burning resistors-why?

[mich]

hi,
o.k,after finding a bad power transistor and replacing it,i still got some problems with this amp....R15 is burning within seconds in the circuit-i cant get why is that.and when i disconect Q6,the other 100R resistor is melting......why?i didnt find any bad traces on the pcb,no shortcuts,i checked all the diodes and polarized capacitors in the circuit-all fine.
also positive and negative DC voltage on the big cap is o.k-(21V and -21V)
here is the schematic:

thanks.

 

[WJS]

check the voltage accross the resistors, if a 100 ohm resister gets, say 20 volts across it, thats 4 watts it has to dissipate. now if you have 1/4 watt resistors, poof. check all of your voltages, and i'd recomend checking q3, q5, q4 &q6 to see if anything is shorting.
I know you found one bad one... could it have caused others to go? or could you have replaced it with another busted one?

This isn't some miracle fix, but just some idea i had.
good luck
Bill

 

[mich]

now if you have 1/4 watt resistors, poof.

I thought those resistors were ment to be 1/4 watt.....and that someone designed it to work like that....
i too think its gotta be a short somewhere.checked the rest of transistors-they show no problems.but maybe ill refrase the question:should there be 20V on them?it seems so....

This isn't some miracle fix, but just some idea i had.

i know it isnt-the only miracle will be if i kick the thing really bad and it will start working......
but thanks for the help.

 

[Mark Burnley]

Hi Mich,

Presumably you'll be firing up this circuit with a load whilst testing- hopefully a resistor and not a speaker you value!

Which transistor did you find faulty- presumably Q5? What was the fault-condition that caused this failure? Before replacing Q5 did you remove Q3 and test it? It's possible that the driver transistor Q3 actually failed originally, and Q5 dying is actually a symptom of a Q3 fault. If Q3 has gone s/c, this would account for the 100R resistor frying.

If Q3 is working okay, it's possible that you've replaced Q5 incorrectly. If Q5 is "missing" (or connected wrongly) when the load tries to draw current from the output stage, the current will be supplied through the emitter of Q3 instead of through Q5's collector (also the quiescent current that flows into Q6 will be drwan) These currents will flow through the 100R resistor in Q3's collector- which will be way in excess of it's regular non-fault current. Q3 is only acting as a base-driver for Q5, and will only ever need to supply a quiescent current + base drive to Q5.

When you remove Q6, this is exactly what is happening- current is being supplied this time by driver Q4 to the load, in turn frying the 100R R16. Removing an output transistor will also mess up the feedback (the horizontal connection running to R8), and the drive to the output drivers will be affected adversely :shock:

I'd definitely check your driver transistors, and whatever load you have connected.

One fault that bugged me once was a problem with a s/c cap in the Zobel network (C12 and R19 in this schem) It's effectively an RC filter to help damp any HF oscillations that occur due to the load on the output. In this case, the C had gone s/c, which introduced a very low resistance load to the output stage!

Good luck with it,

Mark

 

[mich]

hi mark,
i am using a speaker as a load......but one that i dont value..

Which transistor did you find faulty- presumably Q5? What was the fault-condition that caused this failure?

actually it was Q6 (from the thread about testing power transistors)
and i dont know if it came bad or i fried it.- this is a polytone power amp section that i got on ebay-that came with no power transistors-so i bought this couple and installed them and there began the problem...

now ,exacly Q3 is a subtitute part (NTE185).so ill try comparing the datasheets for both.

also when i got the amp,it had D2 conected backwards-i changed it before firing up.

i checked the zobel network of C12+R19 and both are o.k......

thanks for helping!

 

[dale116dot7]

If R15 or R16 burn up, it's because the transistors they are connected to (Q5 and Q6) are open - or they appear open. Maybe NPN where PNP should be, or maybe the replacement transistor has a different pinout or something. A base to emitter voltage drop is about 0.7 volts so there should never be more than 0.7 volts across either resistor. Those resistors are there mostly to speed up the action of the transistors Q5 and Q6 by removing base carriers when the driver transistors (Q3 and Q4) are lowering the drive current. They also act to stabilize the load of Q3 and Q4 during very fast slew rates. Check the wiring and polarity of the transistors, yea, all of them. I usually find it worth my while to check every transistor in one of these circuits, since a single part somewhere can cause the whole mess to not work - and to go up in smoke.

 

[mich]

still strugling with this amp.....
how can one check if the circuit works without conecting the power transistors?
i have burned a few allready....if they are in there they burn,and if i take them out-the resistors burn....

thanks.

 

[Mark Burnley]

Hi Mich,

With a power amp like this, as mentioned above, it's best to work your way through the whole circuit and check that all the transistors are functional before you start powering up. The kind of fault-current involved start taking out components (as you've discovered!)

Throw my transistor tester together:

...and test the transistors. I'd start with Q3-Q4 and Q5-Q6. It's a basic go/no-go tester which tests for "transistor function" i.e. a base current allows a larger collector-emitter current to flow. This flow is shown by the illumination of the LED. If the LED lights before you even press the switch, the transistor has a C-E short. If the LED doesn't light at all, even with the Test switch pressed down, then the transistor is faulty (or you have the NPN/PNP switch wrong)

Once you know the transistors are good, check the resistor values around them, and check for any faulty caps.

Replace as necessary and power up (carefully!)

