what's the voltage on pin 7 of the op amp?
does the offset change if you short C15?
what does the offset do if you lift R20 and the 47Kresistor from pin 7 of the op amp to the base of Q4? be aware that lifting these resistors will set the amp channel to a gain of 1, but for narrowing down the source of the offset, that's what we want to do. of the offset goes to about 0V, then reconnect the 47k resistor and see what happens. if it goes back to -2V, then it's the op amp or one of the caps in the servo circuit. if it stays at -2V when you disconnect the two resistors, then it's one of the diff amp transistors (Q1,2,3 or 4). replace all 4 of them.
Ok, part of the tests done. IC legs:
1 -9V
2,3, 0V
4 -10V
5,6 0V
7 +.9 V
8 +10V
I replaced the IC and the results did not change. Q1-4 had already been replaced before. The -2V offset is still there and the Led is red... probably controlled by the output of pin 7? I'll try the other tests later, but the two resistors you mentioned are on pin 1, not 7 of the op amp...
1 -9V
2,3, 0V
4 -10V
5,6 0V
7 +.9 V
8 +10V
I replaced the IC and the results did not change. Q1-4 had already been replaced before. The -2V offset is still there and the Led is red... probably controlled by the output of pin 7? I'll try the other tests later, but the two resistors you mentioned are on pin 1, not 7 of the op amp...
Last edited:
ok, i was going by the U1A and U1B designations in the schematic, but they must be backwards. something is definitely wrong if pins 2 and 3 are 0V and the output is -9V, what's shown as U1A in the schematic is a differential amp/comparator and generates the signal for the distortion indicator LED. what is shown as U1B is the DC servo amp and feeds a DC correction voltage to the inverting input of the amp module's diff amp. i think one of the electrolytics in the servo amp is leaky or dried out, and that's causing the problem.
if the schematic is correct, the -9V makes sense, because there is a negative difference between the input and output, and the -9V is an indication that there is a problem. the -9V goes through a full wave diode bridge to control the red LED.
the offset adjustment doesn't have much effect with C15 in circuit, and must be set with C15 shorted. it really is an offset adjustment for the servo amp, and not so much for the diff amp.
if the schematic is correct, the -9V makes sense, because there is a negative difference between the input and output, and the -9V is an indication that there is a problem. the -9V goes through a full wave diode bridge to control the red LED.
the offset adjustment doesn't have much effect with C15 in circuit, and must be set with C15 shorted. it really is an offset adjustment for the servo amp, and not so much for the diff amp.
Ok, I took out the 47 K resistor on pin 1 of the IC. the -9V remains there, the output is now at -7. That branch of the circuit goes to Q4. All the caps on the board have been replaced with new ones, and the IC had been replaced too. Any other suggestion? It is getting quite frustrating!
Last edited:
reconnect the 47k, swap the diff amp transistors, Q1 and Q4, and try it again, if no change, then swap Q2 and Q3. the -9 is trying to push the output positive, and it's not having enough of an effect. i would suspect Q3 or 4, but swapping the transistors from inverting to noninverting should reverse the offset (i.e. make it positive) if still no effect, then the next step is the current mirror or one of the cascodes.
Well, I reconnected the resistor, sent a 1KHz wave through the input and found that it goes though Q1 but it comes out with ~-1V offset (same on the good channel). The collector of Q7 looks the same but with ~-9V offset. Then the wave goes into Q2 base and comes out of the emitter but not from the collector that is strongly positive . then it goes into Q3 but it does not come out from the base into Q4 nor from the collector that is also strongly positive with maybe an hint of negative peaks as the B,C, E of Q5 and B, E of Q6 while the C is strongly negative... Looks like a real puzzle wrapped in an enigma, etc....
don't expect to see much on the collectors, as the output from the collectors is a current, not a voltage. also the signal is not passing sequentially from Q1 to Q4. that's not how a diff amp works. Q1/Q2 make up the noninverting input, Q4/Q3 make up the inverting input. the circuit works like a see-saw, with Q7 as the fulcrum (so to speak) Q7 is the current source for the diff amp, and supplies a constant current to the emitters of Q1-Q4. signal-wise, the waveforms on the bases of Q1 and Q4, and then the signals on the bases of Q2 and Q3 should be nearly identical.. think of Q1,2 as the input amp, and Q4,3 as an error amp which samples the output. it is the balance between these two sides (the currents through them) that determines the DC offset, much as a see-saw must have equal weight on the ends to remain level. the collector currents of Q2 and Q3 are the output, and do not have very much voltage variation on them at all. this is why the waveform appears to be very distorted, you're looking at the voltage, not the current. try swapping Q1 and Q4, then testing for a change in offset, then do the same with Q2, Q3 if there is no (or little) change. i'm thinking one of these 4 transistors is leaky, but the only way to find it is by swapping them between noninverting and inverting sides of the diff amp.
