JB Systems Power amp stuck in protect mode

[Robin]

Hello everyone,

I recently got a JB Systems B2-1500 PA Amplifier which has one channel stuck in protect mode. I am trying to repair it, but I’m out of ideas.  The right channel seems to be faulty and the left one seems to be alright.
The most common cause of protection mode is a blown power transistor/mosfet in the output stage of the amp. I checked them all by using a multimeter on the diode tester mode and they seemed to be all fine.
Both the thermostats,(on each channel of the amp), measured the same resistance when I desoldered them.
After that I made sure that the amp is receiving the right amount of voltage. I could compare the voltage of the faulty side to the voltage on the other side. The voltage on the main rails was equal on both channels. So this also shouldn’t be the problem.
Capacitors that have gone bad could also be the cause, (I know it’s a crude way to measure the ripple voltage) but I measured the AC voltage on the main power rails. The AC voltage was exacltly the same on both channels. It was approximately 2.6VAC.
The only difference I could spot between the two channels was the output voltage (AC) at the speaker terminals. There was no voltage on the working channel, but on the faulty channel there was around 4.75 VAC present on the speaker terminals.
I think this is very weird. Normally if an amp goes into protect mode, the speaker relay opens, and there is absolutely NO voltage at the speaker terminals. Could the relay be broken? Is it because of that the amplifier thinks it should go into protect mode?
I hope someone has any ideas, if you do, please let me know.
Thanks in advance!
 

[singapol]

There is no schematic or service manual online so it's us just guessing. Best is you trace the LED to  whatever
it is being triggered by.The relay could be stuck as you said the output transistors are OK.In general from what I have read of JBsystems user manuals protection LED can be triggered when powering or switching off amp.Temperature rise from overheating or DC protection of speakers. It's also odd that there is 4.5V AC on the speaker output.Please trace the signal line to where a fault may be. Also know that this is a class D switching amp so a pulse width modulation IC is  used.The amp be be microprocessor controlled or sound processor.

[gnif]

Check for DC on the faulted output before the relay disconnect, it could be going into protect due to the offset to prevent it damaging speakers. If it is a class A/B amplifier check the DC bias currents, etc.

But as singpol stated, without a schematic it is hard to give any more information. Perhaps some detailed photos of the board (front and back) will help.

[Robin]

Thanks for both of your replies!

I traced the power for the protect relay and got to a 2SA1213 transistor which got 26VDC on the base instead of 25VDC (which DOES make a difference). The base of the transistor is connected to a LM339DG through a couple of resistors.

This is a comparator which has different inputs than the same IC on the other channel. (both channels are built identically)

If I only touch one of those inputs on the working side with my multimeter, that channel goes in protection too. Those inputs come from some smd resistors and capacitors. Could they have gone bad? I think it's not likely.

The DC voltage on the outputs is only a couple 100 millivolts, so I think that's not the problem.

If you guys have other things in mind I could measure, please let me know. Today I even unscrewed all the mosfets from the heatsink to get to the other side of the board to find nothing special. 

[Robin]

Here are some pictures of the defective board

[singapol]

Found something suspicious, black carbon splash (red rectangle) what could have caused it, the blue rectangle
is it a relay spitted out from below it?

[Robin]

It really looks like it on the picture, but in real life it's clearly a 2 written with a black marker. The working board has it too, but there it's more clearly a 1.

[singapol]

Try to measure the voltage across the 0.22 ohm 5W resistors it shows the current each power transistor is biased at.

[Robin]

Thanks! I'll try that today. I know none of them are shorted though, but I'll give it a shot.

[Robin]

I just measured acorss the 0.22ohm resistors. the voltage was always around 10mV, sometimes 20mV, sometimes 5mV.

I also measured the resistance of the trimmer pots on both channels. The working channel was:
trim1 = 54ohms
trim2 = 98ohms

The defective channel was:
trim1 = 36ohms
trim2 = 209ohms

I adjusted those pots to be equal to the working channel, but that did not fix the protect problem.

I do think the transistors are constantly amplifying a smal signal. The defective channel is slightly getting warm, and there is 4.5VAC on that output, so that means that they are amplifying something. I don't know what though.

[singapol]

The voltage across 0.22 ohm should be constant DC without signal. If it varies then something is wrong. Try checking the protect circuit..transistor,diode or resistor..cold joints? I think you are on to something.

Edit: But then again I may be wrong as this is a switchmode amplifier?

[Robin]

I have been able to get a hold of some schematics of the amp. The problem seems to be situated around the LM339D smd IC on page 4 of te schematics. The IC itself seems to be fine as I switched it with the one on the working channel.

Does anyone know what I can do next?

I uploaded the area which I assume contains the faulty part. The output of the optocouplers measures exactly the same as the working channel, so no problems there. I don't know of the difference between -2.075V and  -2.133V matters that much, but from what I learned at school I think it does, but I don't know where it is coming from. The rail voltage +15 volts is the same as on the working channel, so it has to come from the bottom side of the picture I think.

[singapol]

If you have a ESR meter ,check C52,C26 10uF 50V across the optocouplers. Also check C27 2.2uF 50V.
How about the transistors and what is P2?

[Cyberdragon]

I think you mean check for leakage (which an ESR meter can). But if those are ceramic caps then they can't leak (C27 could though). Check the two diodes, and the voltages at P2, the junction of R56 and R58, and the junction of R46 and R62.

EDIT: Didn't notice the electrolytics across the isolators, I was only looking in the box.

[Shock]

As mentioned check both capacity and esr on the smd aluminum electrolytics. If the device is several years old and has seen heat or sat around a bit they could be an issue.

Fading of the black, blue, pink or whatever mark on the top of the can (when others are fine) may indicate there is electrolyte corrosion starting.

[Robin]

First of all, thank you very much for all of your responses! It's awesome to get help from all over the world 

I desoldered C27 to measure it with my multimeter on the capacitance setting (because I do not own an ESR tester).
It was a pain in the but to desolder and I almost ripped a pad off. Nevertheless I got a reading of 2.144µF for a 2.2µF cap. Could I now assume that the caps on the board are okay?

After that I measured the two center diodes (diode 34 and 35) and got a junction voltage of around 0.5V.

The voltage at P2 (Think P2 connects to another board) was 0.000VDC exactly the same as the working channel. This was also the case with the junction of R56-R58.

BUT, The voltage of the faulty channel at R45-R62 was 0.000VDC and on the working channel it was -15.02VDC which is the negative rail voltage of the control board.

My own guess is that the comarator isn't outputting a voltage, because there is an offset at the input. ( see image in my previous post) I don't know where that offset comes from, but I think that's why there is no voltage on the output and therefore transistor BG13 can't drive the protect relay.

[Cyberdragon]

Probably the LM339 at this point. If the diodes are good there's not much else that could cause it to not turn on. If the voltages to the inputs on the first two are close (they may not be identicle if the amp is leaky) then the chip is blown.

[Robin]

I understand what you are saying, but I already switched out the LM339 with the one of the other working channel. So that lets me think that the LM339 isn't faulty.

Couldn't the LM339 be so sensitive that it detects that small difference in input voltage, and therefore does not pull the output high?