Ancient circa-1976 0-25 VDC Power Supply Repair

[Zapped]

Hello,

This is my first post here. I've been watching many EEVBLOG episodes, and Dave's enthusiasm has rubbed off on me. His video blog has re-kindled my interest in electronics, and I am trying to learn and have some fun. To that end, I am attempting to repair a power supply from 1976. It is a home brew job, and I believe it was part of an electronics course offered by Westinghouse in the mid-1970s in Pittsburgh, PA.

The power supply looks to be a fairly low-current (1 or 2 amps max) zero to 25 volt DC unit, with adjustable voltage and current limiting. There is an analog voltmeter, a power indicator lamp, and a current limit LED which is unfortunately lit 100% of the time. It has an on/off switch, voltage, and current controls (both are multi-turn wire-wound pots) plus two banana jacks for the output.

The circuit has a transformer, a full-wave bridge rectifier made of discrete diodes, and 2300 microfarads of filtering. There appear to be voltage regulator and current limiting circuits, but that is about the limit of my circuit knowledge. The active devices include a 2N2905A, a 2N3638, a 78MGT2C which appears to be a regulator. There is also a TO-3 case pass transistor with the markings "466 068001 6621" which I have not been able to identify yet. Testing it with a diode tester shows it to be an NPN type, and the B->E and B->C junctions test OK, biasing in only one direction.

I have attempted to sketch a schematic (see attached) but I don't know if anybody besides me would be able to read it. It's my first attempt at drawing a schematic from a piece of hardware. I could also take some photos of the board if that would be helpful. I tried in vain to find a schematic for it online, but the power supply is sold old that it's probably long forgotten. There are two devices that I am calling RF chokes, but they might be something else. They have some color bands, but the final band is a blue "T" shape. They look like resistors, but have what appears to be wrapped glass case. Any help identifying these buggers would be great. They measure about 15mH on an LCR meter, but the colors on the bands don't seem to correspond with anything I can measure.

The supply puts out proper voltage, and the variable voltage control works. I know that the limit LED should not be lit all of the time. It is very likely that this supply was connected to a load beyond its limits, as it was used in an educational lab setting. I have tested all of the passive devices in-circuit, and I didn't find anything obviously wrong. I could use some help with the following: 1) identifying the pass transistor type 2) identifying the devices with the blue "T" and 3) suggestions on what to test next to get the limit circuit working. If anybody can identify the circuit and perhaps point me to a real schematic, that would be great of course.

Thanks for listening,

Zapped

[Zapped]

Attached are some photos of this power supply.

[TheRuler8510]

The supply puts out proper voltage, and the variable voltage control works. I know that the limit LED should not be lit all of the time. It is very likely that this supply was connected to a load beyond its limits, as it was used in an educational lab setting. I have tested all of the passive devices in-circuit, and I didn't find anything obviously wrong. I could use some help with the following: 1) identifying the pass transistor type 2) identifying the devices with the blue "T" and 3) suggestions on what to test next to get the limit circuit working. If anybody can identify the circuit and perhaps point me to a real schematic, that would be great of course.

I guess i don't follow...if it is putting out the "proper voltage," and there is nothing "obviously wrong," then what is the problem?

We need some more description and analysis of the symptoms if any.

Regards,

[jamesglanville]

The smoothing cap (the 2300uf 50vdc one) might be leaky/old/rubbish. Check by replacing it with one at least as high capacity and voltage rating. I may be wrong, but insufficient smoothing would cause the 2n3638 transistor to switch on and off at 50hz, making the limit led appear to be on the whole time.

[TheRuler8510]

Can you get in there and check it with a scope? Seeing the waveforms can be very helpful. Alternately--you can check with a multimeter on ACV range to check for AC ripple voltage. You can determine if the cap is bad, especially with a load on the pwr supply.

Or you can just replace the electrolyics with new ones. Chances are they are bad after 35+ years.

[jamesglanville]

Something to note is that measuring the capacitance of the caps isn't sufficient - the ESR often increases dramatically as caps age so you really need an ESR meter (or as theruler8510 says just replace them)

[tautech]

Blue banded components look like Zeners. But purple, black, silver, blue. But 7, 0, 1/100, 6 doesn't make sense to me.

Edit
If you have another PSU you could remove them and check for zener action

[Zapped]

Sorry for not being clear enough about the specific problem. I think that when the unit was functioning normally, the current limit LED did not light unless the limit (which was pre-set with the AMPS control) was exceeded. While there is no current meter, a current limit could be pre-set with the AMPS control. The LED would stay off until that current limit was exceeded by the load.

