• Please review our updated Terms and Rules here

Kaypro II

The PSU is probably initially overloaded, but checked off load is over volting! Probably shorted by a tantalum capacitor somewhere on one of the boards. Check the power supply on some sort of dummy load on the 5V rail (e.g. a 12V 50W headlight, or halogen bulb) that should load it sufficiently to check everything.
 
If the psu is ok, You now may have to find the shorted cap. this is where a low-ohms meter, or current tracer probe comes in really handy. Failing that use a "normal" meter & take a best guess.
 
You might try the PSU with a light load on it--say, a small hard drive--but leave the motherboard off for now. I don't expect the PSU to behave differently, however.

Seeing has how this thing is nearing 30 years in age, my first suspect would be the capacitors. They may not be bulging, but they could well be dried out inside. If you have a capacitor tester, that's great, but given the age, I'd probably re-cap all of the electrolytics as a start.

Also, see this excellent guide on trouble-shooting SMPSUs.
 
To what extent are devices such as this:

http://www.peakelec.co.uk/acatalog/jz_esr70.html

good for testing capacitors? It says it can measure the ESR of a capacitor when it's in circuit and I get the impression that they think ESR is an indication of the state of a capacitor?

Im somewhat concerned that given the number of machines I have, I dont want to replace caps that don't need it or over time I'll waste a fortune.
 
Chuck, could you possibly post a copy of the schematic for the PSU board please? I've found the technical reference manual which appears to have the logic board schematics but not the power schematics.
 
I would have thought that the capacitance is more usually a good measure of the state of a capacitor!
Although I'm not sure, I would think that low ESR is down to the construction / materials. If the capacitor is getting hot in service then it's ESR (Equivalent Series Resistance) may have gone up.
Your meter will measure capacitance up to a point, a way of fairly accurately gauging large capacitors in circuit is to charge them up with a constant current supply & time them to a particular voltage Q=CV=IT, although you can get a good impression with an ordinary DMM on diode test, by gauging the time it takes to reach maximum test voltage (obviously reverse bias the rectifier diodes!).
An old fashioned moving coil meter on ohms will also give a good indication, by how quickly the pointer moves, but check polarity on the probes, it isn't neccessarily positive probe positive on ohms.
 
On large electrolytics, I've found that the easiest way to test for a dried-out electrolytic is to tap it. Dried-out electrolytics have a different "hollow" sound that good ones. Not very scientific, I know, but it's a quick, rough test.

ESR is less important in this case, as that's mostly a choice of the original designer. Capacitance and leakage are everything when troubleshooting an existing design. I'd never rely on an in-circuit tester for electrolytics. Quite often, designers parallel filter capacitors, so if you get a low value, you don't know which it is.

I believe that the Osborne 1 uses the same (or very similar Astec) PSU. So if you need a schematic, you can find it in the Osborne service manual.
 
Am I right that capacitors with a higher ripple current value are higher quality?

What does the leakage current value mean?

I can for example in the case of the 100uF 250v caps that the price varies between £0.65 and £2.58. Any tips for chosing the best choice?
 
Last edited:
I'll try to make it simple.

No capacitor (or any other real-world device) is perfect. Any capacitor will exhibit some resistance (and inductance). So, if you charge a capacitor to a voltage, you'll find that the charge will slowly dissipate because of the capacitor's internal resistance. This characteristic is known as leakage current. High leakage currents are undesirable--they cause internal heating as well as negate some of the capacitance properties. Note that we're talking about a virtual resistor in parallel with the capacitance.

However, there is also a virtual resistor in series with the capacitance. Its value is known as the ESR or "effective series resistance".

When used in power supplies, most capacitors are used to smooth out a small AC current superimposed on a DC current. e.g.:

fullWaveOut.jpg


A filter capacitor, being the beast that it is, will tend to resist changes in voltage, so it settles down to alternating charging on the peaks and discharging on the valleys so that the DC voltage settles out to something between the two. Note that this necessarily involves a flow of current in and out of the capacitor. Remember the ESR? Some of the energy from this current flow will be dissipated in the virtual series resistance. Since power P = I²R, a high ESR is not terribly desirable, as more power will be lost as heat in the capacitor, which will shorten its life. (In fact, heat is probably the single most significant factor in the demise of electrolytic capacitors). Note that ESR has no relationship to actual capacitance; it's just a side-effect of living in the real world.

As far as choosing capacitors, I stick with the Japanese brands. Nichicon, Rubycon, NCC, Panasonic, etc. I've been burned by Chinese capacitors more than once, so I generally steer clear of them.
 
Thanks for explaining that!

I've made a mess of things here. Some of the caps I've ordered in have smaller pin spacing than the ones they are replacing. Im assuming I should change them rather than using more lead and making do?
 
As long as the ratings are the same (working voltage and capacitance), don't worry about it. Things get smaller as technology advances.

Do be careful to observe polarity, however!
 
I think you might be off on the wrong track; did you test it with a load as Nige suggested in post #21?

It certainly looks like the PS might be OK, shutting down when connected because of a short somewhere on the system board and with no load cycling on and off due to overvoltage. Try it with a load of some kind, an old hard disk, light bulb, whatever, before perhaps needlessly replacing those caps.

Also measure the resistance to ground of the system board's power supply connections.
 
Having replaced the caps, I powered the PSU on without anything connected to it and C22 fried itself.

There was no sign of a chirping though.

Ok ignore that - human error.

Will try again once I've eaten.
 
Last edited:
Ok fed and happier now :)

I've rectified my error. With the PSU board not connected to anything the results are the same as they were before we started this excercise. Oh well the caps didnt fix it but its all good learning for me.

I connected the system board Gnd pin to the +12v pin and there was no connection. The same applies to the -12v pin and Gnd, however when the +5v pin and ground are measured, I show a resistance of marginally above 300 ohms.
 
Well, I did suggest testing it with a load.

5V / 300 ohms = 160 ma., well within the capabilities of the supply and probably not enough to keep the supply happy. Looking at the schematic, you should be drawing more than that. Is the 5V steady?
 
Well, I did suggest testing it with a load.
As did Nige, but with all that discussion about ESR, leakage, ripple etc. and replacing the line filters, which, although interesting, is not necessarily relevant to finding the actual problem, I didn't see where he had actually tested it with a load other than the possibly defective system board.
 
Ah at 300 ohm reading might be a red herring. I think I misunderstood you. That is the reading between the ground and 5v pins of the logic board with it not connected to anything.

So to clarrify you want the readings at the pins on the logic board when the power board is powered on and connected to the logic board? Do you want the video board disconnected whilst I do this? If so, I assume you want me to remove the edge connector from the video board, and the cable that runs from the video board to the logic board?

I don't have anything to use as a suitable load so Im heading to the store tomorrow to get some suitable bits. I have very little in the way of loose wire etc.
 
Last edited:
Do you have a place near you that carries incandescent bulbs for autos? They're great for making dummy loads. I like the so-called "festoon" bulbs.

Something like a 211 lamp will draw about an amp at 12V and about the same at 5V, which should be enough of a load.
 
Back
Top