DC Offset on ouput of transistor amp (Roland JC-120)

[mjrippe]

I have a newer Roland JC-120 UT on my bench which came in with two blown speakers.  Before connecting two other speakers to test it I measured DC offset with the amp connected to a resistive load.  One output has 20mV of DC and the other has nearly 70mV (these units have two 60 watt power amps).  Is it possible that the DC offset caused premature speaker failure?  20mV doesn't seem high to me but 70mV is a bit much.  Any ideas on how to get rid of the offset would also be welcome!

 

[joaquins]

It doesn't seems like too much, but you could trim it down in many ways.

Usual ways in opamps is to trim both 1k5 resistor, R82 and R87 if I'm reading right. In some power amps it's used a DC servo, but I think it's overkill for your case. Other option would be to add a bias to Q15 base, connecting a trimpot to C48 + side, maybe you want to limit the range putting a trim pot between two bigger resistors, maybe 1k trimpot and two 10k resistors, for a controlled range, maybe less than that is needed to cancel out your level.

Remember the DC will drift with temperature, if you want a better idea of what's happening connect a dummy load and run some signal to get a working temperature, then read the value again so you know how much is drifting, if it goes crazy (which I won't expect) there may be a problem somewhere, you could find it or add a DC servo, which would correct it all the time. In any case 70mV is pretty low I guess, shouldn't harm a speaker, it would probably have ~6Ω of DCR, so it's about 12mA and less than 1mW, hardly a problem.

JS

 

[mjrippe]

Thank you very much for your input.  I work more on tube amps with nice output transformers and no DC on the speakers so I wasn't sure how much is "too much".  One web site made it sound like more than 50mV was bad but they did not have much of an explanation why.

 

[joaquins]

I would prefer to have no DC at my speakers, would be apply with less than 1mV a servo is capable to archive easily, 70mV doesn't seems too much to be the guilty of killing a speaker, and in this case both speakers went bad so doesn't seems to be the problem there... It may not be nice, (those 70mV) but I hardly doubt it's burning speakers. Maybe there is something else besides that, it's strange both speakers went bad together.

JS

 

[PRR]

70mV is less than a milliWatt in 8 ohms and will NOT hurt 50-Watt stage-amp speakers.

That's not the problem.

The simple design will give around 30mV offset "normally". I do not know if 70mV is a fault or just a worst-case. Bring it up with 100 Ohm 10W load and see if it acts normal. If you have not done basic checks, you might find that it only passes "half the wave", that half the push-pull is blown.

Since there is NO protection circuitry, device death is very likely. Remember that one blown device will cascade and blow many other devices (as they try to pick-up the load not carried by the dead device). If you don't get them all the first time, you will blow the replacements. These repairs can be tedious.

 

[JohnRoberts]

70mV should not be enough to fry your speakers  (P=E^2/R) but is more than considered nice.

Check C48 for DC leakage, the DC gain of that amp should be 1x and 70mV is large (but not impossible) DC offset for discrete long tail pair.

Another possibility is a fault associated with Q17, and input shunt I suspect for limiting or turn on transient suppression, I can't tell because the schematic is truncated.

The DC voltage at base of Q15 and Q16 should be close. And base of Q16 should be near 0V. Actually 68k to ground at base of Q16 should generate identical voltage drop to base current of Q17 flowing through the 68k feedback resistor.

Have fun... don't make it worse probing around...  8)

JR

 

[mjrippe]

The amp sounds fine with a guitar at low volume (can't crank it in my shop) but I will check it with sine wave and o'scope tomorrow to be sure we're getting the whole waveform. 

And while I was typing Mr. Roberts chimed in with even more helpful advice!  Thank you, John.  I will check with care

Full schematic for those interested was found here: http://music-electronics-forum.com/t35754/