Somewhat true. the input signal gets phase split and each side of the waveform gets amplifed by the complementary transistors (NPN, PNP). Class AB amplifiers have one or the other transistors slightly conducting with no signal.

troubleshooting using a bi-secting or binary search technique is common, Look at the beginning, the end and the middle. Where is the fault (e.g. begin to middle) Bisect that. etc.

https://en.wikipedia.org/wiki/Binary_search_algorithm

 

Somewhat true. the input signal gets phase split and each side of the waveform gets amplifed by the complementary transistors (NPN, PNP). Class AB amplifiers have one or the other transistors slightly conducting with no signal.

troubleshooting using a bi-secting or binary search technique is common, Look at the beginning, the end and the middle. Where is the fault (e.g. begin to middle) Bisect that. etc.

https://en.wikipedia.org/wiki/Binary_search_algorithm

So confusing ,

I think the reason i struggle with this is because i was a motorcycle mechanic for 13 years and worked with DC a lot the problem here is i now have a 0v rail then -voltave and +voltave respective to the 0v rail, with standard DC if you connect + and - its a dead short and thing go bad real quick , this is strangely different, a few resistors (voltage divider) it seems to be ok, this makes its hard to read the signal path, i know what the conponants are and what there supposed to do and how to test them , problem is this is VOODOO, hats off to all you guys, this isnt as easy as i thought it was going to be but i am going to stick at it and try learn , hopefully at some point i will be able to do this , and finaly repair this thing

 

I am watching this thread with interest. I am learning from it too. I would like to see it posted as a success thread. Hang in there @Rookieme .

 

The part of the amplifier that we are working on pulls the output towards the negative supply rail -45 volts. It deals with the negative half of the signal. In our case this is the negative half of the sine wave. Q23 to Q28 (NPN transistors) is a mirror image of Q40 to Q36 and it deals with the positive half of the signal. As the positive half of the sine wave is getting through we know that half is working. The part we are working on is 4 stages of emitter followers. The signal comes into the base of the transistor of each stage and comes out on the emitter. The emitter follower configuration provides only current gain. The output voltage is very slightly less than the input voltage and has the DC offset of the signal shifted by about 0.6 of a volt. On the negative part of the amplifier that uses PNP transistors the offset is in the positive direction.
While looking trough the tread today I calculated the expected voltage between Q29 (Bias regulator.) emitter and collector. It should be between about 3,6 and 4.2 volts. In one of the posts it was only about 1.4 volts. I did not notice it at the time as the tests then were to see if the bias voltage was too high and could have been the cause of the output transistors failing. I think this could be the cause of the loss of the negative part of the wave form as means that the base of Q 40 is not being biased far enough in the negative direction. I would like you to measure the voltage across the following resistors.

Voltage across resistors
R33
R41
R42
R54
R51
R81
R82
R86
R57
R65
R88
R53

I have started a text file of readings as I found it very time consuming going back through such a long thread. The se readings can be taken without a signal on the input.

Les.

 

When I was in my 20's, I believed in colors. Electrons don't care what the insulation color is.

There was this thing that i though black was negative and red positive. This guy at HP though that you put the white lead to the fuse holder center. I told him, you put the black lead there. White is neutral. Thermocouples got me good. Red was universally negative, but it's not ALWAYS true. Battery cables used to be red and black. Now, forget it. There are built--in fuseholders into the cable ends. I'm not even sure there is 12V written on the battery anywhere. When I did AC, white and black seemed backwards. I still can't remeber which color is which for blue and brown line colors.

HVAC thermostat wiring is another area. The terminals for heat/cool fan etc were at one time wire colors. e/g/ R, W. G and Y (red, white, green and yellow), Now you add common and W1, W2 and Rh and Rc and/or R That's now because you have to power the thermostats, so you need C. You can have separate heating and colling transformers, so different R's. More W's because of heat stages. More Y's because of cool stages.

then there's the heat pump and communicating thermostats.

Power amps are a bad first repair project.

Electronics runs on magic smoke. Once you let it out, you can't put it back in.