Mark

 

[BYacey]

Disconnect your load, and power up gradually using a variac. Set variac to about 10 volts ac and measure the voltage across the resistor that has been burning, adjust so you have about 4 volts across the 100 ohm resistor. Slowly bring up the variac until you have this. At this point you can start measuring voltages on the transistors until you find something that is out of the ordinary. Also measure the voltage on the speaker load terminal with respect to ground. This can also help tell you where the problem may lie.

 

[Mark Burnley]

Yea,

Good tip from Bill there! You can start hunting around with the DMM while it's in a "ticking-over-but-safe" mode.

Remember to be as careful with the AC output from the variac as you would with the wall voltage- a variac is an autotransformer and not isolated from the mains.

:thumb:

Mark

 

[PRR]

If R15 or R16 blow, it is proof that the associated transistor Q5 Q6 is blown-open Base-Emitter. If these transistors were working, R15 R16 voltage could not go over 0.6V, which is 0.004 Watts, and can't smoke any resistor.

Remove Q5 Q6. The amp will "work" without them, just not into speaker impedance. Replace R17 R18 with 10Ω 10W. Don't load the output. Now check your DC voltages. The output voltage should be near zero, but if it isn't, we have to look elsewhere for clues. The voltage across R15 R16 should be near 0.6V. If it isn't, then it is probably zero or ~20V. 20V proves that the big transistor next to it is dead. The voltage across R17 R18 should be on the order of 0.05V, maybe less, not zero, and certainly not as high as 0.1V. If that looks wrong, check the voltage across the D3 D4 string (Q3 Base to Q4 Base) while you turn R14 BIAS: it should be like 1.3V max, and should go all the way to 0.7V or a little less when R14 is turned.

> got on ebay-that came with no power transistors

I would assume that means the amp is fried... why would they sell a working amp without the big-bits?

And when a power-amp fries, the first failed device usually causes other devices to fail.

I used to amuse myself by taking reject amp modules from the Fisher factory, often designs just like this, and figuring out what was wrong. Every failure is part of a chain. If you don't fix every link of the chain, it will fail again, usually instantly.

This amp has Nearly NO protection, either the outputs or the driver. For example: when (not if) Q5 blows open, Q3 tries to carry the load by itself. It can't because of R15, so Q2 tries to carry the load by throwing current through Q3 B-E diode to the load. The only limit in this current is R11 R17, 100.075Ω, I dunno what the supply voltages are, but guessing +/-30V there can be 1 or 2 Watts of power in Q2, which is probably a 0.3 Watt part. Once Q2 is blown, especially if it blows short, nothing is going to work right.

I would assume that Q1 would not fail in service. It might have been killed by idiot repair attempts, but I would replace all the other transistors before thinking about Q1.

120 Watts is an absurd rating for the 4502/802 pair. I've killed them in 65 watt amps. Don't load with less than 8 ohms. I would even suggest abandoning this Collector Follower design and using a pair of big modern audio Darlingtons for output. You lose several volts and a few watts, but they are more stable and a bit more rugged (though original 4502/802s were good).

 

[mich]

hey guys,thank you very much for helping!
i just finished running through the whole schematic with a continuity tester,checking every wire and marking on the paper,and checking ALL the components...i found another faulty resistor (R11)-and i found that the big cap on the negative rail was not making good contact with the pcb(its attached to the pcb with a screwed up screws..).
so ill spend some more time with it following your advices and see where i get too.....thanks again.

 

[PRR]

> another faulty resistor (R11)

Yup. When the outputs blow, it becomes a death-race between R15, R11, and Q2, and it is possible for all of them to lose.

 

[mich]

o.k,got some more facts.

The output voltage should be near zero

it is.im reading -18mV

The voltage across R15 R16 should be near 0.6V

i have 290mV across R16 and 230mV across R15.

The voltage across R17 R18 should be on the order of 0.05V, maybe less, not zero, and certainly not as high as 0.1V

around 0.018V

If that looks wrong, check the voltage across the D3 D4 string (Q3 Base to Q4 Base) while you turn R14 BIAS: it should be like 1.3V max, and should go all the way to 0.7V or a little less when R14 is turned.

EXACLY

 

[Mark Burnley]

mich,

With those readings above, how did you have it set up? Did you have all the transistors in place, or did you do as PRR suggested above?

Mark

 

[jrmintz]

Can I ask what model Polytone this is?

Thanks,

 

[BYacey]

What were the rail voltages when you took these measurements? If the outputs were removed when you arrived at these voltages things don't look to far off at this point. If you are sure the outputs are good, check R8,R9, C6, R20, R21, R22, R23. These affect the negative feedback, and could throw enough imbalance to saturate the outputs if defective. You'll be able to tell by powering up the complete circuit without a load and measuring the speaker output DC voltage. If all is well, it should be less than 20mv.

 

[mich]

MARK,
i had all the transistors in place (after testing them) ,and with an 8ohm speaker as the load,and i got the amp to pass audio-but without much amplification...than i replaced Q3 and burned R16again....

JRMINTZ,
thats a 3788 polytone power amp section,marked 1986 on the transformer,dont know if it was mini brute or mega brain or what exactly...

BYACEY,
i had the rails at +32V/-32V when measuring.

 

[BYacey]

I wouldn't worry about putting a load or passing audio until you have this thing working the way it should. You just run the risk of burning up your speaker coil and amp again if you do. Replace all the parts that are bad, slowly bring up the AC if you have a variac (NO Speaker LOAD Connected!!)
Measur your voltages now and see what you get.

 

[mich]

i wish i had one of those variac.... :sad: