power supply problem - advice appreciated

[larry lurex]

i wonder if anyone on here can give me advice or a pointer where to find more information (maybe another board) on the following problem:

i've had a studiomaster series 4 mixing desk since the late 80s and it's served me well over the years. there is still nothing wrong with it but its power supply is giving me grief.

the power supply  has  four circuits shielded by four fuses. fuses 1+2  (which i think are power sources for the display and amplification) blow. fuses 3+4 (don't know what 3 is there for, 4 is phantom power) don't blow.

when i connect the power supply to the mixer phantom power is on but no display and no sound.

i have checked the circuit and components but couldn't find anything suspicious. i have also tried to find someone locally who could have a look. the only person slightly interested was a tv repairman who 'code checked (?)' the components and couldn't find anything either. mind you he only had a quick look.

i don't wanna spend big money on a repair - it wouldn't be worth it but it would be shame to dump a great mixing desk which works perfectly.

i have uploaded a picture of the power supply unit:
http://www.hammerparkhomes.co.uk/images/powersupply.jpg

the circuit diagram can be found here:
http://www.hammerparkhomes.co.uk/images/circuitdiagram.jpg

[R.G.]

Since the power supply picture and schematic don't show fuses, I'm assuming that the fuses are after the power supply.

The outputs of the power supply look like they're all out of three terminal regulators. Those things self limit the available current to about 1A for the 78xx and 79xx devices and about 1.5A for the LM317 stuff.

If it were mine, I'd measure the outputs of the power suppy all by itself, disconnected from the rest of the circuit. If the voltages look OK, I would surmise that the problem is not in the power supply, but in the circuit itself.

You don't mention fuse values. I'm assuming they're under 1.5A or the 3-terminal regulators wouldn't let them blow.

Tracing down a dead short is semi-difficult. It holds the voltage to 0, so you can't see where the current is going. You can do tricks like putting an AC current through the line and using a tape head to follow the current carrying conductors, but that's relatively a pain.

Again, if it were mine, I'd replace every electrolytic cap in it. If there are a zillion, I'd first replace the ones that are directly from power supplies to ground, and there should be a bunch of those. Then I would temporarily replace the popping fuses with a 15 or 22 ohm power resistor each so I didn't have to buy so many fuses, and then start trying to find what is causing the short. As I said, it's likely to be an electro cap. Look for signs of fluid leakage around them. The nice thing about replacing electros in a unit this old is that it will function better with new ones anyway, so it's not wasted effort.

Then check the amplification stages for shorted outputs.

When you get past that and it's still shorted, either do a tape head current follower, or get ready to start cutting traces to isolate what frees up the power line from the short.

[12afael]

the board for pro audio is www.groupdiy.org

[niftydog]

a tv repairman who 'code checked (?)' the components and couldn't find anything

in other words, he didn't really do anything. God only knows what he meant by "code check" but that doesn't sound like troubleshooting to me.

Anyways...

R.G. has excellent advice as always. I would add this; try to reduce the load just enough so that the fuses stay intact. Then, look for sources of heat. I would expect the regulators to be fairly hot, but if you find a cap or a resistor that's too hot to touch, that may give you an indication of something going awry down the line. A hot component is not necessarily a broken one though, it could be getting hot due to something else failing. It's just sometimes a good indicator of something not right.

Look for bulging tops on electrolytic capacitors, or smelly brown stuff leaking out of same. You could, if you can get an osilloscope, look for ripple on the supply lines. Perhaps it's as simple as a filter cap on the way out (C3, C4 & C8.)

[larry lurex]

Since the power supply picture and schematic don't show fuses, I'm assuming that the fuses are after the power supply.

f1-f4 are clearly shown on the schematic. if you look at the pic - right in the foreground you can see a hole. it's where f1 should be.

The outputs of the power supply look like they're all out of three terminal regulators. Those things self limit the available current to about 1A for the 78xx and 79xx devices and about 1.5A for the LM317 stuff.

If it were mine, I'd measure the outputs of the power suppy all by itself, disconnected from the rest of the circuit. If the voltages look OK, I would surmise that the problem is not in the power supply, but in the circuit itself.

i measured the incoming current for the working circuit (f3-f4) - was ok when circuit was switched on. the other circuit - nothing. but then again - the circuit is broken and that makes sense.

You don't mention fuse values. I'm assuming they're under 1.5A or the 3-terminal regulators wouldn't let them blow.

the fuse values are shown on the diagram. i should mention that i put t1l fuses in all four slots.

Tracing down a dead short is semi-difficult. It holds the voltage to 0, so you can't see where the current is going. You can do tricks like putting an AC current through the line and using a tape head to follow the current carrying conductors, but that's relatively a pain.

Again, if it were mine, I'd replace every electrolytic cap in it. If there are a zillion, I'd first replace the ones that are directly from power supplies to ground, and there should be a bunch of those. Then I would temporarily replace the popping fuses with a 15 or 22 ohm power resistor each so I didn't have to buy so many fuses, and then start trying to find what is causing the short. As I said, it's likely to be an electro cap. Look for signs of fluid leakage around them. The nice thing about replacing electros in a unit this old is that it will function better with new ones anyway, so it's not wasted effort.

Then check the amplification stages for shorted outputs.

When you get past that and it's still shorted, either do a tape head current follower, or get ready to start cutting traces to isolate what frees up the power line from the short.

wow, there's a lot of advice there and i don't get half of it ...

replacing electrolytic caps - what's your advice here ? i'm sure you don't mean using a soldering iron.
replacing the popping fuses with a 15 or 22 ohm power resistor - i've used about ten fuses and runnin low. will it not ruin components to do this ? the fuses blow cos there's a short somewhere and the fuses protect the components. right ?

thanks for takin the time.

[larry lurex]

... I would add this; try to reduce the load just enough so that the fuses stay intact. Then, look for sources of heat. ...

i wouldn't know how to reduce the load - the unit is not connected to anything and fuses blow as soon as i connect to mains.

[R.G.]

f1-f4 are clearly shown on the schematic.

...ooops... sorry, I was looking for the two dots and a squiggle symbol and missed them. They are indeed there.

Interestingly, F1 and F2 are part of the same circuit, the rectifiers on the +/- 15V circuit. That puts a slightly different spin on it.

i wouldn't know how to reduce the load - the unit is not connected to anything and fuses blow as soon as i connect to mains.

And that's the a major hint! It wasn't clear from your description that the power unit was already disconnected and that the unit blows fuses all by itself. I thought that the unit was blowing fuses when connected to the stuff it powers.

That's a much easier fix - fewer things that can be the problem. My first step was to try the power supply outside the unit, and you're already there. The problem is inside the power supply itself, and it further is isolated to the +/-15V circuit.

Here's what to do:
1. Get yourself some of those 15 to 22 ohm power resistors I spoke of and put them in place of the fuses.

replacing the popping fuses with a 15 or 22 ohm power resistor - i've used about ten fuses and runnin low. will it not ruin components to do this ? the fuses blow cos there's a short somewhere and the fuses protect the components. right ?

It will not ruin components. Fuses are to limit current one time. Resistors limit current all the time. Fuses drop essentially no voltage in doing their limiting, resistors drop a LOT of voltage. Using a resistor to find a short is better because the resistor won't just pop. But you can't use the resistor for normal operation. Even better than a resistor is a variable resistor, like a filament light bulb. This has a low resistance at low currents and high resistance at high currents when it heats up. But you have to select the ight bulb, and we don't need that complication right now.
2. By inspection of the schematic, the problem is one of
- foreign matter short on the PCB somewhere
- Shorted capacitor C1, C3, C4, or C8
- Shorted voltage regulator, LM317T or 7915
- Shorted diode bridge KBPC802
- other random, highly unlikely stuff; check the above first.
3. You're going to have to unhook components nondistructively to find this one. Get yourself some desoldering braid, and make sure your soldering iron is well tinned.
4. Hook up your voltmeter with the + lead on the Cathode (bar) of D2 and the - lead on the anode of D4, or to circuit points that are accessible that are connected to them. Set your meter to DC volts, probably a 100 or 200V scale.
5. Turn the unit on. Read whatever voltage you see on your meter and remember it. It's likely 0V, but could just be quite small.
6. Turn the unit off. Desolder and remove C3. Turn the unit back on. Did the voltage come up? If so, C3 is likely bad. Turn the unit off, and then use your ohmmeter to measure C3's resistance with the + meter lead on the C3 + lead and the - meter lead on the C3 - lead.  If its shorted, it reads a low ohms. If it's OK, it first reads low ohms, then rises as the capacitor charges. If C3 is shorted, you've found your problem.
7. If that wasn't it, do the same to C4.
8. If that wasn't it, remove the LM317. If the voltage comes up, the LM317 is bad.
9. If that wasn't it, remove the 7915. If the voltage comes up, the 7915 is bad.
10. If that wasn't it, put your voltmeter leads at the rectifier bridge end of the resistors in the fuse positions, set for 100V AC. Remove the diode bridge. Do you now read about 36Vac? If so, the diode bridge is bad.
11. If that didn't do it, it's one of the weird other things and we can find those, but I'd bet that you have it.

