There was an inductor in parallel with the 15r? output resistor. The drivers were heatsinked but not insulated from the heatsink. Could that have caused the problem I wonder?
Could be that, yes.
Bigger collector plane in strange shape can change the behaviour of the transistor on high frequencies by disturbing the balance between capacitance and inductance in the substrate. It can also pick some external HFs like an antenna, although they should be tamed by the power supply capacitors where the collectors are connected... or may be not, because big capacitors are not good capacitors for high frequencies and can even add their own twist to the situation. Weird!
In addition, this circuit has a tendency to oscillate with some drivers. I tested several types and found MJE243/253 to be suitable.
The oscillation may be generated in the VAS transistor as well, driving very hard the drivers, but then both drivers should have been affected...
For protection you can use the circuit protection as shown on the original amp schematic. It's not perfectly sound transparent, but will cut all excessive currents.
Could be. I have a genuine ncc200 which has the same circuit and similar pcb but has different drivers (mje15030/mje15031) which I did insulate, and different output transistors mj15003s. This has been totally stable. Small apparent differences but quite different in behaviour.
Would the mje243/253s be better drivers for the sc5200s - I want use t0264s rather that T03s
Would the mje243/253s be better drivers for the sc5200s - I want use t0264s rather that T03s
I don't know. I use them to drive MJL3281s, and before that 2SC2835s that originally came with the kit (and most likely were fake).
MJE150xx family should be OK driver either, at least looking at the datasheets. However, just because of its physical size TO-220 package is more frequency limited (and more capacitive) than the smaller TO-225.
I'm still mulling over what might have caused these driver transistors (rated at 4amps) to heat and scorch the board and components around them.
From my reading it appears that if power transistors fail, they can create a short between the collector and emitter, which then passes high voltage and current through the drivers. This apparently tends to happen with fake power transistors. I'm beginning to think perhaps the Toshiba sc5200s I had were fakes and caused this. It all happened abruptly after the amp had been working without problems for hours, then left on and idle for some time afterwards.
From my reading it appears that if power transistors fail, they can create a short between the collector and emitter, which then passes high voltage and current through the drivers. This apparently tends to happen with fake power transistors. I'm beginning to think perhaps the Toshiba sc5200s I had were fakes and caused this. It all happened abruptly after the amp had been working without problems for hours, then left on and idle for some time afterwards.
Remove the blown drivers and the adjacent power transistor, measure them and you'll know. May be the driver drove the power transistor in heavy conduction shorting collector to emitter, or it may be vice versa as you said. I thought you'd mentioned that all parts were of good quality?
I had bought the Toshibas from a reasonably respected source at a good price but had some doubts about them when they arrived.
see
http://www.diyaudio.com/forums/parts/82638-my-transistors-original-copy-41.html
I doubt that the affected drivers are measurable given their condition, but I might dissect one of the power transistors at look at the die.
I would rate this amplifier failure as serious and really want to be clear about what might have caused it before I try again. Thank you for your help and input so far Ruwe.
see
http://www.diyaudio.com/forums/parts/82638-my-transistors-original-copy-41.html
I doubt that the affected drivers are measurable given their condition, but I might dissect one of the power transistors at look at the die.
I would rate this amplifier failure as serious and really want to be clear about what might have caused it before I try again. Thank you for your help and input so far Ruwe.
Last edited:
Still trying to work out the cause of the problem. I've split open one of the power transistors -- see post 426
http://www.diyaudio.com/forums/parts/82638-my-transistors-original-copy-43.html#post2883748
It looks genuine to me but I now seek more informed opinion.
http://www.diyaudio.com/forums/parts/82638-my-transistors-original-copy-43.html#post2883748
It looks genuine to me but I now seek more informed opinion.
I'm sure you are right about this. I've dissected the power transistors and compared them with known genuine ones. They appear to have been genuine as well so they wouldn't have caused the problem. They had shorted but were thermally intact (ie not burnt, melted or showing any signs of overheating).
