Crate GX-212 power amp puzzle

[Steve Conner]

The final straw is that some amp designs are actually unstable at reduced supply voltage. The classic problem is that the output sticks to one of the rails, and as the supply voltage is raised, eventually it snaps to zero. Testing one of these on a light bulb limiter, you would think it was completely hosed, but in actual use it works fine apart from a large turn-on thump.

I've never seen one that is unstable in the common mode, turning both sets of output transistors on too hard at reduced supply voltage. But when a new designer wants to prove to the boss that he's hot stuff, anything is possible.

In this particular case, the diode generates a forward bias to turn the output transistors on. The op-amps have a gain of roughly 8, so 8 diode drops worth of bias are applied to the output transistors. But being Darlingtons they only need a total of 4 diode drops.

The extra 4 diode drops worth of bias are cancelled out by AP1 and its fixed resistor buddies. But the cancellation is dependent on the +/-16V rail voltage being correct. I think at reduced supply voltage there could be an excess of positive bias.

If this turns out to be correct, I guess I'll send my consulting fee straight to Crate.

[J M Fahey]

If this turns out to be correct, I guess I'll send my consulting fee straight to Crate.

You should

And only half joking here.

[DGstrat]

Bill,

I have been rechecking all along, going back to the start, but clearly I'm missing something.
All the advice has certainly helped straighten out my understanding of this circuit.
Time for another fresh pass through.

thanks,
Dave

[Steve Conner]

Like I said, in the light of all the careful measurements and tests we have done, I think it is just a bad design.

The bias current increases if the 16V rails fall, and when running on a lamp limiter, the 16V rails fall if the bias current increases, so you have a positive feedback loop that makes it hang up with the limiter bulb lit brightly.

You may have to lose the light bulb limiter and just whack it with full mains voltage. If I were you I would try to turn AP1 all the way in the direction of minimum idle current first. I think that is minimum resistance.

Or power it up on the bulb limiter with the bias diode shorted out, and then unshort it. If I am right, the bulb should stay out and the bias current shouldn't be excessive.

[DGstrat]

Steve, you hit the nail on the head!

Turns out my light bulb limiter was a false friend in this case... I had something like a 60w bulb in there. Went to a 100 and all was well. No DC on the output and no excessive load on the output transistors.

Feeling pretty dumb about now...

Thank you to all for all the help. Hope to return the favors sometime.

Dave

[Lee_ranaldo]

I have a similar issue with a crate GX-212. The power amp section is similar to the GTX 212 unit. i'll attach an schematic that will be used for the references of this post

The design is slightly different from the old GX-212 design with the NE5532 op-amp driving the power stage.

I've replaced most of the components that were burn or shorted. including diodes (D13) and burnt resistors (R63)

I've measured the voltage across R72 - R74 and it gives me around 5,5 mv which is under the 7 mv specified in the schematic .I kept this under the "standard"
value until i verify that the power stage is stable and it has no issues.

The issue is that R61 & R64 seem to dissipate more heat than the other 2 emiter resistors. The voltage across these are average 5 mv too
the collectors on each power transistors show around +/- 32-33 volts DC being 35V the unloaded voltage i get from the power supply.

If i measure the voltage referenced to ground the 2 TIP142 transistors have around 155 mv on each emiter
and the 2 TIP147 have around 80-88 mv on each emiter

The DC voltage i can measure on the speaker output is around -11,3 mv

also the voltage measured on the base of each pair of transistors is 1,31 votls for the TIP142 and -1.26 volts for the TIP147.

R61 and R64 are ok and not open resistors. in fact they are brand new.

The amplifier plays fine with plenty of bass but this amp had lots of power amp failures.

Any suggestions?

Thanks in advance

[J M Fahey]

The issue is that R61 & R64 seem to dissipate more heat than the other 2 emiter resistors. The voltage across these are average 5 mv too

Why do you say so?
Voltage across/current through them is the same, also ohmic value and rated dissipation

[Jazz P Bass]

TP12 is too high. (TIP147 base voltage)
The chart states : -0.06Vdc.
Your reading is -1.26.

Have you changed the bias transistor Q9?
Inspect the bias pot as it may be dodgy.

You have not explained the failure in any detail, but I find that it helps to highlight the failed components to track the failure path.
This may provide clues to what may or may not have gotten stressed.

[Lee_ranaldo]

yea this amp just cooked transistors like nothing. I replaced them all and set the bias according to the schematic and totally toasted them again after a while.

I replaced the following componentes

Q8,Q10,Q11 and power transistors
D13,D21
R63
C58,C29 (100uf instead of 220uf according to the schematic).

Everything else has been checked and they are within spec except R73 that was around 230 Ohms. I replaced it and still get a high base voltage on both TIP147's
compared to what is shown on the schematic. I was measuring the power amp without a load and no signal being fed into it. I'll put a signal and probe it with my scope and see how that goes.

I know it seems weird that R61 and R64 are slightly warmer than the other 2 emiter resistors since they have around the same voltaje measured across them.

Thanks guys for your replies. i'll keep this updated.

[Jazz P Bass]

About the only thing that can fry both sides (short of massive abuse) is if they both turn on at the same time.

[Lee_ranaldo]

i'll replace the transistors on the differential pair and see what happens. also i plugged a load and still get -1,24 volts but goes down to 0.95 when i insert the signal through the insert jack

at 3 mv average on R72 and R74. I Check the scope and i see a perfect sine wave on the output. with no artifacts.

Also the bases reach +/- 32 volts ac peak to peak on their bases

[ogeecheeman]

Hi enzo
I have the same amp on the bench.customer replaced 2 5W 270ohm power resistors that come off of the bridge.poor solder job= no output and DC hum.cleaned up this mess.amp powers up and plays fine...for a few minutes.snap,crackle and pop but still plays.occasionally clears up but always returns to noisy state. at one point right in the middle of monitoring, the audio quits and I get DC. all components seem fine except the output of the bridge is .032. not good.I have had issues with Crate bridge failure in the past. the schematic isn't clear on the part.can you recommend a replacement bridge? thanks. gordo (ogeecheeman)

[Enzo]

SOunds to me more like you have a bad solder connection somewhere.

[ogeecheeman]

sorry I never got back to you on this one but you were right on the solder joints... Q14 had broken foil trace & C31 was faulty.