Now, the limit LED stays on continuously, even with no load at all, and the current control fully clockwise. I guess that means that it has lost its ability to limit current, or at least display when a current threshold has been exceeded. Looking at the circuit, I guess that means the 2N3638 is conducting all of the time, so I should probably look at voltages around that transistor? I guess I should also check the output current with an ammeter and a variable load to see if there is any current limiting actually happening when I change the AMPS control.

Yes, the cap is probably 40 years old now and needs replaced. I can look at the P.S. output on a scope, and I will measure the output ripple and report back. An ESR meter is high on my list of must have items!

OK on the possible zeners. That sounds right. I will try to test those as zeners and see if that's what they are.

Thanks for the input!

[Zapped]

I measured the ripple, and it was around 1.0 mV, swinging from around 0.3 to 1.6 mV max. It seemed surprisingly clean, given the age of the filter cap. On the DSO scope it didn't look a great deal noisier than an old HP 6215A supply. And yes, these measurements were made with the 20 MHz bandwidth limit ON.

I figured out the cause of the limit LED being stuck on. It turned out that the 2N2905A had shorted out. I replaced it with a 2N1132 from the junk box, and now the limit LED only lights when there is a direct short on the output. The transistor gets warm quickly when the output is shorted, and that is very likely how the original device failed -- with a short condition for too long.

The AMPS control does not seem to have any effect, and there does not appear to be any kind of current limiting happening that I can measure. I ran the supply into on different loads from 0.1 to 3 watts, drawing from 10 to 500 mA of current. For each different load, the AMPS control did not change the current limit, and the limit LED never came on, unless the output was shorted. Perhaps there is another failed component in the current limiting circuit?

Voltage regulation seems to be quite stable under different loads, so at least I've got that going for me. :-) It's still a useful little 5V-20V supply for low current applications. If anyone has any ideas on why the current limiting is not working, I'm all ears. Thanks.  --Zapped

[tautech]

Thats a good start.
Sprargue caps are excellent quality IMO. I have never replaced one.
Did you identify the blue banded device?
Zener of maybe signal diode?
The circuit will make more sense then.

[Zapped]

After some research I'm pretty sure those devices are zeners, but I forgot to try to measure their breakdown voltage. I can't tell from the markings what values they are supposed to be. Apparently there is not a single standard for zener color codes.

I also forgot to mention that the power supply delivers over 5 amps into an ammeter on the 10A range, regardless of the setting of the current limit control. There doesn't appear to be any sort of current limiting taking place. As mentioned before, the current limit LED does turn on into this essentially short circuit, now that I replaced the 2N2905A.

I'm now wondering if this supply only has over-current notification (via the LED) but not actual current limiting? Do supplies like that exist? There is still the mystery of the current limit control not having any effect, either on the limit LED threshold or the output current level.

[Kevin.D]

Going from your unconventional schematic , heres how this should work .The current control works in a opposite sense that you would normally see  ,here the small pass pnp 2n2905 is fed  by a  base voltage derived from your current control (via your brown wire) ,this voltage will be lower than it's emitter voltage at low currents ,so it will be   biased on .
 Now at higher currents through your current sense resistor (big ww resistor ) the  voltage at the emitter (2n2905)  will start to reduce until it aproaches the voltage at it's base (which should be constant) so  it then starts to turn off and so reduce voltage/current to the input of the 78mg . (The large pass transistor  (pwr to-3 ) has it's base voltage controlled by the output of the small 78mg regulator )  .


 So your fault could likely be  the control Voltage from the current control is absent or low  . Measure it (brown wire) to see it adjusts and that it is reaching base of 2n2905 ,if it's stuck at some low value then the zener (if  that is a zener marked as rfc 15mH ? then it's just being used to set current max limit) it could be open or  the current cntrl itself is faulty (open/dirty wiper) .

hope that helps

[Zapped]

Kevin, thanks for the detailed explanation, despite my terrible schematic. I just tried LT Spice and will try to learn it, using this PS circuit as a teaching aid. The transformer alone took some time to research. :-) 

You were correct, one side of the current control pot failed open. Guess I should have checked that before bothering you folks, but this is the beginner's forum. The pot is a ten-turn 100-ohm device, but I'm not sure of its power rating. It is much smaller than the voltage adjust pot. I am going to try some fixed power resistors next, in hopes of finding out if there are any other failed components before replacing the pot. Thanks again for the education.