Back on trying to troubleshoot this GK amp.
Where I stand now;
-Q9 blew.
-Built a light bulb limiter, turned on amp, light bulb lit up and Q9 blew again
-Found that D5 blew along with Q9. Replaced.
-Fired up amp again, light bulb lit up, Q9 and D5 blew again, as well as R32.
-Mains fuse never blew in any of the above. I took the power transformer out of circuit and tested it. It shows 44v from each secondary to the center tap wire. I'm disappointed that the light bulb limiter doesn't prevent anything from blowing, because it takes a week to get parts.😒
Can anyone tell me what would be the possible causes of this? @mbz @huggygood your thoughts would be appreciated.
The two legs on either side of the notch on the rectifier beep for continuity to ground, but my meter beeps when there is some low resistance. It measures about 2.5ohms from each of those legs to ground. I replaced the rectifier and it is the same.
There is no fuse at F1 or F2 in the schematic, BTW. This seems to be just bridged with a jumper.
Is the transformer blown, even though it is showing the right voltages?
first up, reduce the wattage of you DBT bulb by about one third. Depending on amp size you should be using something like 60-75W.
Also the bulb needs to old school incandescent. LED, halogen or others wont "work".
Q9, D5, R32 blowing is somewhat unusual as most of the current flows through Q10. Suggests a near short in the area, maybe C16 has failed. Make sure zener D5 is correctly orientated.
Also the bulb needs to old school incandescent. LED, halogen or others wont "work".
Q9, D5, R32 blowing is somewhat unusual as most of the current flows through Q10. Suggests a near short in the area, maybe C16 has failed. Make sure zener D5 is correctly orientated.
Thanks very much. You're right the incandescent bulb was too high wattage - I think it was 150 watt or so. Will switch that out.
You're right it's a "near short" according to my meter: it beeps for a short below about 50ohm resistance. In several places (mentioned above) the meter actually makes a strange sound because it's right on the point between short and not short, reading 67ohms resistance. I get this on the collector of Q4, the collector of Q9 (base is dead short), that one side of c16, the resistors in the area of r32, etc.
D5 was correctly oriented - I have to order replacements - I could swap out c16.
The screws securing the transistors are a likely suspect I think, but I haven't been able to find one causing this.
Recheck for any damage to mica insulators on those 4 TO220 transistors and make sure insulating "washer" is between screw and transistor. Should not be continuity between screw and heatsink, at least for the TIP31 which has the casing as live/collector...
Hi szegedin,
I don't know what's happened to provoke damage of Q9 and D5, but are you certain R32 is 56K? It takes a lot of volts across R32 to make smoke-- 60V across 56K is only 64mW. Maybe it got replaced with something way smaller and delivered enough current to damage D5 and collateral parts?
To address your original question, some of your earlier posts suggest the opamp is not behaving as designed. Below are posts 17 and 8:
For argument, assume the circuitry following the opamp is working correctly. -12V at the opamp output would drive the discrete stages into the positive rail. With the amp output on the +60V rail, the R4 and R5 divider will deliver about +10V to the opamp + input; the - input should be about 0V. The opamp should be driven to driven to an appropriate voltage to deliver 0V at amp output; with amp stuck at positive rail, the opamp should move to positive voltage (and would be clamped by D1), but you observe -12V. Something at the opamp or surrounding components is wrong.
You can make a rudimentary test of the opamp: remove fuses F1 and F2 (or shorting links) and drive the +/- 15 rails with a bench supply or a pair of 9V batteries. Leave the other circuitry unpowered and you should observe ~0V on both + and - opamp inputs. Depending on offset voltage, the opamp should should drift to either the minus rail or clamp at a small positive due to D1. If you install a temporary short across D1, you will form a unity gain follower and should see about 0V at opamp out, i.e. the voltage on the + input.
I don't know what's happened to provoke damage of Q9 and D5, but are you certain R32 is 56K? It takes a lot of volts across R32 to make smoke-- 60V across 56K is only 64mW. Maybe it got replaced with something way smaller and delivered enough current to damage D5 and collateral parts?
To address your original question, some of your earlier posts suggest the opamp is not behaving as designed. Below are posts 17 and 8:
For argument, assume the circuitry following the opamp is working correctly. -12V at the opamp output would drive the discrete stages into the positive rail. With the amp output on the +60V rail, the R4 and R5 divider will deliver about +10V to the opamp + input; the - input should be about 0V. The opamp should be driven to driven to an appropriate voltage to deliver 0V at amp output; with amp stuck at positive rail, the opamp should move to positive voltage (and would be clamped by D1), but you observe -12V. Something at the opamp or surrounding components is wrong.