Now, you have two things. You have the knowledge of what was busted, and a baggie full of parts that you removed. You have to put the parts back in and you have to replace the broken item.

You could just buy one replacement part and put the old ones back in. Or, for only slightly more money, you could put all new parts back in. IMHO parts are cheap compared to the interruption in my life and time having a piece of equipment down. If it were mine, I'd buy a whole suite of replacement parts and put in all new ones, re-upping it to like-new reliability. I'd also replace the filter caps and other parts in the still-functioning section, but I'm a bit of a nut about things like this.

It's entirely up to you.

Obviously, this outlook suggests another path of repair: don't bother debugging anything, just rebuild it. That works too, but it teaches you nothing about debugging.

This last bit is brought to you by the Keen School of Radical Electronics 8-)

[larry lurex]

wow, thanks very much for that. i see where you're comin from and it all makes all lot of sense.

what i did this afternoon is look at the regulators (tr1-tr4).

i don't have a fancy multimeter that beeps so i switched to diode mode and measured the legs of each transistor.

i started from left leg (1) to right (3). the first digit is the black lead, the second is the red one. these are the readings:

tr1/3
12-120
13-1400
21-120
23-1300
31-800
32-675

tr4
12-120
13-1 (no reaction)
21-120
23-1 (no reaction)
31-800
32-650

tr2
12-980
13-1980
21-1 (no reaction)
23-1 (no reaction)
31-1 (no reaction)
32-699

my conclusions:
tr1 seems ok since it gives the same readings as tr3 which regulates a working circuit.

tr4 gives similar readings (apart from a couple of combinations which don't trigger any reaction ) and also regulates a working circuit.

tr2 is quite different to the others. does this look like a malfunction ? if so, i'll remove it and see if the fuses still blow. i just don't wanna start messin with the board if you think it's not lookin suspicious.

i apologise for my simplistic view of the world of circuit boards but my knowlegde is limited. especially compared to yours.

[R.G.]

That was a good insight, measuring a working and a non-working device and then trying to spot the odd one.

However, in this case, it's not particularly valid. None of these are transistors, they're integrated circuits in transistor packages. Tr1 and 3 are LM317T voltage regulators, TR4 is a ??TL783?? which is a high voltage type for the phantom power. Tr2 is yet a third type, a 7915 negative regulator. The fact that they're different types could easily cause the difference in readings. The comparison between Tr1 and 3 is valid, though, so it points to TR1 being OK.

I can appreciate being conservative about messing with the board. However, right now, it's useful only as a somewhat bit odd bit of decoration or mini-sculpture - it's already dead. At a guess, the most likely parts to be bad, in order are:
1. C3 and/or C4
2. Diode bridge
3. Tr2
4. TR1

That order is based on (1) known slow failure of electrolytics (2) amount of voltage/current stress.

If you're after minimal disturbance, pull off C3 and C4, then see if the fuses still blow. If they do, it's most likely the diode bridge.

The only subtle way to be sure you have the correct part without desoldering and removing parts is to use a current tracer, which is the device I mentioned with a tape head pickup. These will let you follow a current path with the probe until you find someplace where the current is going that it shouldn't. Even these have their problems with capacitors, which are supposed to act like AC shorts. A Hall effect magnetic probe and a DC current might trace DC, but that's even more expensive.

If you are worried about damaging the board, find someone to help you.

[larry lurex]

update and quick question:

i have remove tr2 (which i suspected to be faulty) and d4. this has changed nothing. fuses are toast as soon as i plug in.

i've now got only three fuses left and goin to the electronics shop to get resistors would take me hours.

however i've got lots of potentiometers. pots i have: 1k, 100k, 220k, 500k and 1m ohm.

could i use any of those as variable resistors ? if so, which one should i use ?

[larry lurex]

i've run out of fuses so i'll just post what i've done:

removed tr2 - f1 blows

removed d4 - f1 blows

removed c3 - f1+f2 ok

put tr2 back in - f1+f2 blow

put d4 back in - f1+f2 are ok separately but blow badly when both fuses are in

[R.G.]

Here's my advice:

Go buy a new set of semiconductor  and electrolytic capacitor components for everything from F1 and F2 to the pin 1, 2, and 3 outputs. Replace them all.

