quartz pulse to synch up. The fact that there may be more pulses than used shouldn't be a problem if the pendulum was not in position to use the coil and magnet. Tinker Dwight
I've repaired several of these clocks by substituting a silicon plastic transistor in place of the original germanium and using a pair of AA batteries instead of a single cell. The batteries do not last as long, but then so what? The clock is working again without too much messing around and there is no leakage current draining the battery as with the original germanium rubbish.
