Gallien Krueger 200MB just started shorting

[ggjgjhkihkjhiuyiy8788]

Hi there. A guy brought me his G-K 200MB because it was distorting on low notes (he plays an upright bass through it). Here's the schematic:

G-K200MB.pdf

So I check it over and all seems fine, although the bias voltage is a tad low (like 4mV) so I adjust it as per the manufacturer's specs (to 5mV). No problems. So I'm just checking some more voltages on the output transistors (Q11,12,16,18) when I notice the power transformer is chattering and the secondary lines from it are actually shaking (this is without a speaker load attached, BTW)! I turn the power switch off immediately and it stops (I should note that the power switch is on the secondaries in this amp- the primaries stay powered so long as the amp is plugged in!). So I measure voltage on the secondaries thinking they've shorted, but they're fine- each side putting out about 30VAC. Next I think maybe the rectifier is shot- I remove it and test all 4 diode points and they seem fine. I then thought I would be clever and only hook up one side of the secondary at a time to see if the -ve or +ve rail was to blame, but unhooking either results in the same problem. I figure I'm stressing the PT out a lot with these tests, so I'm currently using a step-down transformer to run it on 120V mains (this is an Australian 240V amp).

I scoped the unloaded secondary voltage and it seems fairly clean but when I switch the power on the waveform compresses on both sides, with flattened peaks (I'll attach pics later) and this is the same on both ends of the secondary.

None of the 5W or 2.7K 1W resistors test as open circuit and both filter caps seem to charge up when doing a resistance test. So I'm trying to figure out what would cause this and why would it start happening all of a sudden? I'm certain I didn't short anything out while I was probing and a close inspection of the board doesn't reveal anything out of place. I might replace the filter caps since they're about due anyways, but shotgunning parts never solves anything, IMO.

Any ideas on how to isolate the problem? I'm thinking of maybe disconnecting the +/-12V and +/-15V from the power supply first, since I'd only half to lift 4 resistors.

I honestly don't know why I keep accepting solid state jobs. I can't seem to get my head around them and they seem to go from bad to worse before (hopefully) they get better.

-Jaret

[ggjgjhkihkjhiuyiy8788]

Here's the trace for the unloaded PT secondary:


And the same but when the power is switched on:

[km6xz]

What do the 43 volt rails look like in value and wave form. Is the amp pulling excess current? Any problem that would cause the wires to shake are probably either pulling high current of it is a clue that does not point to any defect. Without a load on the power amp, what do you see at the output with the scope, DC? Until you have a stable amp don't connect any load to it.

[ggjgjhkihkjhiuyiy8788]

The unloaded output has about 500mV DC on it. I scoped the 43V but I can't get a photo b/c R317 + R275 (both 0.33ohm/5W) start getting very hot within seconds and I even heard a sizzle from around R275 so I shut it off immediately again. You can smell something wrong as soon as you turn it on but nothing else seems to be getting as hot as these two resistors. What I saw on the scope looked like a small AC signal but I'll have to double check to get the freq and test it with the meter to get a voltage reading. I noticed that the power LED is not lighting up when the amp is on, either.

[ggjgjhkihkjhiuyiy8788]

Ok, did some more checks. Running the amp on 1/2 mains voltage, the 43V rails are both about 4.5V DC and the 1/2 rail has 1.5V DC, which seems okay. The caps also seem okay because you can see the 43V rail voltage slowly drop when you switch the amp off.

Here's the AC + DC scope on the -43V rail (5V/div & 2ms/div):

Looks to be about -5VDC and a clean 4VAC PTP. Should there be AC here or is the rectifier leaking? This is without a test signal through the amp, mind you.

Next, here's the AC + DC scope on the output (2V/div & 2ms/div):

Looks like a bit less than -1VDC and around 3.5VAC.

[ggjgjhkihkjhiuyiy8788]

I checked the output transistors and I think Q11 is shorted. Measuring resistance of base or collector to the emitter shows 0.3ohms vs. 262 b/w C+B. Q12 has 1.1ohms to E vs. 262 b/w C+B (Q11 and Q12 have their C's + B's connected). A diode test on these gives 0mV to E and 140mV b/w C+B for both Q11 and Q12. I'm not sure how comparable these measurements are between NPN's and PNP's, but measuring Q21 and Q22 on the opposite rail gave 10ohms to the emitter, 560ohms from base to collector, 6mV to emitter and 272mV B to C for a diode test function- much higher than the other side!

Given that R275 is getting sizzling hot within seconds of switch on and it's connected to Q11's emitter, I'm thinking this adds up to a shorted transistor, right?

Is my reasoning wrong (again)?

