Can't get it to 1A - max is 670mA....and that is with 0 ohms!!!! transistor still measures out the same as before.
Yes - continuity is good.
I take it your power supply can give more than 1A and it's not being limited by your power source.
If OK then the transistor seems dodgy. Can you measure a base to collector forward voltage (does it respond as a diode would?). I've not known a device to self-limit before, but a collector-emitter short can also pass the base-collector diode test.
Do you think your latest test shows a ded or half-dead device? Try one of the other 3055's you have (hopefully on another heatsink) for a similar test.
If OK then the transistor seems dodgy. Can you measure a base to collector forward voltage (does it respond as a diode would?). I've not known a device to self-limit before, but a collector-emitter short can also pass the base-collector diode test.
Do you think your latest test shows a ded or half-dead device? Try one of the other 3055's you have (hopefully on another heatsink) for a similar test.
Mooly- the drivers I referred to are the pre-drivers if you like. They were in TO-18. ST once manufactured a BC394-BC395 pair (might have been BC393) (probably in sgs days) that were 180V but I'm not sure the U17219 etc were those.
Hi John, I'm not familiar with any of those U devices either but I would agree with what you said earlier and that there should be lots of possible options for replacements.
There has to be something very basic amiss with this example, particularly that clipping that occurs at just -6 volts or so. An open collector path on the output did give exactly that observed effect in simulation with the driver doing all the work... up to the point it expires 😉
There has to be something very basic amiss with this example, particularly that clipping that occurs at just -6 volts or so. An open collector path on the output did give exactly that observed effect in simulation with the driver doing all the work... up to the point it expires 😉
It's based on a lot of experience. An individual example can work, but my general experience was negative wth them and positive with Mj15003.
The originals were 2N3055s (made via a different process) selected for larger SOA. Ergo vanilla 2N3055s are already a backward step. I curve-traced a bunch of them years ago and Vceo leakage started at 60V, which can be attained by this circuit.
Ok going to do this again - I found that the '10 ohm' power resistor is in fact 145 ohm!- thats gone in the bin!!
So now TR2 L hits 1A at 450 Ohms
TR1L hits 1A at 1K1
The 3055 hits 1A at 1K4
TR2L measures out as it should do on diode test.
I have ordered some MJ15003 units to try - but they won't be with me until Tuesday - also stupidly (don't know how) I seem to have broken C102. I know its 0.1uF but what type is it?
My service manual shows it as a 0.01uF (so 10nF) and not 0.1uF.
It is in the audio path and so a small film type would be best choice but whatever you fit would not effect operation at all.
It is in the audio path and so a small film type would be best choice but whatever you fit would not effect operation at all.
@ejp - curious, but you may have tested or worked with more 2N3055's than I have. But the 2N3055 seems to be on a third generation of production process now. I'm pretty familiar with the "H" original device, and when the first epi devices appeared, the second breakdown limit was 60V 0.2A with a 40V breakpoint. Now that On Semi publish 60V 0.9A it seems a different beast - again. The measurements I made showed devices from about 10 years ago achieved breakdown voltages of about 80V or more. I attributed that to the need to meet the 100V Vcbo spec, which the epi technology would have needed, because the epi tech had a sharper breakdown than the old 3055's, so needed a higher BVceo to achieve the BVcbo. In most Class AB amps, you can get by with the Vcer spec (70V) at least. I have to say that the RCA epi devices were the worst, and their MJ2955's in particualar would not survive a conventional Vceo test. So I stuck with ST or ON semi after the change. Now ST no longer manufacture it anyway.
In fact I suspect the latest 2N3055 shares a line not too different from the MJ15003.
In fact I suspect the latest 2N3055 shares a line not too different from the MJ15003.
Last edited:
Mark - A curious result, because the pointers suggested something wrong with that transistor. I don't know if Quad matched the transistors, but being quite low compared with Tr1L still seems that there may be something wrong. Mooly's simulation seemed to confirm a suspected low gain/non-functional device and the overheating driver.
Let's hope your MJ15003's fix the problem.
Let's hope your MJ15003's fix the problem.
OK MJ15003s arrived -
Put 2 in for the L channel - checked all drivers - all good.
Turned on with lamp limiter - lamp comes on full brightness - turned off power and checked devices - TR1L & TR2L measure out OK.
TR106 Short C-E, TR105 Short C-E.
R122 almost vaporised!
Replaced TR105 and 106 with MJE 350 and MJE 340.
Turned on again - light stays out. Full rail and mid rail voltages are spot on, but voltage across emitter resistors is 145mV minimum.
TR105 get Very hot VERY quickly, as does R122.
Also no audio passes on this channel.
Put 2 in for the L channel - checked all drivers - all good.
Turned on with lamp limiter - lamp comes on full brightness - turned off power and checked devices - TR1L & TR2L measure out OK.