from the collectors of Q2,3 all the way to the collectors of Q9,11, voltage waveforms are pretty much irrelevant, since the signal exists as a current, not as a voltage. if you look at Q5, 6, and 8, you will see that their emitters are connected (for all intents and purposes) to the positive rail, so the voltage on anything connected to their bases will only be 0.7V or so below that. it is the B-E currents in these transistors which is being amplified, not a voltage. the signal is converted back into a voltage by Q9 and Q11.
from the collectors of Q2,3 all the way to the collectors of Q9,11, voltage waveforms are pretty much irrelevant, since the signal exists as a current, not as a voltage. if you look at Q5, 6, and 8, you will see that their emitters are connected (for all intents and purposes) to the positive rail, so the voltage on anything connected to their bases will only be 0.7V or so below that. it is the B-E currents in these transistors which is being amplified, not a voltage. the signal is converted back into a voltage by Q9 and Q11.
Thank you for the explanations! But isn't Q7 connected to Q2& Q3 rather than Q1 and Q4?
I don't have to switch transistors as I have enough spares. Therefore, I'll leave Q1 and Q2 alone and I'll substitute Q3 (D01) and Q4 (C1345) and check. The waveform shoud show on Q4 base and on Q9/Q11, correct?
I don't have to switch transistors as I have enough spares. Therefore, I'll leave Q1 and Q2 alone and I'll substitute Q3 (D01) and Q4 (C1345) and check. The waveform shoud show on Q4 base and on Q9/Q11, correct?
Last edited:
on the base of Q4 you may have some offset from the servo amp, this is normal. on Q9 collector, if everything is working right, you should have a couple volts positive offset, and a couple volts negative offset on Q11 collector. but on those collectors, you should have a waveform of about 30x what your input signal is. the offsets on Q9C and Q11C are due to the DC voltage drop across the bias circuit.
i always replace diff amp transistors as pairs (or in this case sets) you might as well, because the D01s likely came from the same batch, and the C1345's probably came from the same batch, and have been operating under identical conditions for the same amount of time, and if one of a pair is failing, the other one won't be too far behind, so you might as well replace all 4 while you have the amp module apart. it doesn't pay to take short cuts when repairing amps. 98% of the time, offset problems are caused by one of those 4 transistors being leaky. at the very least, if you're going to replace a pair of transistors, replace the D01s first, as they are under the most voltage stress, and are more likely to develop leakage. most diff amps are just one pair of transistors, and in that case, i replace both. in this amp the D01s are considered a pair, and the C1345s are considered a pair, while a D01 and a C1345 make up a side of the diff amp, and are not technically (at least in terms of the way a diff amp works) a "pair". if you want to get really technical about it, all 4 transistors constitute a "pair" because Q1,Q2 are a darlington transistor, and Q4, Q3 are the other darlington transistor in the diff amp.
so, best course of action: replace all 4
next best: replace the D01s and see if that fixes it, then replace the C1345s if it doesn't.
after you get them replaced, put a short across C15 and trim the offset with the offset pot. let the amp run idle for about 5 minutes first, and try to keep air currents away from Q1,2,3,4. you can, for the purpose of setting the offset, have the amp module standing with the board facing up. Tom Holman believed in building the amp so that repair would be as easy as possible.
once you get the amp working properly again, you will have to go through the distortion "tuning" procedure. if need be i can explain how that's done.
i always replace diff amp transistors as pairs (or in this case sets) you might as well, because the D01s likely came from the same batch, and the C1345's probably came from the same batch, and have been operating under identical conditions for the same amount of time, and if one of a pair is failing, the other one won't be too far behind, so you might as well replace all 4 while you have the amp module apart. it doesn't pay to take short cuts when repairing amps. 98% of the time, offset problems are caused by one of those 4 transistors being leaky. at the very least, if you're going to replace a pair of transistors, replace the D01s first, as they are under the most voltage stress, and are more likely to develop leakage. most diff amps are just one pair of transistors, and in that case, i replace both. in this amp the D01s are considered a pair, and the C1345s are considered a pair, while a D01 and a C1345 make up a side of the diff amp, and are not technically (at least in terms of the way a diff amp works) a "pair". if you want to get really technical about it, all 4 transistors constitute a "pair" because Q1,Q2 are a darlington transistor, and Q4, Q3 are the other darlington transistor in the diff amp.