TR2 (7915) is clearly bad. D4 may be. But the reliability of the diode bridge, the filter caps, and other parts has been compromised by the electrical abuse. You might get by with only a diode bridge, TR2, and D4, but since you're already there, why not up it to full reliability?

Three terminal regulators like the 7915 and LM317T are under US$0.50 each, diodes are under a dime, and the diode bridge is under US$1.00. New filter caps might be a couple of dollars each, but you're going to get out of this for under $20.00 if you just replace them all.

[larry lurex]

ok, i'll replace the ds, the cs, the trs in that circuit. i won't replace the resistors, though. alright ?

i don't mind the cost, i just hope i can get the parts. the only place here in the uk is maplin. i checked online this morning and they didn't have all of them. i'll travel to their store tomorrow and see what i can get.

i'm not sure what the kbpc802 one is but i assume it's the big one with four legs. the part left to d9 would appear to be a resistor they forgot to number. right ?

and finally: what about using pots instead of resistors in place of fuses ? if it's not a good idea exactly which resistors should i get - you said 15 to 22 ohm ...

many thanks.

[gez]

the only place here in the uk is maplin

Didn't follow the parts list you need, but these suppliers are worth checking out:

www.rapidelectronics.co.uk
www.esr.co.uk

Even if they don't have what you need (and I'd be surprised if they didn't), they're cheaper than, and have a better range than, Maplin, so it's worth knowing about them.  If you use ESR just phone them (very helpful bloke runs it).  With both companies delivery is usually the next day.

[R.G.]

i'm not sure what the kbpc802 one is but i assume it's the big one with four legs.

It's an integrated diode bridge. You **could** replace it with four 1N4004s properly soldered into the four holes in the PCB if you can't get a replacement.

You don't need to replace the resistors - just use an ohmmeter and verify that they are what the schematic says they are.

the part left to d9 would appear to be a resistor they forgot to number. right ?

Yes, that's what it looks like. It's not likely to be what's killing your fuses, anyway.

and finally: what about using pots instead of resistors in place of fuses ? if it's not a good idea exactly which resistors should i get - you said 15 to 22 ohm ...

The resistors are a temporary measure, only for debugging. I picked low-ohms values so they could pass some significant current. Using a pot for that would likely burn out the wiper contact on the sliding surface of the  resistive part of the pot. If you are rebuilding, don't bother with them.

When you rebuild:
(1) remove all the semiconductors and electro caps.
(2) Put in the diode bridge. Fire up the power supply and measure for DC at the + and - terminals of the diode bridge. You should see a DC level of the correct polarity.
(3) shut it down, put in C3, fire it back up, and verify about 25Vdc across C3.
(4) shut it donw, put in C4, fire it up, and verify about 25V across C4
(5) shut it down, put in the parts in the negative side (T2 and associated), fire it up, and verify -15V output.
(6) shut it down, put in the parts for the positive side, and verify +15V
(7) have a beer  

[larry lurex]

thanks gez and again thanks rg. i'll start with step 7 and see how i get on.

[larry lurex]

ok, here's an update.

i finally managed to get hold of all the parts required this week. i rebuilt the circuit the way recommended on here.

put in diode bridge, got correct readings. put in c3, measured, c4, measured again. spot on so far. then i moved to the next stage and put in parts on the negative side, fuses blew. took them out again and put in parts on the positive side, fuses blew.

it's all cost me too much time and effort (not so much money) so i've decided not to follow the problem up anymore.

what i have got is a separate power supply which provides me with the required 15v i need for the circuit in question.

however, i'm confused yet again.

according to the schematic, pin 1 of the connector to the mixer should be 0v, pin 2 -15v, pin 3 +15v.

i connected the output of the power supply and all i get is pin 2 -15v. the other two are 0v. what am i missing here ? any ideas gratefully received ?

[NaBo]

Wow R.G. ... this has given me a lot to think about with my current hopelessly messed up project... my talkbox/head amp.  Thanks for the indirect help yet again

Debugging power supplies/amps is certainly a whole other beast compared to 9V stompbox circuits, especially when there's an apparent short somewhere in there... cooking a 30 dollar xformer would not be fun.  If you can give me any other tips specific to my situation in that thread, that would be great too.  I think im about ready to stop swearing at the thing and just giving it even more "electrical abuse", all cuz it was "supposed" to be working.  :wink:

Sorry to interject here Larry  :oops:, i wish you the best of luck fixing your problem!