You can make a rudimentary test of the opamp: remove fuses F1 and F2 (or shorting links) and drive the +/- 15 rails with a bench supply or a pair of 9V batteries. Leave the other circuitry unpowered and you should observe ~0V on both + and - opamp inputs. Depending on offset voltage, the opamp should should drift to either the minus rail or clamp at a small positive due to D1. If you install a temporary short across D1, you will form a unity gain follower and should see about 0V at opamp out, i.e. the voltage on the + input.
Last edited:
Okay thanks. I recently replaced all those insulators, but maybe that's where something went wrong.
With the hot leg of that blown resistor (see below) and diode lifted now, I am not getting that near-short described above, but the two side legs of the rectifier are short to ground. Is that correct? It will be a week before I have the new parts.
BTW can I replace the IN966 diode with another that is 16v and 400-500mw? The IN966 is scarcely available.
@BSST Thanks for your comments. You're right, it's not 56k, it is 10ohm?? Good catch. 🙏 I can't find this resistor on the schematic. It is connected to Q9 collector and Q10 collector. Maybe it is there in lieu of the F1 fuse? Sorry for the confusion.
Unfortunately I don't have a bench power supply. The best I would be able to do is replace that opamp and try it. It's obsolete so I will add an equivalent to my order.
Yep - they are using those resistors for fuses. Make sure Q10 actually has GAIN and that it’s not open or shorted base to emitter which will overload the components driving it. If that resistor is on the collector of Q10 only, blowing it effectively takes it out of the circuit and the drivers take the load (which is bad). TIP31 gain spec is so wide anything can happen (and MJE15030 is better here should you need to change it). Also check the op-amp supplies for shorted bypass caps on the preamp board. It not the usual failure mode, but must be ruled out. Those regs have no current limiting so an overload down the line can take them out. A resistor for a fuse only prevents the place from burning down - it doesn’t actually protect transistors.
@wg_ski Ah ha, good to know, thanks. Glad to know I can use MJE15030.
Q10 tests good and is not shorted B-E, although it's possible it was shorted through the screw. Didn't blow in these last two rounds tho.
Last time I checked, the preamp section works perfectly. I had used a separate power supply to use this amp as just a preamp when I gave up on fixing the power amp section. Now I'm back on it. I can't find any shorts in caps on the board.
So far, fuses, resistors and light bulb limiter have not protected transistors. Man, I wish I could buy these parts locally.
Q10 tests good and is not shorted B-E, although it's possible it was shorted through the screw. Didn't blow in these last two rounds tho.
Last time I checked, the preamp section works perfectly. I had used a separate power supply to use this amp as just a preamp when I gave up on fixing the power amp section. Now I'm back on it. I can't find any shorts in caps on the board.
So far, fuses, resistors and light bulb limiter have not protected transistors. Man, I wish I could buy these parts locally.
Before I fire this up again, is there any way to logically determine what had to be short to cause the D5 and Fuse 1 (10 ohm resistor) to blow?
It seems that Q9 did not blow, it just showed short with the other two in circuit.
I can't find a short to ground other than the two legs of the rectifier which are 2.2ohms resistance to ground.
I replaced D5 with IN5246 instead of IN966 since the latter is not available. Should be same 16v 500mW - please stop me if this will blow things up.
I have a 60watt bulb in the light-bulb limiter, but I had tried 25w and 150w before. Hopefully 60w is right, but I'm not expecting it to protect anything.
I bought an equivalent to the op-amp but have not put it in yet because it's difficult to get that out and in without breaking anything.
Kind of expecting it to blow in the same place but I can't think of what else to test.
It seems that Q9 did not blow, it just showed short with the other two in circuit.
I can't find a short to ground other than the two legs of the rectifier which are 2.2ohms resistance to ground.
I replaced D5 with IN5246 instead of IN966 since the latter is not available. Should be same 16v 500mW - please stop me if this will blow things up.
I have a 60watt bulb in the light-bulb limiter, but I had tried 25w and 150w before. Hopefully 60w is right, but I'm not expecting it to protect anything.
I bought an equivalent to the op-amp but have not put it in yet because it's difficult to get that out and in without breaking anything.
Kind of expecting it to blow in the same place but I can't think of what else to test.
I share your fear.
The only path to destruction of D5 that I can see is failure of Q9 and or Q10. Unfortunately, there is no current limiting for the Q9, Q10 regulator circuit, so shorts on the +15V (or -15V either) rail lead to catastrophic destruction of Q9, Q10, and then D5.
I suggest unplugging the cable supplying power to the preamp stages. Check for shorts to ground of the +/- 15 preamp rails and also the power amp board. Let us know if you find anything suspicous. It might be a good idea to incorporate regulators with current limiting.
The only path to destruction of D5 that I can see is failure of Q9 and or Q10. Unfortunately, there is no current limiting for the Q9, Q10 regulator circuit, so shorts on the +15V (or -15V either) rail lead to catastrophic destruction of Q9, Q10, and then D5.
