What's causing my old-school Galien-Krueger GMT 300B to drop volume when i pick hard?

[AndyDufresne]

Hello everyone, I have a Gallien-Krueger GMT 300B from the '70s. The problem is that with lower volume and gentle picking the amp sounds great. However, when you strum hard and increase the input signal the amp kind of fades and the volume drops in proportionately with the strumming.

When I run the "effects send" to the "effects return" I get the same problem. If i put a signal into the "effects return" the amp works fine, so I'm assuming that it's a problem in the preamp. When I run the "effects send" to another amp it gives me the same dropping volume effect.

Thanks.

[Enzo]

Without me looking it up, does that one have a compressor? And can it be turned off?

[AndyDufresne]

It has switches for "Contour" "Current Lim." "Bright" and "Inp. Atten.", but there is nothing labeled compression.

[J M Fahey]

The "Current Lim." might be a limiter or Compressor.
And yes, old FET compressors (the kind used way back then) usually had such problems.
That's why most manufacturers stopped using them.
SWR continued for many years, had the exact same problem, and most bass players preferred to switch them off and cope with the buzzy clipping.

[AndyDufresne]

I'm not familiar with how a FET compressor would work. Is it just a transistor that it will only allow a set amount of signal through?

The switch has 10, 5, and 2 as settings. To me it looks like the "10" setting is off, because the terminals on the switch have nothing connected to them on the "10" side. Would these numbers represent amperes? So it is limiting the current to either 10, 5 or 2 amps?

Any advice on how I could test the circuit? Any advice on how I could disable the circuit? I've found it hard to locate any schematics.

[J M Fahey]

Email them stating Model and Serial Number and politely ask for the schematic.
They usually send it.
Post it here, so we all talk about the same.
Then we continue.

[teemuk]

There is no built-in compressor/limiter in those amps.

The current limit feature is an user adjustable VI limiter for the power transistors and therefore naturally in the power amp section.

There are lot's of switches and jacks in the signal path (e.g. "input attenuator", "boost/distortion") that could have corroded along the years.

I'm more inclined to believe it's a failing component like a capacitor, JFET or a generic bipolar transistor.

I have a faint memory I posted some links to archive of old GMT/G&K schematics into the schematic requests section. There should be three ungodly large .pdf files titled "Old Products" from which you'll also find schematics for 300B model. As said, the files are simply way too large to attach here. Search for the link.

[AndyDufresne]

I cleaned the all the pots, switches and jacks with electronic cleaner.

Here are the schematics that GK sent me.

300B Bass Preamp 60002.pdf300B-300G Early Power Amp 60000.pdf

[J M Fahey]

Yes, no compressors there.
And the Curr Lim *is* a variable current limiter.
Not clear what it was used for.
It allows signal clipping in the power stage at less than maximum power, not too interesting in an SS amp.
Oh well.

[AndyDufresne]

Any advice on which components or circuits I should check? Any theories on what could cause this?

[J M Fahey]

Yes.
Rather than check "for bad components" , of which there may be none, we must make a functional test.
It's a 6 gain stage preamp.
Some stages provide raw gain, some provide frequency sensitive gain, meaning equalization, but all have the same core structure: a single stage, common emitter transistor (Bipolar or FET), with approppriate biasing resistors and a coupling capacitor in series with the input.
All of them are capable of rectifying the audio gignal and blocking, or at least losing gain .
It will be difficult to happen on the first two, Q1 and Q2, because they have 220K resistors in series with their gates (and Q1 is DC coupled) plus they have a very "strong" biasing system which is not easy to disturb.
But from Q3 on, it may happen.
I suggest you scope Q3/4/5/6 collectors and play your bass, at stage levels.
I'm sure one of them (or more) will show an important DC voltage shift in sync with your playing.
You might also see signal clip, assymetrically, and with a varying clipping point or duty cycle.
Such unpredictable behaviour made me switch to Op Amps already in the early 70's, when most anybody else was still going discrete.
Of course, do not use either the Distortion nor the Boost, which will make the problem worse.
I also think that your bass signal must be hotter than what GK imagined way back then.

PS: what about the Contour switch?
On/Off, does it affect the volume "pumping"?

[AndyDufresne]

Unfortunately, I don't have a scope. Is it feasible that just going ahead and replacing the Q3/4/5/6 transistors will give me a good chance at solving the problem?

Are the coupling capacitors just there to block DC?

I'm guessing that if one of the biasing resistors was bad then the transistor would just be constantly open or shorted and wouldn't cause this kind of effect. Is that correct?

Also, the schematics note that these transistors should be MPSA18, but on my unit they are TIS97. The TIS97 seem harder to locate than the MPSA18.