I believe the sequence was that oscillation caused the power transistors to pass a large current and then become short circuit. Somewhere in this process the speaker protection then detected dc and disconnected the speakers, resulting in an open circuit at the amp output terminals. The drivers (one in each of the two channels) overheated badly kept passing large currents to the base of the power transistors causing the track to this base to burn out, scorch and melt the components around the driver.
Still not sure what caused the oscillation. Bad wiring layout perhaps causing a feedback oscillation?
I'm still puzzled over the fact that you had one negative and one positive side driver blown on different channels. It seems like the fault is not consistent, which is disturbing ...
Ruwe, I can really only speculate on what really happened but perhaps both power transistors in both channels were under stress, and in each channel the most marginal power transistor went short circuit first.
In the channel I looked at, the power transistor on the negative side of the board was showed a short between b,c, and e. The associated driver was thermally and I think electrically intact.
On the positive side of the board the power transistor was open circuit, and the associated driver was more than badly overheated.
I ran a simulation to try to replicate what happened and the sim shows that a shorted power transistor and the disconnecting of the amplifier load will cause massive power dissipation in the driver of the "other" power transistor.
I have thought long and hard about why this amp burnt (I believe through oscillation) , yet another I had built that was very similar was rock stable . I had discounted a faulty component being the reason because both channels went - though of course it could have been a bad batch. I think the power transistors were "kosher" and not the cause.
I thought it might be some incompatibility between the drivers and power transistors (mje15034/35 and sc5200s) but I don't really think so.
In the end though I think it was wiring-I was trying to be neat and hadn't really thought enough about interactions and feedback between inputs and outputs.
I will try another build soon but am still unnerved and cautious after this experience. From what I read, the only definite way of detecting oscillation is with an oscilloscope, and I may yet invest in one.
I take this opportunity to give the highest commendation to Rod Elliott's speaker protection module, without which my expensive speakers would have not survived intact.
In the channel I looked at, the power transistor on the negative side of the board was showed a short between b,c, and e. The associated driver was thermally and I think electrically intact.
On the positive side of the board the power transistor was open circuit, and the associated driver was more than badly overheated.
I ran a simulation to try to replicate what happened and the sim shows that a shorted power transistor and the disconnecting of the amplifier load will cause massive power dissipation in the driver of the "other" power transistor.
I have thought long and hard about why this amp burnt (I believe through oscillation) , yet another I had built that was very similar was rock stable . I had discounted a faulty component being the reason because both channels went - though of course it could have been a bad batch. I think the power transistors were "kosher" and not the cause.
I thought it might be some incompatibility between the drivers and power transistors (mje15034/35 and sc5200s) but I don't really think so.
In the end though I think it was wiring-I was trying to be neat and hadn't really thought enough about interactions and feedback between inputs and outputs.
I will try another build soon but am still unnerved and cautious after this experience. From what I read, the only definite way of detecting oscillation is with an oscilloscope, and I may yet invest in one.
I take this opportunity to give the highest commendation to Rod Elliott's speaker protection module, without which my expensive speakers would have not survived intact.
I recently came across another post where someone decided to change their amplifier driver / power transistors combination from mje350 & 340/tip3055 &tip2955 to mje350 & mje340/2sc5200 & 2sa1943 with a bad outcome.
http://www.diyaudio.com/forums/solid-state/132110-p3a-blowing-2sc5200-2sa1943-problem.html
Prior to the change, during several hours of usage, there were apparently no problems with the amplifier. After the change the 2sc5200 &2sa1943 power transistors shorted within a few minutes. The general view in the thread was that oscillation was the cause, due to some incompatibility between the drivers and power transistors.
Even though my wiring was a bit careless and may have contributed, I now think that this incompatibility may have been the dominant cause of the oscillation in my amplifier.
http://www.diyaudio.com/forums/solid-state/132110-p3a-blowing-2sc5200-2sa1943-problem.html
Prior to the change, during several hours of usage, there were apparently no problems with the amplifier. After the change the 2sc5200 &2sa1943 power transistors shorted within a few minutes. The general view in the thread was that oscillation was the cause, due to some incompatibility between the drivers and power transistors.