TR106 Short C-E, TR105 Short C-E.
R122 almost vaporised!
Replaced TR105 and 106 with MJE 350 and MJE 340.
Turned on again - light stays out. Full rail and mid rail voltages are spot on, but voltage across emitter resistors is 145mV minimum.
TR105 get Very hot VERY quickly, as does R122.
Also no audio passes on this channel.
None of this is making any sense at all tbh... its bizarre 🙂
Can we try a totally different approach.
Build the amp up again as you have been doing but this time remove fully from the board TR103 and TR104.
The amp in that state has to power up with zero current flow in the output stage because the outputs and TR105 and TR106 are very firmly OFF like that.
Lets see if that actually happens.
Can we try a totally different approach.
Build the amp up again as you have been doing but this time remove fully from the board TR103 and TR104.
The amp in that state has to power up with zero current flow in the output stage because the outputs and TR105 and TR106 are very firmly OFF like that.
Lets see if that actually happens.
Just to keep you up to speed...
TR1L & TR2L are now replaced on a heatsink with short flying leads so I can get to them easily if I do something stupid and blow them up!!
So TR103 and 104 are out of the board - the board powers up (bulb not lit) - mid rail voltage SLOWLY creeps up to 33.5V
TR1L & TR2L are now replaced on a heatsink with short flying leads so I can get to them easily if I do something stupid and blow them up!!
So TR103 and 104 are out of the board - the board powers up (bulb not lit) - mid rail voltage SLOWLY creeps up to 33.5V
Last edited:
Well its a start, and it seems to show there is no weird error around the outputs and their drivers. So far so good on that score.
I just wonder if this thing is going totally unstable when it burns up. I can't just get a handle on what is going on here tbh. I'll have to think.
I just wonder if this thing is going totally unstable when it burns up. I can't just get a handle on what is going on here tbh. I'll have to think.
I can only think of two things to try at this point, and this is really just to get a feel for what is going on.
With everything built up again lets try shorting the base of TR103 and TR104 together. This should prevent any conduction in the output stage.
Lets see if it survives that. No load present and no input signal for this.
The second thing was (and with it all back to normal again) to try adding a cap (say 22 or 47uF) between the base of those drivers TR103 and TR104.
I really don't know which to suggest you try first...
With everything built up again lets try shorting the base of TR103 and TR104 together. This should prevent any conduction in the output stage.
Lets see if it survives that. No load present and no input signal for this.
The second thing was (and with it all back to normal again) to try adding a cap (say 22 or 47uF) between the base of those drivers TR103 and TR104.
I really don't know which to suggest you try first...
With base of TR103/104 shorted amp turns on and mid rail rises slowly to 33.5V - minimal voltage across emitter resistors.
Same thing with capacitor across TR103/104. Voltage across emitter resistors is 18mV with a slight adjustment ( upto 23mV) with RV101. Capacitor does start to get warm.
Voltage across base of TR103/104 is 4.4v
Same thing with capacitor across TR103/104. Voltage across emitter resistors is 18mV with a slight adjustment ( upto 23mV) with RV101. Capacitor does start to get warm.
Voltage across base of TR103/104 is 4.4v
So just to be clear on this...
Two separate tests, one with the base's shorted as one test and then another test with them not shorted but now a cap between them.
For both, nothing burns up.
Assuming that is the scenario then that points to the former burn ups being caused by instability of some kind.
Is the cap (I assume you mean the one you added) getting hot in itself or is it getting heated by something nearby? It definitely should not get hot or even warm in itself.
Two separate tests, one with the base's shorted as one test and then another test with them not shorted but now a cap between them.
For both, nothing burns up.
Assuming that is the scenario then that points to the former burn ups being caused by instability of some kind.
Is the cap (I assume you mean the one you added) getting hot in itself or is it getting heated by something nearby? It definitely should not get hot or even warm in itself.
Indeed, two separate tests.
Cap was starting to get warm - not hot, but definite change in temp
OK - so board built up as normal, output devices ( on separate heatsink) connected with the exception of BASE.
Amplifier powers up, mid rail voltage appears almost immediately, emitter resistor voltage is variable between 18 and 20 mV.
Connect BASE of both output devices and amplifier turns on - all voltages correct same emitter resistor voltage range as before.
Add a 1KzH signal and the mid rail V starts to drop - with a -3dB input signal mid rail voltage is about 6.6v
Cap was starting to get warm - not hot, but definite change in temp
OK - so board built up as normal, output devices ( on separate heatsink) connected with the exception of BASE.
Amplifier powers up, mid rail voltage appears almost immediately, emitter resistor voltage is variable between 18 and 20 mV.
Connect BASE of both output devices and amplifier turns on - all voltages correct same emitter resistor voltage range as before.
Add a 1KzH signal and the mid rail V starts to drop - with a -3dB input signal mid rail voltage is about 6.6v