[J M Fahey]

No.
As I said, I don't think there are "bad" parts there, specially the transistors.
Don't change parts needlessly.
You will solve nothing, and probably cause real problems.
If you don't have a scope, sorry, this is one of the cases where it's needed.
*Even* if you have one, this basically confirms what you hear, but knowing where and how much it happens may lead to some suggestion ... which may work or not.
This amp is not broken, is "as-is", an old design.
You didn't answer what effect, if any, had the controls on the problem.
If possible, post some gut picture, would love to see the internal chassis space distribution, whether the preamp is on a separate PCB, etc.
The coupling capaitors are there to block dc, correct.
No biasing bad, or the amp would sound funky all the time, not just on signal peaks.
And I still suspect a bass signal hotter than what was common in the early 70's, plus today's bass players pop, tap, slap, do all kinds of percussive and peaky stuff on bass, while in that age most (as in 99%) just played smooth bass lines.
Try to find which control settings trigger that problem.
Good luck.

[AndyDufresne]

Here are some pics. I'm seeing if I can get a hold of a scope.

[J M Fahey]

Hi Andy, thanks a lot.
INCREDIBLY well built amp.
Parts of the best quality.
I had asked for pictures to see the internal space, "modern" GK are ultra-compact and do not allow for Mods, other than altering some part value.
Yours has a lot of space and, more important, separate boards so you *may* build, say, a modern GK preamp or whatever you want and stick it there .
Only, I don't think it's worth it.
I see a killer amp now, with a minor inconvenience.
Maybe by scoping we find what happens where and maybe, by adding a diode or something we can compensate for it , but otherwise I wouldn't touch it.
Remember Led Zeppelin'd Bass player, after using his wall of Acoustic 361 from '69 to '77 (their "Golden Age") , then switched to GK600B (same as yours but higher power) for 3 more years.
Not a bad endorsement !!
John Paul Jones » Gallien-Krueger GMT 600B AMplifier

[52 Bill]

Do you have a voltmeter? Have you tested the 35 volt power supplies to the preamp? Try monitoring the preamp power supplies when the signal drops out. Does the supply voltage drop with the signal?

[AndyDufresne]

That's good news. I originally bought it because our bass player needed a head and we couldn't afford a working one. I bought this one for $30 and was hoping that it was going to be a more obvious fix, like a bad solder joint or a big leaky cap.

I really like the sound I kind of wish it was the G300 with the reverb. It would make a good head to keep at the house. I really don't mind the troubleshooting though because I'm learning a lot. I love to fix things anyways.

Could you give me an explanation of how the diode could help? The diode would block current in one direction right? So you think that current might be flowing the wrong way somewhere? Are you thinking that the base lead on one of the transistors might need a diode in front of it?

[J M Fahey]

Don't hurry up.
I'm just thinking aloud, and, of course, I can imagine a couple things which can happen there, but can't suggest a medicine unless I'm certain about the sickness.
Example: short ago I went to see the Doctor for a numb pain in my left leg.
He examined me , then wrote a list of tests he wants done: various body fluid chemical analysis, XRays, ElectroMyography, renal and liver tests, the works.
I asked : "Dr, what may it be" he answered "bring the test results and we'll talk"
Repeat dialogue 3 times.
The 4th time he gave me a 600 page book covering about 2500 or 3000 illnesses: "it's *probably* there ... you want me to be more specific? ... there's the door ... go away and come back with those tests."

And so far my (presumed but not confirmed) diagnostic is "signal rectification shifting the bias point" ... but I need your results.

[AndyDufresne]

I do have a voltmeter. But I was wondering if you could help me a little with the schematic? Tell me if I have this correct.

So, there is a rail that runs throughout the preamp that is labeled +36 V. After this voltage is reduced by resistors to 14, 17 or 18 volts it goes to the collectors of all the transistors.

Voltage on the base of these transistors controls the flow of current through the transistors and out the emitters. The emitters are all connected to chassis ground.

This is where I get a little confused. I thought that the transistors acted like a gate of sorts. So, there is gain coming out of the emitter? Then shouldn't there be a load between the emitter and ground?

For Q1 and Q2 the emitter goes to -36 V, while Q3,4,5,6 all go to chassis ground. I'm not sure why.

[AndyDufresne]

The contour switch does not seem to affect the problem. Neither do any of the other switches. It appears that the problem is linked to the distortion. Initially, the problem did not surface. But after a few minutes of playing the problem showed up again.

I had not noticed before, but when the volume pumping occurs the sound is only distorted. When the volume problem is briefly not there, the sound is only clean. The distortion pot does not change the tone in either circumstance.

The volume pumping issue can be resolved by turning down the volume on the guitar, If the signal input is low enough then the problem won't occur.

I tried to test the +35v rail, but I got zero volts, so I obviously didn't get it right.

[J M Fahey]