Even though my wiring was a bit careless and may have contributed, I now think that this incompatibility may have been the dominant cause of the oscillation in my amplifier.
Yeah, it seems like an unfortunate combination of oscillation, due to board layout, and fake parts.
NCC200 clone bias problem
I have built a couple of these already, but on this built I am running into a new problem. I can't set the bias current in any way. I am posting my measurements hoping someone can help me sort it out. The trimpot works and both resistance and voltage change but I get zero voltage drop, so no current, across the 100ohm emitter resistor. Everything seems to check out at the LTP and current source. I have bc546 (with twisted leg) as Q5. When I turn the trimpot the emitter voltage changes. Collector voltage in Q5 remains constant at ~0.170V, I am not sure if I should see that change together with Q4 collector. All transistors and diodes checked. I have BC546 for LTP, 2SA1145/2SC2705 for VAS, MJE243/253 for drivers, 2SC5200 for output
Any help is greatly appreciated
I have built a couple of these already, but on this built I am running into a new problem. I can't set the bias current in any way. I am posting my measurements hoping someone can help me sort it out. The trimpot works and both resistance and voltage change but I get zero voltage drop, so no current, across the 100ohm emitter resistor. Everything seems to check out at the LTP and current source. I have bc546 (with twisted leg) as Q5. When I turn the trimpot the emitter voltage changes. Collector voltage in Q5 remains constant at ~0.170V, I am not sure if I should see that change together with Q4 collector. All transistors and diodes checked. I have BC546 for LTP, 2SA1145/2SC2705 for VAS, MJE243/253 for drivers, 2SC5200 for output
Any help is greatly appreciated
Hi,
It seems like you have some problem in the driver transistors. If no current flows through the output emitter resistors, a voltage of 1.4V across the Baxendall diode is not normal. If the components check OK (including that diode), you may be having an 'open'. See whether everything is soldered and that you have correct resistor values in the driver Q1 area.
Another suspicious thing is that the Q3 current source is turned on too hard for only 10-ish mA of current. May be you should replace Q3 first and check the driver Q1 after.
It seems like you have some problem in the driver transistors. If no current flows through the output emitter resistors, a voltage of 1.4V across the Baxendall diode is not normal. If the components check OK (including that diode), you may be having an 'open'. See whether everything is soldered and that you have correct resistor values in the driver Q1 area.
Another suspicious thing is that the Q3 current source is turned on too hard for only 10-ish mA of current. May be you should replace Q3 first and check the driver Q1 after.
Hi Ruwe,
thx, very helpful as always! Will start debugging as you suggested. The voltage drop across the Baxandall diode is actually 0, as I measure same voltage at both ends (and collector of lower output transistor). So if I turn the trimpot voltage arise at both ends with no drop. Same thing with voltage at collector of Q3, it goes down if I turn the trimmer but nothing else changes. The collector voltage of Q5 is supposed to change turning the trimpot or stay close to the same value?
thx, very helpful as always! Will start debugging as you suggested. The voltage drop across the Baxandall diode is actually 0, as I measure same voltage at both ends (and collector of lower output transistor). So if I turn the trimpot voltage arise at both ends with no drop. Same thing with voltage at collector of Q3, it goes down if I turn the trimmer but nothing else changes. The collector voltage of Q5 is supposed to change turning the trimpot or stay close to the same value?
This is weird. What about the voltage on the output. I assumed it's something close to the right LTP transistor base, i.e. close to 0V. But then, if you have -1.4V at the collector of the bottom output transistor, I guess the output is close to -1.4V as well, and that means that the fault may be at completely different place. For example the LTP, the feedback capacitor, grounds not connected etc.
When you measure diodes if some current flows and you don't see drop of 0.6-.7V the diode is probably shorted. If there is no current and no drop across the diode I'd say there is an open in the circuit, but the diode itself may be OK. Did you check the orientation of the Baxandall diode?
Note that both driver transistors have drop B-to-E of about 0.4V which is pretty normal...
You'll have to investigate some more. I noticed that the currents in the LTP is low, about 350uA each, but you have close to 1mA in the current source. So, it doesn't sum very well and something is missing.