 

The part of the amplifier that we are working on pulls the output towards the negative supply rail -45 volts. It deals with the negative half of the signal. In our case this is the negative half of the sine wave. Q23 to Q28 (NPN transistors) is a mirror image of Q40 to Q36 and it deals with the positive half of the signal. As the positive half of the sine wave is getting through we know that half is working. The part we are working on is 4 stages of emitter followers. The signal comes into the base of the transistor of each stage and comes out on the emitter. The emitter follower configuration provides only current gain. The output voltage is very slightly less than the input voltage and has the DC offset of the signal shifted by about 0.6 of a volt. On the negative part of the amplifier that uses PNP transistors the offset is in the positive direction.
While looking trough the tread today I calculated the expected voltage between Q29 (Bias regulator.) emitter and collector. It should be between about 3,6 and 4.2 volts. In one of the posts it was only about 1.4 volts. I did not notice it at the time as the tests then were to see if the bias voltage was too high and could have been the cause of the output transistors failing. I think this could be the cause of the loss of the negative part of the wave form as means that the base of Q 40 is not being biased far enough in the negative direction. I would like you to measure the voltage across the following resistors.

Voltage across resistors
R33
R41
R42
R54
R51
R81
R82
R86
R57
R65
R88
R53

I have started a text file of readings as I found it very time consuming going back through such a long thread. The se readings can be taken without a signal on the input.

Les.

Hi LesJones,

I agree its time consuming going back and forwards through the forum,
I will do this tomorrow after work and post my results,
Also following the base waveform readings Q34 may be faulty, using the diode and ohm check its a bit off,
The PCB tracks for Q34 are also bad, i did try repair them, but that is on another forum, i will try post pix to see if there is abetter way of doing this repair
Pete

 

Q34 underside

 

When I was in my 20's, I believed in colors. Electrons don't care what the insulation color is.

There was this thing that i though black was negative and red positive. This guy at HP though that you put the white lead to the fuse holder center. I told him, you put the black lead there. White is neutral. Thermocouples got me good. Red was universally negative, but it's not ALWAYS true. Battery cables used to be red and black. Now, forget it. There are built--in fuseholders into the cable ends. I'm not even sure there is 12V written on the battery anywhere. When I did AC, white and black seemed backwards. I still can't remeber which color is which for blue and brown line colors.

HVAC thermostat wiring is another area. The terminals for heat/cool fan etc were at one time wire colors. e/g/ R, W. G and Y (red, white, green and yellow), Now you add common and W1, W2 and Rh and Rc and/or R That's now because you have to power the thermostats, so you need C. You can have separate heating and colling transformers, so different R's. More W's because of heat stages. More Y's because of cool stages.

then there's the heat pump and communicating thermostats.

Power amps are a bad first repair project.

Electronics runs on magic smoke. Once you let it out, you can't put it back in.

Your the second person to call it magic smoke, when i started this project on another forum i was taking advice and tuition from a great guy called KrisbluzNZ, unfortunately the poor guy passed away before we could complete it, and i completely lost interest for a while, god bless him

 

Your the second person to call it magic smoke

https://en.wikipedia.org/wiki/Magic_smoke

 

Here are the resistor readings
R33=2.58
R41=2.53
R42=48.69
R51=69.48
R54=0.17
R81=3.86
R82=3.93
R86=1.38
R57=1.23
R65=0.13
R88=0.00
R53=27.08

 

@LesJones

What's your gut feeling?

@Rookieme

Let me ask you about what you know about heat sinking? i.e. How would you mount the power transistors?

 

@LesJones

What's your gut feeling?

@Rookieme

Let me ask you about what you know about heat sinking? i.e. How would you mount the power transistors?

The power transistors are mounted on a fibre pad as to not make contact with the actual aluminium heat sink, along with thermal paste for better heat transfer , although these are old ones, do you think the main outputs could be shoted to the heatsynk?

 

All 4 To-3 transistors measure open circuit from the jacket to heatsynk with a continuity test

 

@LesJones

In post #184 you say the signal comes in through the base and out the emitter?