Basically, bipolar transistors (Q3 to Q6) need their base around 650mV above the emitter voltage.
Some (biasing) resistors are used to provide that.
R21/23 , R30/31/33 , R34/35/39 , R45/43.
So far, so good.
Since the audio is either on a collector with around 15VDC or in a pot or jack with 0VDC (because it's referred to ground) you can't connect it direct to the base which in this case rests at 650mV DC above ground (because the emitter is at 0V) so you need an element which blocks DC but passes AC (audio), which is a capacitor: C10/26 plus a lot more in the other transistors, because they not only block DC but also make complex networks so they have special frequency response curves (active filters).
But what most forget or ignore is that *every* transistor base is part of a diode (base: anode ; emitter: cathode) and this, if fed with a *strong* audio signal will rectify it and charge said capacitor with its peak voltage , which will absolutely disrupt its biasing and make it distort, lose gain, whatever .
Ring a bell?
That's why I said that there are not bad parts there, simply it's an old design.
You need a scope to *see* what's happening on every collector when you play loud and then, *maybe*, I can suggest a cure ... or not.
Otherwise, you will hear what happens, at the end of a chain (through the speaker), but will not know which is the weak link, nor what's exactly happening.
And, of course, do not use the crappy "Distortion" nor the "boost" available there, it will onlñy make the problem worse.
My doubt is: *without* using them, does the problem still happen?
At what volume and tone settings does it appear?
More precisely: with Lo and Hi mid on 5 (on a 0/10 scale), bass and treble on 10 , Dist and Boost on 0 , at what volume setting does the problem appear?
Is the amp clean or already distorting/crunching?
Please answer these doubts.
Thanks.
PS: that's why GK switched to Op Amps, they do not have this problem.
PS2: you may build a new preamp, if you dare, and mount it inside that amp ... but the sound will be quite different.
Or use some Bass pedalboard or processor as a preamp and plug it straight into the Power Amp in.

[teemuk]

That's why I said that there are not bad parts there, simply it's an old design.

J M, I think you are missing some pretty important clues in this:

Initially, the problem did not surface. But after a few minutes of playing the problem showed up again.

I had not noticed before, but when the volume pumping occurs the sound is only distorted. When the volume problem is briefly not there, the sound is only clean. The distortion pot does not change the tone in either circumstance.

- The problem seems to appear only occasionally, sometimes the amp sounds fine, then it drifts into misbehaving again
- The tone becomes distorted when the volume begins to pump
- It is not something related to the built-in distortion feature
- It is triggered by high amplitude input signals

To me all this sounds exactly like a part going bad. I find it hard to believe they would have let a design pass if it did this in "stock form". With "part" I mean everything from plain components (caps, resistors, semiconductors, etc.) to parts like solder joints; it could be an intermittency issue. If the volume pumping keeps appearing frequently then locating the stages where it originates shouldn't take long with proper tools.

We can discuss days and days about this but only thing that gets your amp troubleshot and fixed is getting into the action: pop up the hood and start finding out what causes the issue. Don't have a scope? Built an audio probe, loan one, rig up one of those "soundcard scopes", anything. You NEED tools to troubleshoot and fix the amp.

[J M Fahey]

That's the point I'm trying to stress from day 1.
My "no bad parts" recommendation is meant to try to stop our friend Andy from butchering that PCB by changing parts at random, with no clue.
As you also agree, this is one case where a scope is needed.
That's why they sit on troubleshooting benches, not only because they look cool
About one year ago my trusty old Kikusui scope died ; of course on a Saturday afternoon and me, as always, with a lot of backñogged work pending.
I downloaded a PCScope, the kind that uses the PC card as input.
Worked acceptably well, only audio frequencies of course , no problem with that.
In fact I wrote something about it, including two attenuators , a 5 step one and a fixed one, because they only accept a few hundred mV signals . I guess it was in SSGuitar.
The *BIG* problem in this case is that software PC Scopes do not read DC.
And here I want to clip the probe to different collectors, from Q3 to Q6, play loud, and not only watch the distorting waveform but the DC level dancing wildly.
I'm sure it happens at least in one of them and is the cause of the perceived sound problem.
And once I find that, I'll clip that probe to that same transistor base and also watch for rectification and bias shifting.
And only then, decide on what to do, *after* seeing and confirming the problem.
Shotgunning an amp (changing parts at random hoping to find the bad one) often causes *new* problems.
Yes, maybe some part drifted value or some cap became lossy , making the preamp more sensitive to higher level signals.
It's possible.
But, which one?
In a nutshell, this problem deserves being scoped.

EDIT: download
Winscope
As they insist, you *need* an attenuator or you will kill your Soundcard input.
USE THE SOUNDCARD LINE INPUT, NOT THE MIC INPUT
The full attenuator I suggest is:

or you can use a fixed one, a simplified version:
.022uF>220k>output to mini plug>2K2>Ground.
Don't forget the protection diodes across the output leads. You can use 1N4002 or 1N4148.
They will clip any signal above 700mV peak (500mV RMS), that´s why my attenuator is calculated to supply signals around 200mV to the PCCard Line input.
Within its limitations, it works and has saved me more than once.
You expect to find some stage (transistor collector) where you see signal distort in a funky way.
Unfortunately this PCScope will not show DC drifting in response to the audio signal, which is *the* telltale symptom.
Oh well. Better than nothing.

[kscience]

The real issue may be the cheesy connectors used- with 18 awg single conductor wire. The weight of the wire when vibrated will move the connector; the connections to the pc board can be fractured. Also the connectors can be oxidized; either way this can cause issues you speak of.

That said- you really should replace the electrolytic caps in the old beast, as they dry out over 40 years. Especially if left unused for any length of time.

[AndyDufresne]

Hello, I finally have access to an oscilloscope and some free time. I know it's been a few months but I was hoping someone was still up to help me out. How should I test the transistors? What should I be looking for? Thanks.