I suspect the same from what I saw on the photos of the wave forms on the scope.
If you don't put some load on the output, you won't see Nelson suggested.
David.
Hi Nelson,
Shows you what I get for assuming anything! And yes, that would certainly push the efforts in a different direction, although I think that is also what Pete may suspect too.
Since there is no diff pair right now, I'm not at all sure what the information could be gleaned from this experiment. Wherever the resistor load ends up pulling the offset to, I would expect a load to draw DC current. The voltage amp is shut down, but not the current amplification stage. Am I missing something here?
These output stages tend to lose other low value resistors that still look fine after they have failed. For me, walking through with an ohmmeter is the easiest way to confirm or deny anything. Things get more tricky once power is applied and you have to interpret what you are reading. An open emitter resistor will still have what looks like the expected voltage drop, only a little higher. Keeping it simple seems to be the best way to approach this to me.
The IT-18 will definitely show any signs of leakage in any of these transistors. If anything, leakage on the PCB will cause a reading more than a good transistor. Warming the transistor up as you measure it will also show any weirdness with it.
Thanks for chiming in Nelson, Chris
Shows you what I get for assuming anything! And yes, that would certainly push the efforts in a different direction, although I think that is also what Pete may suspect too.
Since there is no diff pair right now, I'm not at all sure what the information could be gleaned from this experiment. Wherever the resistor load ends up pulling the offset to, I would expect a load to draw DC current. The voltage amp is shut down, but not the current amplification stage. Am I missing something here?
These output stages tend to lose other low value resistors that still look fine after they have failed. For me, walking through with an ohmmeter is the easiest way to confirm or deny anything. Things get more tricky once power is applied and you have to interpret what you are reading. An open emitter resistor will still have what looks like the expected voltage drop, only a little higher. Keeping it simple seems to be the best way to approach this to me.
The IT-18 will definitely show any signs of leakage in any of these transistors. If anything, leakage on the PCB will cause a reading more than a good transistor. Warming the transistor up as you measure it will also show any weirdness with it.
Thanks for chiming in Nelson, Chris
Chris
David.
I,m confused about what state this amplifier is in right now. What state is it in?
David.
Last edited by a moderator:
what would be useful is not to change components ad nauseam,
but to check the strategical points of the amp and display
the voltages that are measured at theses points..
A circuit that is half working is still a working circuit, so the failure
can be easily deducted from the measured voltages..
Voltage at :
base , emitter , collector of transistors from Q1 to Q8..
Q9 and Q10, the current limiters can be omitted at this point,
and also all the output stage.
Measurements can be made with the negative probe to ground, and
hot probe (+) used with a resistor in serial , of about 10K , to prevent any
oscillation that could be triggered by the measurement influence..
but to check the strategical points of the amp and display
the voltages that are measured at theses points..
A circuit that is half working is still a working circuit, so the failure
can be easily deducted from the measured voltages..
Voltage at :
base , emitter , collector of transistors from Q1 to Q8..
Q9 and Q10, the current limiters can be omitted at this point,
and also all the output stage.
Measurements can be made with the negative probe to ground, and
hot probe (+) used with a resistor in serial , of about 10K , to prevent any
oscillation that could be triggered by the measurement influence..
Hi David,
Hi wahab,
When checking voltages in the output stage, you are further ahead to use the output of the amplifier as a reference for your meter common. That way you can see true voltage relationships without circuit noise and drift moving your numbers around. In the voltage amp stage, then putting your meter common on input ground does make sense.
Also, there are times when these measurements don't clearly point to a problem. That's when a technician shows the stuff they are made of.
-Chris
From what I understand, the diff pair has been pulled. It's confusing to follow.
Hi wahab,
Generally, yes. I would modify your plan by measuring suspect resistors without power applied. Normally, a 100 ohm resistor will decouple the capacitance of a probe, but 10 K may help reduce signal injection. Mind you, at the diff pair bases and the Vas collector, normal meter probes would really disturb normal circuit operation, even with a 10 K resistor for decoupling.
When checking voltages in the output stage, you are further ahead to use the output of the amplifier as a reference for your meter common. That way you can see true voltage relationships without circuit noise and drift moving your numbers around. In the voltage amp stage, then putting your meter common on input ground does make sense.
Also, there are times when these measurements don't clearly point to a problem. That's when a technician shows the stuff they are made of.
-Chris
hi, anatech..
Amps are an easy task generaly, as i can t remember one that i couldn t repair,
altough some of them made me just wanting to throw them trhough the window...
Amps are an easy task generaly, as i can t remember one that i couldn t repair,
altough some of them made me just wanting to throw them trhough the window...
Thread with SPICE Simulation Files
http://www.diyaudio.com/forums/solid-state/163616-spice-simulation-adcom-gfa-555-a.html
http://www.diyaudio.com/forums/solid-state/163616-spice-simulation-adcom-gfa-555-a.html
Hi David,
I split out the posts dealing with our musings of the way service was in the past.
The new thread is here
-Chris
I split out the posts dealing with our musings of the way service was in the past.
The new thread is here
-Chris
Hi wahab,
There was only one amplifier I could not repair, save a couple that suffered extensive physical damage. We got a Conrad Johnson Premier 1 amplifier in for service - power damage. The mains went about 2X the normal level and held for a while, the amp was left running all the time. The distributor couldn't fix it either, but I finally figured out what the problem was (I think) after I had left that shop.
I have created new circuit boards, stitched boards together again, replaced large burned out areas (ie - no PCB material left, a hole) and other challenging work. Some other service people called me "the court of last resort" because I got the pieces no one else could repair. That goes for FM tuners, CD players and Cassette decks and other audio nightmares. In my spare time, I like to fix things everyone else gave up on. Keeps me sharp and I still do that to this day.
Not all amplifiers are easy to repair, although they might seem to be more simple. I have seen the craziest problems that I can't fault the previous tech from finding, then there are those "struck by technician" jobs. Other times, products have been modified to death, and then there are those things that were never designed properly to begin with. I do generally look forward to repairing amplifiers, but it's the same with most audio devices for me. No, I don't fix discmans, walkmans, MP3 players or anything else like that. They have no hope of sounding really good.
One thing is for certain. Technicians can be a pretty proud bunch, but it's easier when they can work together.
-Chris
You just haven't lived long enough then I guess! 😀
There was only one amplifier I could not repair, save a couple that suffered extensive physical damage. We got a Conrad Johnson Premier 1 amplifier in for service - power damage. The mains went about 2X the normal level and held for a while, the amp was left running all the time. The distributor couldn't fix it either, but I finally figured out what the problem was (I think) after I had left that shop.
I have created new circuit boards, stitched boards together again, replaced large burned out areas (ie - no PCB material left, a hole) and other challenging work. Some other service people called me "the court of last resort" because I got the pieces no one else could repair. That goes for FM tuners, CD players and Cassette decks and other audio nightmares. In my spare time, I like to fix things everyone else gave up on. Keeps me sharp and I still do that to this day.
Not all amplifiers are easy to repair, although they might seem to be more simple. I have seen the craziest problems that I can't fault the previous tech from finding, then there are those "struck by technician" jobs. Other times, products have been modified to death, and then there are those things that were never designed properly to begin with. I do generally look forward to repairing amplifiers, but it's the same with most audio devices for me. No, I don't fix discmans, walkmans, MP3 players or anything else like that. They have no hope of sounding really good.
One thing is for certain. Technicians can be a pretty proud bunch, but it's easier when they can work together.
-Chris
Wow, you guys have been busy while I was in the woods! I must say it was nice not to encounter any technology more complex than an axe for a couple of days (illegal snowmobilers notwithstanding). A wood-stove is something I can troubleshoot successfully!
Seems like it's time to re-group and simplify. So here's my plan for proceeding:
1. Finish testing the output stage as per Pete's suggestions. Replace any bad components. Leave in place the poorly matched 30/31 drivers for now.
2. Reinstall all components in the input board. At this point, almost every component will be new. I will post a list of what was replaced and with what.
3. Re-wire and reinstall everything. (By re-wire, I mean to replace the wires that connect the input board to the output stage boards and to the speaker jacks. FWIW, none of these use any Molex connectors, they are wire-wrapped--not soldered--to posts.)
4. If it's not working, measure and document voltages at the critical points in the circuit, get some scope traces, compare voltages and traces to the good channel, and see what the brain trust has to say 🙂
So, brain trust, how does this plan sound?
Seems like it's time to re-group and simplify. So here's my plan for proceeding:
1. Finish testing the output stage as per Pete's suggestions. Replace any bad components. Leave in place the poorly matched 30/31 drivers for now.
2. Reinstall all components in the input board. At this point, almost every component will be new. I will post a list of what was replaced and with what.
3. Re-wire and reinstall everything. (By re-wire, I mean to replace the wires that connect the input board to the output stage boards and to the speaker jacks. FWIW, none of these use any Molex connectors, they are wire-wrapped--not soldered--to posts.)
4. If it's not working, measure and document voltages at the critical points in the circuit, get some scope traces, compare voltages and traces to the good channel, and see what the brain trust has to say 🙂
So, brain trust, how does this plan sound?
Hi Fred,
Sounds like a good way to regroup and get things clear.
The very first thing I would do? Carefully inspect the PCB for possible solder splashes or debris, use a toothbrush to release bits of stuff on the foil side. Next, inspect all the solder connections - even the ones you are positive are fine. Next, measure each resistor on the output boards and the voltage amp section (other board). Since the power has been off, the unit is discharged and safe to work on. May as well check the voltage drop across the signal diodes.
Test each part you have pulled. If the part is good, install it, it's best to have minimal part changes until you actually have the amp running with no other problems. After that, you can proceed to replace parts, checking for proper operation occasionally as you work.
The idea behind my advice here is to eliminate as many technician induced factors as possible. This will actually make your life easier.
-Chris
Sounds like a good way to regroup and get things clear.
The very first thing I would do? Carefully inspect the PCB for possible solder splashes or debris, use a toothbrush to release bits of stuff on the foil side. Next, inspect all the solder connections - even the ones you are positive are fine. Next, measure each resistor on the output boards and the voltage amp section (other board). Since the power has been off, the unit is discharged and safe to work on. May as well check the voltage drop across the signal diodes.
Test each part you have pulled. If the part is good, install it, it's best to have minimal part changes until you actually have the amp running with no other problems. After that, you can proceed to replace parts, checking for proper operation occasionally as you work.
The idea behind my advice here is to eliminate as many technician induced factors as possible. This will actually make your life easier.
-Chris
I have mentioned the odd bias circuit and should have noticed that it is Allison's design from 1972 but applied to the driver stage alone:
http://www.diyaudio.com/forums/soli...terview-error-correction-282.html#post1356833
I don't think this results in the best thermal compensation for the output stage since the intent is to hold the drivers at a constant bias. I believe that thermal compensation on the outputs alone would also be required. The fairly large emitter resistors on the outputs probably helps with this issue since the amp does not seem to have a problem with thermal run away.
http://www.diyaudio.com/forums/soli...terview-error-correction-282.html#post1356833
I don't think this results in the best thermal compensation for the output stage since the intent is to hold the drivers at a constant bias. I believe that thermal compensation on the outputs alone would also be required. The fairly large emitter resistors on the outputs probably helps with this issue since the amp does not seem to have a problem with thermal run away.
Hi Pete,
The bias circuit does lag the real temperature and bias changes by a bit, no doubt about it. But, that's just the nature of this beastie. They don't go thermal, so I'm happy with that as a technician.
The circuit you referred to looks like it really needs something along the lines of 0.1 ohm emitter resistors, maybe 0.22 ohms maximum. Then it runs with gobs of bias current! I can't see how it differentiates between emitter current from signal and that which is due to bias current. Otherwise it will merrily cut off the bias current from the idle side. Pretty much the same thing that happens with most current bias circuits. The 10 nF capacitor just aids in a quick transition between on and off states.
I guess I'll have to try it out and see.
-Chris
The bias circuit does lag the real temperature and bias changes by a bit, no doubt about it. But, that's just the nature of this beastie. They don't go thermal, so I'm happy with that as a technician.
The circuit you referred to looks like it really needs something along the lines of 0.1 ohm emitter resistors, maybe 0.22 ohms maximum. Then it runs with gobs of bias current! I can't see how it differentiates between emitter current from signal and that which is due to bias current. Otherwise it will merrily cut off the bias current from the idle side. Pretty much the same thing that happens with most current bias circuits. The 10 nF capacitor just aids in a quick transition between on and off states.
I guess I'll have to try it out and see.
-Chris
Back at it today. I'm not as clear on testing the output stage components as I thought I was. The current state of the boards is drivers installed (the temp 30/31) and three of the four power transistors on each side taken out of circuit by lifting a leg of the emitter resistor and removing the case screws.
I can use my IT-18/DVM to check leakage/gain on the out-of-circuit power transistors, but Pete had suggested using the PS to provide a higher voltage leakage test. I'm not quite confident I know how to set up that test.
Can some kind soul walk me through it?
cheers.
I can use my IT-18/DVM to check leakage/gain on the out-of-circuit power transistors, but Pete had suggested using the PS to provide a higher voltage leakage test. I'm not quite confident I know how to set up that test.
Can some kind soul walk me through it?
cheers.
I'm not sure what Pete had in mind.
I take it you have on set of output connected?
Have you tried powering the channel up on the variac and puting a signal through it now that you have it back together?
From the photos I looked at in a much earlier post of yours, it looked to me like only half of the output transistors were being driven. Like Nelson suggested.
If that is the case then you still have a damaged component on the input board. We would be looking for open or partially open resistors, shorted capacitors, open, or shorted, or partially damaged diodes or transistors.
I'm sorry but there is just too much thread here for me to read to see what has already been tried.
David.
I take it you have on set of output connected?
Have you tried powering the channel up on the variac and puting a signal through it now that you have it back together?
From the photos I looked at in a much earlier post of yours, it looked to me like only half of the output transistors were being driven. Like Nelson suggested.
If that is the case then you still have a damaged component on the input board. We would be looking for open or partially open resistors, shorted capacitors, open, or shorted, or partially damaged diodes or transistors.
I'm sorry but there is just too much thread here for me to read to see what has already been tried.
David.
You can check the current of the current source/sinks by measing the voltage across the emitter resistor. Check at q3 by measuring across r14. My schematic shows 4ma here so you should get roughly .66V across r14 at 332 ohms and for q4 shows 4ma so .52V across r16 130 ohms. Do this at idle, no signal. Check to see that T1 thermal switch is closed. If it is open then these two sources will collapse.
I know, it is an epic thread. When the process is concluded, one way or the other, I'll summarize it all as a thank-you to all who contributed.
Anyway, all I'm trying to do right now is figure out how to do a high-voltage leakage test of Q13-16 and Q17-20 using the amp's PS.
Anyway, all I'm trying to do right now is figure out how to do a high-voltage leakage test of Q13-16 and Q17-20 using the amp's PS.