-Jaret

[g1]

Sounds right, the output device is probably shorted. Check them out of circuit to be sure. If Q11 is bad you should also replace Q12. If one of the lower pair is bad they should both be replaced. Make sure the drivers and their 100R resistors are ok. (Q9 and what looks like Q18 or 19, can't read it)
Are you sure you couldn't have shorted something while probing? Are the outputs encased in plastic or do they have a metal back? If metal, a bad insulator could have allowed the metal case to short to the heatsink due to pressure from the probes.

[ggjgjhkihkjhiuyiy8788]

Hey, thanks for backing up my diagnosis! I'll pull Q11 and Q12 and test them again and replace the pair if necessary. I didn't test the 100R's, but I did test Q9 and Q19 and I think they were fine (I'll double check). The lower pair test fine, I think, but maybe I'll pull them and test them out of circuit to be safe too.

I've learned about shorting transistor leads with the probes the hard way in the past, so I'm extra careful now and I don't think that caused this. The output insulators could have failed, but I didn't measure any zero/low resistance from their leads to the heatsink and I think they're plastic cased anyways. I suppose probing a loose connection could've caused it (?), but all the joints look good and I retouched a few just to be safe with no effect on the problem.

I'll let you know how it goes. Thanks for your help!

[ggjgjhkihkjhiuyiy8788]

Okay, time to dredge this one back up again, unfortunately.

So, I replaced both output pairs on the GK-200MB and the 0.33R's just to be safe. I fired it back up, and the same problem- dancing transformer secondaries. At this point, I realized when I previously tried subbing out the bridge rectifier, I re-installed the original 90 degrees to what it should be. After kicking myself in the ass, I installed it properly, turned the amp on again and everything was great. Adjusted the bias to get just under 5mV across the 0.33R, checked the waveform on the output at different volumes- all good and clean. Played through it a bit and sounded tight and had the customer play through it and he said it sounded as good as the day he bought it.

Here's the kicker- he takes it to a gig the following weekend, after playing it only a bit at home. He's on stage, jamming away, and the amp starts smoking and a flame shoots out the power switch on the back. He brings the amp back to me, rather upset (rightly so, IMO).

So I open it up to find the plug connecting the PT to the board is cooked around one pin and the traces leading from it are burnt away. I've cleaned the charcoal off the board to see what needs what, and the damage is definitely focused between the power switch, the rectifier and the PT plug. I initially thought the insulation had failed on one of the output transistors I replaced, but nothing tests as shorted there. What IS testing bad is the rectifier. One half is reading a dead short across both diodes. What I'm wondering is, was the rectifier failing what caused the fire, or was the fire what caused the rectifier to fail? I have this theory that when I installed the rectifier incorrectly it weakened it to the point it later failed. Plausible? Or should I be looking elsewhere? I don't want to just slap in a new rectifier in case the same thing happens again.

In the meantime, I'm going to keep testing components in the power supply to see if I can find any other culprits.

[Steve Conner]

I have this theory that when I installed the rectifier incorrectly it weakened it to the point it later failed

Seems very likely. I personally would not have sent it out the door with the same rectifier bridge. Even applying 120V would have been enough to severely overload it.

This is a case where a light bulb limiter would have earned its keep many times over.

[ggjgjhkihkjhiuyiy8788]

True enough, but you live and learn and I've certainly learned a painful lesson from this one. Thanks for the response, Steve.

Interestingly, the schematic also shows the rectifier in the wrong orientation, with the secondaries connected to the '+' and '-' outputs, rather than the corners. The layout and PCB outlines definitely show it in the conventional orientation, with the notch to the +43V rail. This is not why I installed it incorrectly to begin with, it's just an interesting coincidence.

[Jazz P Bass]

Subliminal.
You saw it & without "thinking", installed it that way.
Scarey!

[ggjgjhkihkjhiuyiy8788]

I started tallying up the cooked parts and I noticed that one of the 1N4002's from the -ve rail to the output looks like it got pretty hot too. Although it's testing okay out of circuit I'm going to replace it and its twin on the +ve rail. At this point, I noticed the output transistors were testing weird so I pulled them too- 3 out of 4 are testing shorted across all three terminals (diode check & resistance). Only Q22 was safe.

This doesn't bode well for my theory that the rectifier caused the fire-breathing current draw. If one leg of the rectifier shorted (and it did- one of the rear -ve diodes is testing shorted out of circuit), wouldn't the the current just take the path of least resistance and circulate from the PT to the rectifier to ground to the PT centre tap? I'm wondering how the output transistors could've been smoked in the process, unless the problem started elsewhere and then caused the rectifier to fail. As I mentioned, I had replaced all 4 outputs and ensured the bias was within limits. None of the three legs of any output transistor tests as being shorted to the heatsink.

What gives?

[ggjgjhkihkjhiuyiy8788]

<bump>

[riz]

Back in my time as a bass player, I remember GKs sounding great, but I shot fire out of at least two of them myself, not counting all the ones I heard about from others. Gee, that probably doesn't help, does it?! You may have just gotten really unlucky with the timing, I guess is what I'm trying to say...