I suggest unplugging the cable supplying power to the preamp stages. Check for shorts to ground of the +/- 15 preamp rails and also the power amp board. Let us know if you find anything suspicous. It might be a good idea to incorporate regulators with current limiting.
Well; now at start up, the light bulb does not light up, and nothing exploded - that's good.
However, when I plug in a speaker, the light bulb does light up. I get the same loud hum from the speaker output with 47v DC on the speaker output. What does this point to? At least I'm back to square one, which is better than square -1. 😒
Measuring from one side of the R21 resistor to the other, I get nothing for bias. I installed a new bias pot some time back and expect to have to bias again.
What to check now?
However, when I plug in a speaker, the light bulb does light up. I get the same loud hum from the speaker output with 47v DC on the speaker output. What does this point to? At least I'm back to square one, which is better than square -1. 😒
Measuring from one side of the R21 resistor to the other, I get nothing for bias. I installed a new bias pot some time back and expect to have to bias again.
What to check now?
Last edited:
Don't hookup speaker or other heavy load until we get amp output bias back to 0VDC.
Did you replace U1 and what device did you choose? Would you report voltages at U1 (pins 2 through 7), voltages at Q1, and Q2, bases of Q3 and Q7, amp output voltage and any bias drops across R21 and R24. The schematic shows four 0.33 ohm, 5W resistors, but I see only three in your photo. Can you clarify?
Did you replace U1 and what device did you choose? Would you report voltages at U1 (pins 2 through 7), voltages at Q1, and Q2, bases of Q3 and Q7, amp output voltage and any bias drops across R21 and R24. The schematic shows four 0.33 ohm, 5W resistors, but I see only three in your photo. Can you clarify?
Here goes:
U1 pins:
1) -14.74v
2) -05.3mv
3) 8.45v
4) -14.74v
5) -14.74v
6) -11.89v
7) 14.61v
8) null
Q1 base: 47v
Q2 base: 47v
Q3 base: 48v
Q7 base: 53v
Across R21 and R24 I get nothing, 0 mA. There is a guitar version of this amp which accounts for the missing third 0.33phm resistor, so it's not really missing anything. I think that's the only difference on this schematic.
Voltage on speaker out: 47v
I bought a LME49720 op amp but did not put it in because of the high possibility of fudging up the traces taking the old one out.
Last edited:
Many things may have gone wrong in the meantime
and it is hard to follow here.
In post 1 I read that you replaced blown Q2, Q3, Q7.
Did you check associated resistors R20, R25 and D2 ?
Supply voltages of 55 V are fine and indicate that your
mains is about 10% lower than nominal. Do not even
think about the transformer being the culprit here.
Of course check both 15 V supplies after replacement,
but why check the rectifier if supply voltages are fine ?
It is obvious that two of the rect leads read low ohms to
ground, this means checking the secondary winding ..
I suspect that fuses F1 and F2 are replaced by resistors.
Many companies would use 10 ohms. 10 ohms for R32,
R33 is wrong.
On the photo I see that the collector leads of all 5 TO-220
transistors are cut and not connected. This means their
electrical contact is provided by the screw (collector is on
the metal tab also) to the copper side of the board. Have
a look at these connections. At the same time all transistors
must be isolated from the heat sink. So I do not agree with
the remark in post 25.
I began writing before seeing your last post.
and it is hard to follow here.
In post 1 I read that you replaced blown Q2, Q3, Q7.
Did you check associated resistors R20, R25 and D2 ?
Supply voltages of 55 V are fine and indicate that your
mains is about 10% lower than nominal. Do not even
think about the transformer being the culprit here.
Of course check both 15 V supplies after replacement,
but why check the rectifier if supply voltages are fine ?
It is obvious that two of the rect leads read low ohms to
ground, this means checking the secondary winding ..
I suspect that fuses F1 and F2 are replaced by resistors.
Many companies would use 10 ohms. 10 ohms for R32,
R33 is wrong.
On the photo I see that the collector leads of all 5 TO-220
transistors are cut and not connected. This means their
electrical contact is provided by the screw (collector is on
the metal tab also) to the copper side of the board. Have
a look at these connections. At the same time all transistors
must be isolated from the heat sink. So I do not agree with
the remark in post 25.
I began writing before seeing your last post.
Bunch of issues. We'll start with the opamp.
Supply voltages (pin 4 and 7) look reasonable. Pin 2 should 0mV, so voltage is a bit strange. Pin 3 looks reasonable given the PA output is about 47V. But these input voltages should force pin 6 to a positive voltage, so the output at -12V means the opamp has failed.. (With a good opamp and if pin 3 remained at 8.45V, the opamp output would clamp at about +9V when D1 conducts.)
U1 is an LF351 single opamp, right? I'm afraid the opamp you've selected is a dual opamp so it will not be compatible. Any singles in your drawer?
Would you report b, e, c of Q1. It may have failed as well. Note nominal voltages noted in the schematic.
Supply voltages (pin 4 and 7) look reasonable. Pin 2 should 0mV, so voltage is a bit strange. Pin 3 looks reasonable given the PA output is about 47V. But these input voltages should force pin 6 to a positive voltage, so the output at -12V means the opamp has failed.. (With a good opamp and if pin 3 remained at 8.45V, the opamp output would clamp at about +9V when D1 conducts.)
U1 is an LF351 single opamp, right? I'm afraid the opamp you've selected is a dual opamp so it will not be compatible. Any singles in your drawer?
Would you report b, e, c of Q1. It may have failed as well. Note nominal voltages noted in the schematic.
@as_audio
The resistor off the Emitter of Q7 (R25) does seem to be a 100ohm resistor (not a failed 22ohm). Does this reasonably account for anything?
Indeed, it is hard to follow. But it seems that it is now right back at the place I was when I started this thread. It is of course very helpful to eliminate factors that are not the problem. I did check the secondary leads of the transformer and get the same roughly 2.4 ohms between them. I disconnected the transformer from the board and found that it is delivering the correct 44v between each of the two leads to the center tap. It's helpful to rule it out.
I did replace all the insulators off the To-220 transistors noted, and checked that that the screw is not shorted to the heatsink, which it can easily do so if you get it a little wrong. You're right the contact from below the board (a terrible design) is made and checks out.
@BSST
Okay. Do we have a consensus that the opamp is failed for sure and it's not a symptom of a failure elsewhere?
Is a UA741CP a compatible single opamp? I have one of those.
Other than that I only seem to have NE5332N , RC4558p, 5218A, which are all "dual."
Q1 tests good.
If these don't work can you recommend a compatible opamp since LF351N is no longer available? I will have to order and start another episode of this saga next week.
The resistor off the Emitter of Q7 (R25) does seem to be a 100ohm resistor (not a failed 22ohm). Does this reasonably account for anything?
Indeed, it is hard to follow. But it seems that it is now right back at the place I was when I started this thread. It is of course very helpful to eliminate factors that are not the problem. I did check the secondary leads of the transformer and get the same roughly 2.4 ohms between them. I disconnected the transformer from the board and found that it is delivering the correct 44v between each of the two leads to the center tap. It's helpful to rule it out.
I did replace all the insulators off the To-220 transistors noted, and checked that that the screw is not shorted to the heatsink, which it can easily do so if you get it a little wrong. You're right the contact from below the board (a terrible design) is made and checks out.
@BSST
Okay. Do we have a consensus that the opamp is failed for sure and it's not a symptom of a failure elsewhere?
Is a UA741CP a compatible single opamp? I have one of those.
Other than that I only seem to have NE5332N , RC4558p, 5218A, which are all "dual."
Q1 tests good.
If these don't work can you recommend a compatible opamp since LF351N is no longer available? I will have to order and start another episode of this saga next week.
I suspect you may have occasion to remove Q1. With Q1 absent, the output should switch to the negative rail. I see from you latest post that Q1 is ok. The in-circuit voltages must be very strange given base at 47V. Would you report all three terminals?
Re replacing U1, I suggest TL081 mini-DIP as a suitable replacement. A low risk way to remove the opamp is to snip the leads one at a time as close to the IC body as you can manage. Then with soldering iron, remove each lead one at time. I'd recommend an IC socket for flexibility in testing. A 741 opamp might serve as a quick test, but it's too slow for a respectable sub.
Re replacing U1, I suggest TL081 mini-DIP as a suitable replacement. A low risk way to remove the opamp is to snip the leads one at a time as close to the IC body as you can manage. Then with soldering iron, remove each lead one at time. I'd recommend an IC socket for flexibility in testing. A 741 opamp might serve as a quick test, but it's too slow for a respectable sub.
I'll do that then, snipping it out. I do have a socket that I can put in there. I will try the 741 opamp because one week wait time and $8 shipping between every part I need to diagnose is more of an issue than long-term quality - I'll have a socket in there if it works.
Okay, revision to the above. Sorry for the confusion, but the schematic does not precisely match the amp. I realized that I have two extra transistors; an MPSA06 and an MPSA56. Unfortunately a "Revision C" schematic of this amp (from 1981) does not exist, but the one posted here is pretty close. In any case it's not the Crane Supercomputer, there's only so many parts and I still have to find the defective part and replace it.
So yesterday I was probing the wrong MPSA06 transistor. Q1 voltages actually appear normal; 0v base, -.5v Emitter, 52v collector.
All other data from post #34 stand.
In light of this, does the opamp explain all outstanding anomalies in voltage? I haven't snipped it out yet.