so, best course of action: replace all 4
next best: replace the D01s and see if that fixes it, then replace the C1345s if it doesn't.
after you get them replaced, put a short across C15 and trim the offset with the offset pot. let the amp run idle for about 5 minutes first, and try to keep air currents away from Q1,2,3,4. you can, for the purpose of setting the offset, have the amp module standing with the board facing up. Tom Holman believed in building the amp so that repair would be as easy as possible.
once you get the amp working properly again, you will have to go through the distortion "tuning" procedure. if need be i can explain how that's done.
Ok, I was saying that because all 4 transistors had been replaced already with new ones. I checked the collectors of Q9 and Q11 and there is a strange, peaky waveform on both, as everywhere else ( drivers, pre drivers) up to the output plus there is a ~ 1V ripple on the + and - 85V rails (caps have been already been replaced). Maybe the problem is after Q9 /11? and is more than the offset? a nice waveform is on IC legs 6 & 7. Any way to split the circuit to try to isolate the problem better?
the ripple on the rails is normal. this amp has excellent PSRR (power supply rejection ratio) so that shouldn't be a problem.
can you show a picture of the "strange peaky waveform"? i've seen just about everything that can go wrong with this amp and might recognize the problem if i see what you're trying to describe. (we even had a failure we called the "Veejur Effect" because the waveform collapsed in a way that resembled the way that the Klingon ships and other things disappeared in the first Star Trek movie when scanned by Veejur.)
it's possible you replaced a leaky transistor with a leaky transistor, or one of the resistors in the diff amp is defective (changed value, open)
you aren't operating the amp in bridged mode are you? if you have a bad op amp on the protection board, it could leak DC through the coupling cap to the input transistor (if it's reverse biasing the electrolytic capacitor, the capacitor will leak DC and cause offset). make sure the bridging switch is OUT (off).
can you show a picture of the "strange peaky waveform"? i've seen just about everything that can go wrong with this amp and might recognize the problem if i see what you're trying to describe. (we even had a failure we called the "Veejur Effect" because the waveform collapsed in a way that resembled the way that the Klingon ships and other things disappeared in the first Star Trek movie when scanned by Veejur.)
it's possible you replaced a leaky transistor with a leaky transistor, or one of the resistors in the diff amp is defective (changed value, open)
you aren't operating the amp in bridged mode are you? if you have a bad op amp on the protection board, it could leak DC through the coupling cap to the input transistor (if it's reverse biasing the electrolytic capacitor, the capacitor will leak DC and cause offset). make sure the bridging switch is OUT (off).
No DC voltage at Q1 base. Signal is 1V pp. there is the top half of the waveform at Q3 (just replaced)input. There is a much larger waveform at the collector of Q3 that rides on the ripple. Half a waveform (larger but not as much as the other) is on collector of Q2, also riding on the ripple. The ripple is ~1V. spikes appear in the second half of the circuit with power transistors removed.
you have no feedback with the power transistors removed. the speaker line in the Apt-1, unlike most amps of it's day, does not connect to the drivers. if you're trying to run the amp without power transistors, you must wire a 1.5k/2 watt resistor in place of the B-E connection of one of the + side outputs, and one for the B-E connection of one of the - side outputs. this will provide negative feedback for the inverting side of the diff amp. troubleshooting without the 1.5k resistors will be next to impossible, as the amp will run open loop, and produce a terribly distorted waveform at best (or just latch to one rail or the other). we used to have 1.5k resistors soldered in to gutted output transistor cases as a troubleshooting tool, so we could just plug a pair of resistors in place of one + side and one - side output transistor after removing all 4 of the output transistors. you didn't tell me before that you were trying to run the amp without output transistors... that could explain a lot... get a pair of 1.5k 2 watt resistors installed and then test it again for signal and offset WITHOUT a load connected to the speaker terminals. connecting a load will burn up the 1.5k resistors. i thought you had read the rest of this thread, i'm sure i've mentioned this before...
Last edited: