Help repairing Gallien Krueger 400B

[substitute]

Good morning, for sentimental reasons I picked up a Gallien Krueger 400B, it's an earlier version of the popular 400RB but it's all discreet along the lines of an old Sunn or Acoustic.  Something is wrong, while I know the amp has a really aggressive sound right now it sounds like a distortion pedal regardless of where the volume/gain is set.  Also, it's generally not as loud as it should be.  I've gotten pretty good with tubes but I don't know anything about transistors so looking at the schematic and the circuit board it's not adding up to me.

Here's what I've done so far...

- Run the preamp out into another power amp.  Sounded great.
- Bought some schematics online. (these are pretty rare)
- opened it up and noticed the power transistors have been replaced.
  The originals were PMD10k80 and PMD11k80, the replacement parts are NTE251 and 252.  So first question, are these compatible?  I've been looking over the data sheets, and seems close, but another set of eyes that's familiar with transistor specs would be greatly appreciated.
http://www.farnell.com/datasheets/1029424.pdf
http://www.nteinc.com/specs/200to299/pdf/nte251.pdf

I've attached a screen shot of the power amp schematic.  Let me know what else I can do to help you guys help me.

 

[substitute]

Things are better.  Despite not knowing much about transistors I'm pretty sure they're supposed to be soldered to the circuit board.  One of the replacement transistors was mounted but never soldered, next to it was a fried resistor.  Fixed that up, things are pretty good now, but it still won't bias right.  I don't have much faith in the work that was done to it before I got it.

 

[PRR]

> I don't know anything about transistors

And we know very little about what prior techs and you have done, and what the present state is. Can't see what you done from 550 miles away. Tell more.

> things are pretty good now, but it still won't bias right.

What "things"?

What's wrong with the bias? Too hot? Too cold? Wobbly? Votes for the wrong politicians?

MEASURE.

Basic:
voltages on power rails
voltages at D1 D3 or the "16V" points
voltages at the "1.5V" points (with polarity)
voltages on big-transistors Q12 Q13 Q14 Q15 Emitters
(we should be seeing near +60V +30V 0V 0V -30V -60V)
* If not near-Zero (<1V) at Q13 Q14 emitters and OUT jack, do NOT plug in speaker! *
Voltage *across* R44 R45

> the power transistors have been replaced.

Not a shock. 200 Watts in just four devices is pretty bold. Specific: it will hold-up forever on a test bench, but die on-stage, sooner or later.

 

[PRR]

> it still won't bias right.

I believe the bias never worked as intended. (All those unlikely 1% resistors are a bandaid for the fact the Q9 bias sensor isn't really doing much.)

If the voltage *across* R44 is less than 0.1V, and the DC voltage at the speaker is less than 1V, it's OK to rock. 0.2V across R44 is not a panic. More will cook, today or next summer.

 

[substitute]

And we know very little about what prior techs and you have done, and what the present state is. Can't see what you done from 550 miles away. Tell more.

It's working now, meaning amplifying bass without sounding like a distortion pedal, I turned it up and it gets pretty loud, but living in a row home there's a limit to what I can test at home volume wise.  However there's still more distortion than I'd like, the attack of the note sounds good but as the sustain fades it fuzzes out. 

Prior to acquiring it looks like all the transistors in the power amp have been replaced, maybe the 270r resistors around the power transistors too.  There's what looks like 1 watts in there now.  Doesn't seem to have been recapped.

What I've done, soldered Q15 (the power transistor down near the bottom of the schematic), and replaced R51 (it was burnt up).  Adjusted R1 for DC on the output jack, the lowest it would go is around 5mv.  Per the tech note on the schematic measured across Q13E and Q14E, adjusting R49 did very little, 9mv is about the highest it will go, 7mv the lowest.

I'll check the other voltage measurements as soon as I can, weekend went by too fast.  As always, thanks so much for the help PRR.

 

[CJ]

you can wrap up the speaker in some blankets to spare the neighbors,

sometimes the transistors that drive the big power transistors are strong enough to fool you into thinking that the amp is working, when really it is running off the driver transistors, so watch out for that,

 

[pucho812]

If your wondering about the NTE parts use their online cross reference


http://nte01.nteinc.com/nte/NTExRefSemiProd.nsf/$$Search

which it says

Original part PMD10K80 replace with NTE247 So looks like NTE251 and 252  Are the wrong replacements.

 

[CJ]

getting a direct coupled transistor output circuit to work can be a real pain, unless the designer knew his stuff,

especially if it has obsolete parts and you are forced to use non original replacements,

i remember suffering thru this while repairing car stereos in the back of the TV shop,

try using RCA  SK series to get a car stereo for a 1969 mercury cougar to work,

NTE and SK series, their catalogs say that one p/n can replace 10,000 different transistors, so you can imagine the odds of getting the base currents and emitter voltages matched up right on a direct coupled job, 

 

[PRR]

> DC on the output jack, the lowest it would go is around 5mv.  Per the tech note on the schematic measured across Q13E and Q14E, adjusting R49 did very little, 9mv is about the highest it will go, 7mv the lowest.

5mV DC is fine for a 200W bass speaker; 1V would be high-ish.

