Peavey Mark IV: dc on the speaker and preout jacks

[lespaul924]

Hello

I have an old peavey mark IV bass amp that initially was putting 50v dc out on the speaker. After opening it up the -15v rail was down and I found and replaced a bad resistor that got it back up. No more dc on the output and checking the main +-52/+-15/+-16 rails On power and pre and limiter sections all seems well.

However; playing a guitar through it results in pretty terrible distortion and weak output both preamp out into another poweramp and also poweramp in. further investigating shows that with a signal generator putting in around a volt of 440hz results in 35v dc on the speaker out and little on the pre out jack. 1volt of 1k hz puts out 109 volts dc on the speaker jack and around 12 on the pre out. All with the pre and volume pots cranked up. lowering the volume lowers the dc.

Any tips on the best course of action?

I've replaced every opamp in the pre as they were in sockets, and my first guess to the culprit is the -15v rail, but I'm just not finding any obvious failed parts or voltage rail faults. Only when a signal is present am I getting the DC on the output.

Thanks, this place is a great resource
Paul

[Gunny]

From the symptoms you described my first advice is NEVER connect a speaker to this amp until the DC is eliminated from the output...or poof! speaker will be destroyed.
Second, DC on the main output tells me that at least one power transistor is shorted. When that happens you usually find one or more of the smaller transistors associated with it are also gone.
I'd identify the power semiconductors that are bad, unsolder them and check for DC at the output. You'll also have to study the schematic and satisfy yourself how far the damage has gone.

[52 Bill]

Welcome to the board.

...a volt of 440hz results in 35v dc on the speaker out and little on the pre out jack. 1volt of 1k hz puts out 109 volts dc on the speaker jack...

As Gunny pointed out, that much dc voltage on any speaker will toast the voice coil, but I don't know how you can get that much voltage out of a power supply that is + and - 52v DC.

If both pre and power amp output is distorted, your diagnosis of power supply is probably a good one. What happens to the power supply rails when you drive the amp with a signal?

[Enzo]

With 52v rails, there is no way to get 100v on the speaker output.

Isolate the problem. The rear panel is the powr amp and power supply. FLop the panel down and disconnect the cable to the front panel - the one with +/-15 and the signal. Yours is a 400BH rear panel, right? That power amp will run alone. Verify no DC on the output, then if OKO, connect a speaker. Now with your finger touch the end pin where the preamp cable was plugged on. SHould make hum out the speaker. If the power amp does that, then clip a test signal to that input pin and ground. Does that come out OK?

Demount the preamp - the front panel - and while not powered, go down the board with your meter on diode test, and check EVERY little glass diode for shorted. MAny of them are in pairs. Offhand I see about 30 of them. SOunds like a lot, but really, they take seconds apiece, so it won;t take but a few minutes. Look at the schematic, most of these paired diodes are protecting jacks - either inoput or output jacks. WHat they do is clamp the line serving the jack from going outside the supply rail voltages.

Replace any shorted ones with plain old 1N4148 or similar.

Now remove the speaker load, reconnect the preamp board to the power amp board and fire it all up. Check at each and every jack at its tip contact. Is ther DC at any? Ought not to be.

[icefloe01]

With 52v rails, there is no way to get 100v on the speaker output.

I haven't seen the schematic so I have a question; is that a true +-52VDC as in is there a +52V in respect to ground and a -52V in respect to ground or is there only a +52V and ground? If there is both a plus and minus that could easily give him 100VDC if it's a PNP/NPN output and at least one in each half is shorted.

Eric.

[lespaul924]

Thanks for all of the replies

yes, it's a 400bh module, here is scheme with a trace of where the excessive dc is present.The red trace was 75volts DC with a 1k sine wave that was putting around 35v AC on the output.
the proper +or- 52 volts is on the collectors on top and base/emitters on bottom.The area circled in purple I was getting different voltages. at the leg of Q8 and r45 was -27volts. The circled R68 was the bad resistor.

I have done the diode test on a couple, but not every, power trans on each side and they tested fine. Also, Q2 Q12 Q1 Q7 Q8Q9 and Q10 were fine with the diode test. CR12 and the other diodes tested well. I will go back and test every power transistor. Can only one bad part be this catastrophic?

I know that dc on a speaker is bad, but in my experience it's always been a bad voltage rail or blown output transistor(s) and was always present despite operation. I've never encountered this situation before, where there was only dc on the output with the presence of a signal. In theory isn't something acting as a rectifier? Does it have to do with that this is not an npn pnp output ,but a "quasi-complimentary?"

Enzo, I did isolate the boards a few days ago and I have checked most of the clamping diodes, but I will go back and systematically check every one. when the boards were seperated and touching the pin there was hum in the speaker; however, at the time I wasn't aware that dc was only present with a signal. I will go back and check tomorrow. With both boards connected just putting a thumb on a jack shows around 128v or more on the output.

Thanks again for the responses
feel free to ask any questions or recommend places/parts to check/recheck.

