This is my understanding, I should let Fred answer you also:
It also draws a lot of current when it was all together clipped with oscillation which led us to suspect the output stage.
The VAS had an open collector and was the only transistor that measured bad, a few others had low beta.
The diff pair transistors are removed now, in order to let the VAS turn off, the current source then pulls the output stage negative, and when he puts a 1K resistor from C-E on the VAS in order to pull it to the pos rail just for DC testing purposes, it clips and oscillates. I was expecting this to be a simple DC test, to see if the output stage could swing to the rail. Opening T1 to turn off the current source stops the oscillation and then the output stage does go to the positive rail also.
I have 2 of these amps, and one makes a fuzztsss sound on powerdown so perhaps he is just hitting this voltage where it oscillates.
I welcome your input as I thought we would be done, ~50 steps ago.
Last edited:
What I do with my amps if I get a hard to find fault is remove all the semiconductors and test them for HFE.
While they are out I check every resistors value.
If you still cant find the problem then it could be capacitor fault.
A way to get around output transistors blowing fuses is to remove them and feedback the vas output into the ltp so all you get is the driver running.
Have a good look around the voltages and look for something wrong.
PCB faults can sometimes arise where the pcb is cracked.
My best fault was after I shorted out my amp and blew the output transistors it didnt work after putting new ouput transistors. So I removed all the semiconductors and an ltp transistor had an hfe of 1 ! I would never have guessed that one....
I am just trying to advise here since I also have this amp and plan to work on it in the near future. I have not had my hands on the actual amp.
I have done that myself years ago - removing all the semis and testing them when the amp had some serious problems.
I also like to test them at full voltage since I've seen some marginal parts over the years.
I have done that myself years ago - removing all the semis and testing them when the amp had some serious problems.
I also like to test them at full voltage since I've seen some marginal parts over the years.
Last edited:
That's largely correct, it actually starts oscillating when I get to around 40 volts on the rails. Below 40v it seems to act normally.
However, I may have missed something Pete said, because the drivers ARE still installed.
To be absolutely clear, then, with no load and everything installed EXCEPT Q1 and Q2, Q7-C tied to the + rail, it oscillates.
With Q1 and Q2 still not installed, and Q17-19 removed from circuit, it still oscillates.
Same state, turn down the bias pot to zero, it still oscillates.
Same state, bias at factory setting, remove one leg of T1, it STOPS oscillating.
I too am vastly grateful for your input since like Pete, I thought we'd have it figured by now. Needless to say, I'm also grateful he, and others, have stuck with me so long. Dogged crew.
Last edited:
That would be excellent Fred about the BOM.
When I say output stage, depending on context it might include the drivers. I think removing them was our next step, but since it works with the current source disabled we should be able to proceed.
Again, I welcome Nelson's input.
When I say output stage, depending on context it might include the drivers. I think removing them was our next step, but since it works with the current source disabled we should be able to proceed.
Again, I welcome Nelson's input.
No load, no info. Good chance that you can only swing one
polarity with a load, which will explain the oscillation.
😎
polarity with a load, which will explain the oscillation.
😎
I'm not following can you explain where the oscillation is coming from? And are you saying it will not swing to the rails with no load?
Try feeding a 1KHz signal into the amp at about 30 or 40 VPk-pk. Then check what frequency it is oscillating at? Is it breaking into oscillation at the top of the waveforms, or is it just full oscillation across the entire waveform? The mechanisms are quite different and the frequncies as well. If its loop related, it will be approx from 300k or 400k to 1MHz. If its output stage parasitic oscillation, then probably 1-3MHz (and then usually along the tops of the waveform).
Intersting circuit BTW - they really did squeeze the cost out, but this amp got some nice reviews when it came out.
Good luck with your project.
Intersting circuit BTW - they really did squeeze the cost out, but this amp got some nice reviews when it came out.
Good luck with your project.
Last edited:
Maybe I've forgotten something, but I don't believe I ever tested for the oscillation with the diff pair out, so we don't know that. That's what I meant about maybe missing something you said earlier.
Even tech writers sometimes experience FRWS (Failure to Read the Words on the Screen).
I did test for the oscillation after replacing the pair with the temporary, poorly matched 30/31 set and it was, of course, still there.
You can see a scope trace of the oscillation at post #190 (p.19). If you enlarge it you can see how the scope is set (ch.2 is the output showing the osc.).
Post #296 isn't that with the diff pair out?
If you have that oscillation on the scope again at any point, increase the sweep
speed until you have just a few cycles on the display so that we can see the shape.
Make sure that you don't have X10 magnification on the Y (sweep). and read
the period of one cycle, the reciprocal is the freq. It seems to be very low
which again makes no sense.
Voltage Amplification Stage (VAS) is Q7.
If you have that oscillation on the scope again at any point, increase the sweep
speed until you have just a few cycles on the display so that we can see the shape.
Make sure that you don't have X10 magnification on the Y (sweep). and read
the period of one cycle, the reciprocal is the freq. It seems to be very low
which again makes no sense.
Voltage Amplification Stage (VAS) is Q7.
Last edited:
I took a look at post #190. I can only comment once the waveform is fixed - looks like the top halves are chopped off (lower trace), so I think this needs to get sorted out first and then the oscillation can be looked at. Again, the frequency of oscillation is a pointer to the root cause.
I don't think its a valid debugging technique in this specific case to worry about oscillation when components (the diff pair) are removed. You should have everything in circuit.
BTW, why don't you measure the voltages at all the nodes and then put these onto the cct diagram and post this up (compare also between good and bad channel).
I don't think its a valid debugging technique in this specific case to worry about oscillation when components (the diff pair) are removed. You should have everything in circuit.
BTW, why don't you measure the voltages at all the nodes and then put these onto the cct diagram and post this up (compare also between good and bad channel).
No, post 296 trace was definitely taken with the drivers installed. I failed to properly capture the scope settings, but it was it at least .5ms/div and no slower than 2ms. It was set to 1volt/div. I'll do better next time I'm set up for testing.
I thought it was Q7, but I lack confidence in my ability to read the schematic/understand the circuit. There must have been a typo in an earlier post. It just made me think, uh oh... But no worries there.
Fred, did you pull and test Q6 yet? If you did, sorry I missed the results. To me it seems suspicious since the oscillation goes away when T1 is pulled. Also, did you end up replacing D5 and D6 or at least test them with a leg pulled? Those are the nearest components to T1.
Q7 is the VAS with a 5mA current sink, Q4. The only data I've found on the 2SA1270 says it's a 35 volt transistor but that can't possible be right as it has full power supply rails - 1.6 or so --- 160 Volts. A 35 volt transistor would be toast. The purpose of Q6 is to turn on the current regulators Q2 and Q4. If either power supply rail fails, the current sources turn off thereby shutting down the amplifier. If the thermal switch opens, again, it turns off the amplifier without actually turning off the main rails. This is often used in Motor drive Amplifiers (MDAs) to disable a motor.
For Q6 I'd look at a Zetex ZTX796 or something like it. Outside of providing a shutdown function, it has nothing at all to do with the audio. It operates at only 0.95 mA , 160 Volts and dissipates 0.15 Watts so it's not a difficult device to sub. ON semiconductor BF421 / BF423 would be OK. I would use either of these devices in my amp if needed. That 2SA1270 can't possibly be right but I can't find the typo.
G²