Ruwe, thx again. I rounded my measurements at the LTP. The sum of the two currents is actually ~1ma equal to the current source. Your guess is right, voltage at the top output emitter is close to 0 equal to that measured at both ends of the 100ohm resistor. I pulled out the Baxandall diode and it seemed okay, I will replace it anyway. By open you mean a short in the circuit? I power it up with a light bulb in series with the rails and it behaves as it's supposed to indicating no shorts. Grounds are connected properly, all transistors and diodes rechecked....Will report again after replacing the diode
Hi,
"Open" = "disconnected", "not connected" or "no continuity"
So, what's the verdict so far? If output is at 0V and the Baxandall is OK, check the driver ...
Hi Ruwe, I replaced the diode (not faulty but just in case) but continue to get the same strange readings in the same areas . I am attaching another set of readings for anyone so kind to try to help solve the puzzle.
All transistors are checked for 0.5-0.7 fwd drop, resistors+diodes are okay, all grounds connected to common 0 point at the caps. I sample checked for continuity, don't see opens. I get no voltage/current at the output emitter (and across output emitter resistor), or across the Baxandall diode (and the 100 ohm resistor to the collector of the lower output transistor)
a) input stage/LTP. Currents across the 100 ohm resistors seem okay (0.40ma+0.63ma=1.03ma = sum of currents on the upper part of the LTP = current across the 620ohm resistor in constant current source) b) VAS + Vbe stage. I think I have proper current readings at constant current sources Q3/Q4 (7.97ma across 27ohm at collector of Q4 and across 68ohm at collector of Q3) c) Vbe multiplier. I realized I have mounted the trimpot the other way around, but this should not matter except for turning the screw in opposite way. When I turn the trimmer, the 1K resistor at the trimpot changes from ~0 (min bias ?) to ~750 (max bias?) while voltage at emitter of Q5 goes from -1.2v to -2.1v. Voltage at the collector of Q5 does not change so Vce goes from -1.4v to -2.3v when I turn the trimpot (is this correct or should Q5 collector voltage change to keep Vce at 1.4v?). Finally, lower o/p collector measures -0.9v (same as the Baxandall) but the emitter of upper o/p is close to 0.
If I connect my DMM in series with positive rail I get a 9.3ma reading that does not change adjusting the trimpot. I tried to connect an old speaker, I get a tiny very distorted music at the output. Something is definitively wrong, but can't figure it out.
Apologies for the long/boring post, need some advice. Thanks for any help
All transistors are checked for 0.5-0.7 fwd drop, resistors+diodes are okay, all grounds connected to common 0 point at the caps. I sample checked for continuity, don't see opens. I get no voltage/current at the output emitter (and across output emitter resistor), or across the Baxandall diode (and the 100 ohm resistor to the collector of the lower output transistor)
a) input stage/LTP. Currents across the 100 ohm resistors seem okay (0.40ma+0.63ma=1.03ma = sum of currents on the upper part of the LTP = current across the 620ohm resistor in constant current source) b) VAS + Vbe stage. I think I have proper current readings at constant current sources Q3/Q4 (7.97ma across 27ohm at collector of Q4 and across 68ohm at collector of Q3) c) Vbe multiplier. I realized I have mounted the trimpot the other way around, but this should not matter except for turning the screw in opposite way. When I turn the trimmer, the 1K resistor at the trimpot changes from ~0 (min bias ?) to ~750 (max bias?) while voltage at emitter of Q5 goes from -1.2v to -2.1v. Voltage at the collector of Q5 does not change so Vce goes from -1.4v to -2.3v when I turn the trimpot (is this correct or should Q5 collector voltage change to keep Vce at 1.4v?). Finally, lower o/p collector measures -0.9v (same as the Baxandall) but the emitter of upper o/p is close to 0.
If I connect my DMM in series with positive rail I get a 9.3ma reading that does not change adjusting the trimpot. I tried to connect an old speaker, I get a tiny very distorted music at the output. Something is definitively wrong, but can't figure it out.
Apologies for the long/boring post, need some advice. Thanks for any help