8mV bias will "work", and is safe. But will be a little rough-throat sounding around the 1W level. The obvious solution is: don't play a 200W stage-amp near 1W. You didn't drag a hundred pounds to under-whelm the bar's ventilation rumble.

 

[PRR]

That bias circuit reads wrong to me. Can someone check me?

There's a transistor, the trim-pot, two 1K resistors, and two more resistors. *What values??* I can't make it out on the image; I don't think values are actually given.

If this 667 ohm resistance is to make significant change in bias, it must be a fairly large part of the total resistance divider. Assume the total is 1K to 2K.

Q4 Q5 flow about 4mA (0.4V/100r). Worse: this current *rises* with temperature (we need bias to fall with temp).

We need about six times 0.6V (two triple-Darlingtons) or 3.6V total bias.

3.6V/4mA= 900 ohms. A 900 ohm resistor would be "right", except it has an all-wrong temperature coefficient. We need the resistance to be *much* higher than 900 ohms.

Q9 is a small Darlington with gain over 2,000. If it actually takes most of 4mA, the base current is less than 2uA.

The resistance divider current must be "much" higher. Say 100uA?

3.6V/100uA= 36K divider resistance.

That's a long-long way up from 1K||(2K+2K).

I think the divider values are about 50 times too small.

The bias voltage is dominated by Q4 Q5 current, which increases with temp, instead of a multiple of Q9 voltage, which decreases with temperature.

(For a fairly constant idle current in Q13 Q14, we need a bias that DEcreases when the amp gets hot. If this does not happen, the amp gets hot, idles at higher temp, gets hotter yet, and goes into "run-away".)

I'd want to do it different.

> getting a direct coupled transistor output circuit to work can be a real pain

Not just a pain. When things go wrong, it gets VERY expensive fast.

In the old days I had 100 and 10 ohm resistors at 10 or 20 Watts. I'd put these in series with the power rails. For a correct bias, the drop in the resistors was small. When things went wrong, the voltage would fall very fast, the resistors get hot instead of the transistors. However in a 200W amp you need 100W resistors to handle disaster gracefully.

For a stage-amp, I'd go crude. 5 or 6 1N4007 diodes (mounted ON the main heatsink) might give a useable bias, and no adjustment to go bad (never trust a pot wiper). However if connected wrong, the amp will burn-up at turn-on. You need some mildly expensive lessons before you try this on a 200W, especially a "beloved classic".

 

[substitute]

Pucho, thanks for the tip on the NTE stuff, I had no idea that existed.  CJ, thanks for the context regarding the topology.  Since I really like how the preamp sounds I've been thinking about ditching the power amp and installing something else.  Maybe one of these hypex class D modules.  http://www.hypex.nl/downloads/product-downloads.html

PRR, man that got deep fast... here's a close up of the bias section.  In the mean time I'll have to read that a few more times.

 

[PRR]

> here's a close up of the bias section

That's not from the same plan you posted before. See Attached.

The resistor-string is now 14.65K, in sight of my 36K suggestion.

Ass-uming that Q9 is exactly 1.200V base-emitter, this trim will allow: 3.43V min, 3.80V center, 4.27V max.

Ass-uming that the output (two triple-Darlingtons) needs six times 0.600V or 3.6V, plus two times 0.025V on the resistors, we need 3.65V. The final transistors idle at 1/100th of their maximum current so may need less than 0.600V. Different transistors have different areas (and doping) so the exact voltage needed will vary.

You are unable to reach 0.025V on the emitter resistors. It sure looks like you should have plenty. This makes me wonder if one of the other parts is still sick.

Each transistor base-emitter drop should be near 0.6V. Under 0.5V is OFF, over 0.7V won't happen except at high current, much over 1V is blown.

PNP and NPN doing the same job should have very-similar base-emitter drops. (PNP will tend to be 0.020V more than NPN.) A great difference, say between Q8 and Q10, is suspicious.

The output transistors are double-stacked so each gets less voltage. The bases of the outer devices are nominally half the supply voltage, noted as 28.7V.

 

[substitute]

Whoops!

The 2nd one is the right one.  The schematics I got have several versions of the power amp. 

I'll sit down this weekend and measure about, that's really helpful info PRR.

 

[substitute]

Since my last post I've been enjoying this amp, the output was a little hairy at power up but I'm using it in a punk band.  So maybe that just made us sound more authentic    However the amp shut down the other day in the middle of practice.  Opening it up R50 had burnt up.  Here's a list of what I've done and what the results were thus far...

-Replaced R50, started smoking immediately
-Replaced Q8, 16, 12, R50 is fine but R40 started smoking immediately
-Replaced Q13, R40 and 50 are fine but another resistor (I think R29) started smoking immediately.  It's a little hard to identify which resistor it is since at this point that area of the pcb is fairly blackened.

Any insights would be appreciated.  I'll keep replacing resistors but if anyone could help me figure out a way to keep it powered on long enough to measure some voltages that'd be a big help.


I've attached a screen shot of the schematic problem area.

and here's a link to the complete service manual.  My version of the power amp is page 7.
Thanks!
https://www.dropbox.com/s/2o17kkxgrk15w0m/G-K%20400B%20Service%20Schematics.pdf?dl=0