[Enzo]

Here is a pdf of the 400BH. I find it a lot easier to work with.

This amp uses +/-52v rails. No point in the circuit is more than 52v from ground. That is the highest voltage the amp can produce.

If there is both a plus and minus that could easily give him 100VDC if it's a PNP/NPN output and at least one in each half is shorted

Eric, would you like a moment to reconsider that answer? If a transistor shorted the output to one or the other 50v rail, there would be 50v on the output. If both sides shorted, the +50 and -50 would then be connected together, and the power fuse would blow HARD. Even if the fuses somehow maintained, +50 and -50 don;t add up to 100.

There just isn't any way to get 75v or 128v out of this amp. WHat kind of meter are you using, and where are you connecting your ground probe from the meter? I would be conecting my black probe to chassis.

[lespaul924]

I'll have a chance to recheck everything tomorrow. I have a new battery, but I will bring another meter and check with that also. I'm using a higher quality RCA meter, but have a good fluke to use also. black probe is on the chassis and i was getting expected readings on the power rails. Will try again though. Thanks again

[icefloe01]

Eric, would you like a moment to reconsider that answer? If a transistor shorted the output to one or the other 50v rail, there would be 50v on the output. If both sides shorted, the +50 and -50 would then be connected together, and the power fuse would blow HARD. Even if the fuses somehow maintained, +50 and -50 don;t add up to 100.

Yes, if the +52 and the -52 short together. But, suppose that the red/yellow wire from the transformer has lifted from ground and also the main board and output jacks. If the output jack's sleeve gets tied to the -52V by some means and the tip to the +52V, -52V becomes the new ground reference on the output jacks giving you 104VDC on the output. But this will only happen if the grounds get lifted. Another explanation is maybe the meter is on the DC millivolt scale and it's not immediately obvious that it's set to that...

Eric.

[Enzo]

Sorry, I thought you were the one who proposed the transistors were shorted.

But I'll listen, how might all these odd connections occur?

[lespaul924]

thanks for all the comments,

trying with a different meter showed no dc on the output.

Kind of odd that the original meter I was using makes accurate readings except on this one amps output. It consistantly shows anywhere from 0-140v dc depending on the strength of the signal and I definately have it grounded right and good contact with the probes. Any ideas on what would cause a meter to give false readings under only these conditions? Have been using it for 4/5 years and this is it's first major error.

feels good to have the amp up and running and not chasing anymore ghosts

Thanks again for all the comments

[icefloe01]

Sorry, I thought you were the one who proposed the transistors were shorted.

I was, but I prefaced that by saying I hadn't seen the schematic so I didn't know what the output structure was, although the PNP/NPN idea was definitely incorrect.

But I'll listen, how might all these odd connections occur?

In looking at the board layout in the pdf you posted, I can see that the -52V side of R48 is almost surrounded by ground. A solder blob or loose piece of wire there would easily provide the altered ground reference. That should blow the fuse provided the transformer center tap is connected. If it's not then it possibly won't blow the fuse. Now the whole bottom bank of transistors is rendered useless and distorts the signal horribly. The +52V is the tricky part especially since he said the amount of DC varies with the volume setting... I'll have to ponder that some more. But I'm not saying this is neccesarily what is definitely happening, but in order to get over 50 volts DC out, the ground reference somehow has to change. I also say they need to double check the meter and make sure it's on volts and not millivolts. Especially since they came back with a reading of 128V...

[Enzo]

Is the first meter a really cheap one? It is probably reading AC volts as DC somehow. I used to have a meter that didn;t block DC from its AC scales. Often got very confused readings.

[lespaul924]

The faulty meter is a radio shack, but one of the bigger ones. has capacitance and transistor sockets, not that i really use those features. It's from the early 90's and this is the first time it hasn't kept up with my fluke or benchtop unit. Good to know though.

But on a different issue, the original low distorted tone is probably a fried speaker. Once I was sure there was no dc on the output i tried it through a cab and it sounded fine, but it was not the first cab i tested it through. that speaker could very well have met the 52v. I'll have to go check it out.

It's just a no name thing in a small pa cab, but is there a practical do it yourself fix, or are you better off getting a new speaker?
and for future reference what's the voltage threshold and amount of time it takes to nuke a 100-200watt speaker? my friend said he saw a spark when he pulled the jack out. the sign of a killing blow?

Thanks again

[Enzo]

You can see sparks if you pull a speaker cord out while playing. All the spark means is that some voltage was present. COuld be unwanted DC, could be perfectly normal signal voltage.

If your meter has those features, then it should be sophisticated enough not to be fooled. How is the battery inside it?

what's the voltage threshold and amount of time it takes to nuke a 100-200watt speaker?

REpair the speaker? if the loudspeaker inside the enclosure is screwed up, replace the loudspeaker. It is not probably worth reconing. If the enclosure cab is OK, we put a new loudspeaker into it.

How long is a piece of string?