This has been resolved. The problem were fake transistors.
I recently build myself a discrete single supply 15w amplifier and it worked great but after about a week of using the pnp output transistor blew up and shorted for no reason(it wasnt overheating and it was running at 40% volume) after that i replaced the transistor with a new one and the same thing happened again after about a week of use(again it didnt overheat and the amp wasnt overloaded) I tried troubleshooting but i failed to find a reason why the transistors keep failing? I think its because they are fake but im not 100% sure(i bough a couple of them at once so maybe all of them were fake but im not sure).If anyone can help me find the issue i would apreciate it!
(sorry if my english isnt the best)
The red circle is the transistor that keeps blowing up
Note: the 2k resistor in the bias spreader is a trimmer to adjust the bias
I recently build myself a discrete single supply 15w amplifier and it worked great but after about a week of using the pnp output transistor blew up and shorted for no reason(it wasnt overheating and it was running at 40% volume) after that i replaced the transistor with a new one and the same thing happened again after about a week of use(again it didnt overheat and the amp wasnt overloaded) I tried troubleshooting but i failed to find a reason why the transistors keep failing? I think its because they are fake but im not 100% sure(i bough a couple of them at once so maybe all of them were fake but im not sure).If anyone can help me find the issue i would apreciate it!
(sorry if my english isnt the best)
The red circle is the transistor that keeps blowing up
Note: the 2k resistor in the bias spreader is a trimmer to adjust the bias
Last edited:
If your outputs are real, they should both blow. You may simply be driving the output too hard, depending on supply voltage and bias current (and heat sinking). 40% volume may be the point where it dissipates the most heat for a sine wave. 1/3 maximum output is accepted as that point for a constant sine.
What bias current are you running, supply voltages? How hot is the heat sink normally (°C)?
What bias current are you running, supply voltages? How hot is the heat sink normally (°C)?
And what is the impedance of speakers? Are commercial speakers or are made by you?
You know that the schematic is incomplete?!?
Another reason can be high frequency oscillations. The R18 from the schematic is hot? It is connected?
However, the SOA on these amplifiers is not extraordinary and if they have a larger power supply, the available current is significantly reduced.
You know that the schematic is incomplete?!?
Another reason can be high frequency oscillations. The R18 from the schematic is hot? It is connected?
However, the SOA on these amplifiers is not extraordinary and if they have a larger power supply, the available current is significantly reduced.
Hi thanks for the reply can you please explain what you mean by "If your outputs are real, they should both blow"? Also the supply voltage is 50V supplied by a bench power supply and the bias current is about 55-60 ma. And for heatsinking they never really get too hot.
the impedance is 8 ohms commercial speaker. Also what do you mean that the schematic is incomplete?! The resistor in the output filter doesnt get hot
The resistor in the output filter doesnt get hot
the impedance is 8 ohms, commercial speaker. Also what do you mean that the schematic is incomplete?!
The resistor in the output filter doesnt get hot
the supply voltage is 50v and the voltage before the cap is very close to half of the supply voltage (its off by about half a volt)
TIP41/2 should hold up on +/-25V with 8 ohm load. 4 ohm is a little iffy, but that would be making he heat sink hot. Those types are supposed to handle full power to 40V. If they are shorting before heating up, that usually indicates secondary breakdown occurring. And s/b should not be occurring at 25V on real ones. Hard to imagine that happening on fakes either, but fakes also have other reliability problems, and could be TIP32’s in 42’s clothing.
At 50Hz and 50V (+/-25V) trey have less than 1A current capability before second breakdown. At the peak of the sinus current, the opposite transistor will see the full supply voltage.
Replace them with the classic pair: 2SC5200 with 2SA1943 which have over 10A at 50Hz and 50V and I don't think you will have any more problems. Or any other pair similar in parameters.
Replace them with the classic pair: 2SC5200 with 2SA1943 which have over 10A at 50Hz and 50V and I don't think you will have any more problems. Or any other pair similar in parameters.
At full supply voltage, the transistor sees very little if any current. High currents can only really exist with 25V or less on it. At resistive load the max instantaneous dissipation occurs at 12.5V. Into a fully reactive load it happens at 25. With a real world speaker, somewhere in between.
If you follow your logic, C5200’s wouldn’t be able to run at +/-40V. They can and do, even at 4 ohm.
If you follow your logic, C5200’s wouldn’t be able to run at +/-40V. They can and do, even at 4 ohm.
ok but the tip41/2c says in the datasheet that it says that at a 25v supply it can handle about 2.5A? How did you get that 1A current capability? isnt each transistor supposed to see +/- 25v accordingly?
I do not think so, but that is just an assumption.
These TIPs are slow devices.
If in doubt, do a simulation or real measurements.
This is probably not your present problem, but... The trimmer can best be put in the other branch (location R11), so that the quiescent current decreases rather than increases when the wiper loses contact.
@Angel23
I was about to reply to your original thread last night and got as far as actually typing it all out... and then you deleted the thread.
Short version...... you need an output coil, you need base stopper resistors and C9 also fell under my suspicion given the closed loop gain. It looks far to high in value. The vbe multiplier must be on the main heatsink.
You have done a simulation so look at stability margins.
I was about to reply to your original thread last night and got as far as actually typing it all out... and then you deleted the thread.
Short version...... you need an output coil, you need base stopper resistors and C9 also fell under my suspicion given the closed loop gain. It looks far to high in value. The vbe multiplier must be on the main heatsink.
You have done a simulation so look at stability margins.
The vbe multiplier is on the heatsink already. Also regarding C9 you said it might be too high so what value woud you think will be better and is there a way to calculate it? Also i researched a little about base stopper resistor but cant figure out a proper value ive seen values around 1.5/2.2 ohms but im not sure
Something in the 4.7 to 10 ohm region for base stoppers. C9 I would have thought would be around 22pF to 47pf as a typical value for your component values and closed loop gain. Also don't run to high a bias current, around 50ma should be fine and give some protection against thermal runaway.
Post your .asc sim file (it will attach directly to the forum) and I'll run it later.
Post your .asc sim file (it will attach directly to the forum) and I'll run it later.