Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

July 18, 2024, 12:28:29 PM

Login with username, password and session length

Recent Posts

 

Repairing a Gibson G-105 Guitar Amp

Started by Timko, May 15, 2024, 09:46:49 PM

Previous topic - Next topic

Timko

Hello!  I've been a pedal builder for the past 10 years, building both from PCBs in the DIY community as well as designing some of my own.  Last winter, I began to dabble in amps.  I first took on this Gibson G-105 from the 70s. I first worked at re-capping the thing.  In a poor decision by me, I used some cheap caps to replace the two 3000uF caps located in the rectifier section of the amp.  It sounded great for a while.  It was super clean unless you pushed it, and was phenomenal hooked up to an electric piano or guitar.

One night, the amp began to totally fuzz out, producing this gated, starved sound.  Sort of like when you start screwing with the voltage going into old Germanium fuzz pedals. Luckily, the amp has a signal line out right before the long tail pair phase inverter, and I validated that it sounded fine.  I did some sanity checking around the speakers to ensure that wasn't the issue. I changed the crappy caps in the rectifier to some nice Mallory screw in terminal caps.  I even changed out the 4 power transistors. 

I've gotten a little over my head in terms of what step to take next.  I'm pretty capable of testing, but I'm unsure of where to look next.  I'm hoping the people in this forum could help point me in a good direction to begin to look.  Thanks!

Schematic for reference:  https://el34world.com/charts/Schematics/files/Gibson/Gibson_g_105.pdf

DrGonz78

#1
List voltage reading for Q8 & Q9 at  all pin's connection points. Measure all power supplies. What are your -/+40v rails? -/+12v rails? Can you scope a signal through the unit and figure out where the signal goes bezerk?

Edit: If you scope the output is the fuzz actually crossover distortion?
"A person who never made a mistake never tried anything new." -Albert Einstein

Timko

Quote from: DrGonz78 on May 15, 2024, 11:01:19 PMList voltage reading for Q8 & Q9 at  all pin's connection points. Measure all power supplies. What are your -/+40v rails? -/+12v rails? Can you scope a signal through the unit and figure out where the signal goes bezerk?

Edit: If you scope the output is the fuzz actually crossover distortion?

Thanks for the reply, and sorry for my belated response.  I had someone at my company walk out, which pretty much threw my last two weeks into a tailspin. I checked the power rails, and both have -/+40v and -/+12v I expect.  As for the long tail pair, here are the readings I got from them with a 1kHz signal at 5v:

Q8
C -0.704
B -0.127
E 38.83

Q9
C -0.713
B -0.137
E 39.35

I'll admit my oscilloscope skills are lacking in terms to how to best use it to figure out  where the signal goes off of the rails.  I did put a scope at the end of the chain as suggested, and I can see that the bottom part of the sine wave appears to be clipping as it appears to be pretty flat. I'm currently looking into how to measure crossover distortion via watching some youtube videos, but any suggestions you can provide to an amp noob would be greatly appreciated.

g1

5V 1Khz signal is way too much.  50mV would be a better level.
But for DC static voltages you do not want to be applying any signal at all.
Recheck DC at Q8 and Q0 with no signal applied.  And check the pinout, it looks like you have the emitter and collector readings reversed.

Timko

Thanks g1, that was helpful.  The pinout shape on my schematic and what the redistor looks like are a bit different (still the correct value though), and you're indeed right that the pinout was backwards.

So with no signal, this is what I got:

Q8
C 39.26
B -0.128
E -0.713

Q9
C 39.79
B -0.149
E -0.709

g1

Those seem good.  Can you do the same checks (DC volts, no signal) for e,b,c of Q13 and Q14.

Timko

I captured some readings for Q10-Q18

Q10
C 1.175
B 39.94
E 39.42

Q11
C 0.941
B 0.095
E 0.091

Q12
C 0.012
B 0.088
E 0.088

Q13
C 39.72
B 1.097
E 0.561

Q14
C -39.74
B 0.270
E 0.082

Q15
C 39.83
B 0.565
E 0.80

Q16
C 0.082
B -39.89
E -39.84

Q17
C 39.83
B 0.565
E 0.97

Q18
C 0.082
B -39.89
E -39.80

g1

Q14 doesn't seem to be turning on, probably resulting in cross-over distortion which Dr.Gonz suggested.
Check through the components (and connections) on the path shown in red here:


You cannot view this attachment.

DrGonz78

Q11 & Q12 caught my eye too but understanding each transistors role in a circuit is difficult. I don't know what Q11 & Q12 actually do but it never stops me from guessing. Are they current limiting the circuit in case of overload? Not sure if they are pre drivers for the outputs or what. I'm curious to know how bad of a guesser I am.  :o
"A person who never made a mistake never tried anything new." -Albert Einstein

Timko

#9
I did some testing on the part of the circuit g1 suggested - I noticed that the diode D6 tests in both directions at a very low value of -0.051v.  However, that may be due to the Thermistor in parallel.  Both D4 and D5 have a value of 0.51 (around a 1n4001) and don't go both directions.

The Thermistor in parallel with D6 is connected off of the circuit board to the chassis. Is there a way to tell whether or not that is functioning correctly?  On the bench with no power it reads around 47 ohms, but with power, it reads 0.

g1

That sounds normal for the thermistor.  Make sure neither of it's legs are touching the heatsink.

Quote from: DrGonz78 on June 06, 2024, 01:20:00 PMI don't know what Q11 & Q12 actually do but it never stops me from guessing. Are they current limiting the circuit in case of overload?
Correct

Timko

Can I replace D6 with a 1n4001 or other 1n diode?  I assume that section of diodes is just keeping the power rails separated?  I took D6 out off of the board, and it kind of crumbled when I tried to test it. 

g1

That string of parts marked in red sets the bias for the drivers and in turn, the output transistors.
Was D6 the same type as D4 and D5?  Black plastic like 1N4001 rectifier diode, or glass type like 1N4148 signal diode?
Replace with same type it was (signal or rectifier).
While you are there I would resolder all connections in that bias string.

Timko

I did some tracing around the parts mentioned and noticed that the connection for R53 was loose.  However, there wasn't any change in the issue.

g1

Measure DC voltages (no signal) at points A through G.

You cannot view this attachment.