I thought the base was the point at where the transistor was made forward bias by applying around 0.6 to 0.7vdc, and the signal came in through the collector (collecting the signal) then out the emitter (emitting the signal) i know some of the signal flows in opposite directions depending if its NPN or PNP, as the diode inside controls this,

Also are you sure Q23 to 28 is the negative side as this is on the top half of the schematic and thats posative,

I could be completely wrong here

 

Q23 to Q28 deal with the positive half of the signal. See line 2 of post #184 " Q23 to Q28 (NPN transistors) is a mirror image of Q40 to Q36 and it deals with the positive half of the signal. "
There is something wrong around Q29. I would expect the voltage across R65 to be about 0.6 volts and the voltage across R57 to be about 3 volts. The current trough R54 is about 2.5 mA So Q29 is conducting when it should not be. It has leakage between collector and emitter. (But it is not fully short collector emitter short circuit which is a common mode of failure for transistors.)
This is the bias regulator circuit so this could be the cause of the problem. I suggest replacing Q29. Use the correct part number devise as this part of the circuit is quite important. (NOTE It needs to be in thermal contact with the heat sink so it can track with the base emitter voltage of the output transistors which changes with temperature.)
Re post #194 You are correct that the transistor needs to be biased with about 0.6 volts between emitter and base but the rest of your understanding is wrong. The most common configuration is grounded emitter. The signal comes in on the base and goes out on the collector. (The signal is inverted.) This gives voltage and current gain. The emitter follower. (Which is being used on Q40 to Q36.) The signal comes in on the base and goes out on the emitter. (The signal is not inverted.) Grounded base is another mode but I have only seen it used in RF circuits.)

Les.

 

Q23 to Q28 deal with the positive half of the signal. See line 2 of post #184 " Q23 to Q28 (NPN transistors) is a mirror image of Q40 to Q36 and it deals with the positive half of the signal. "
There is something wrong around Q29. I would expect the voltage across R65 to be about 0.6 volts and the voltage across R57 to be about 3 volts. The current trough R54 is about 2.5 mA So Q29 is conducting when it should not be. It has leakage between collector and emitter. (But it is not fully short collector emitter short circuit which is a common mode of failure for transistors.)
This is the bias regulator circuit so this could be the cause of the problem. I suggest replacing Q29. Use the correct part number devise as this part of the circuit is quite important. (NOTE It needs to be in thermal contact with the heat sink so it can track with the base emitter voltage of the output transistors which changes with temperature.)
Re post #194 You are correct that the transistor needs to be biased with about 0.6 volts between emitter and base but the rest of your understanding is wrong. The most common configuration is grounded emitter. The signal comes in on the base and goes out on the collector. (The signal is inverted.) This gives voltage and current gain. The emitter follower. (Which is being used on Q40 to Q36.) The signal comes in on the base and goes out on the emitter. (The signal is not inverted.) Grounded base is another mode but I have only seen it used in RF circuits.)

Les.

Can you suggest a good replacement for Q29 on digikey, i will buy a few just incase the left chanel needs replacing also

Pete

 

I advise you to use the correct transistor. It might be worth seeing if kiss agrees with my conclusion. Farnell have the BC549C in stock but ordering from there the postage would be more than the item. You could see if you have a similar rated transistor in stock and try that while you have the current limiting resistors in place of the fuses. (But fit the correct item when we are ready to remove the current limiting resistors.)

Les.

 

The power transistors are mounted on a fibre pad as to not make contact with the actual aluminium heat sink, along with thermal paste for better heat transfer , although these are old ones, do you think the main outputs could be shoted to the heatsynk?

This is more of an aside than anything else. Mica and Kapton (orange) require heat sink compound. Silpads are greyand squishy and they don't. Your final outputs are TO-3.

Generally, you only use a very thin layer of heat sink grease.

I'll try to come back to Les's questions.

 

I advise you to use the correct transistor. It might be worth seeing if kiss agrees with my conclusion.

here's https://www.mouser.com/datasheet/2/395/BC546A_series_C1703-2487965.pdf the datasheet.

The suffix matters and it's not on the schematic. Your going to have to inspect the part number and the suffix of the original parts. The suffix determines the gain.

Current gain group :
A hFE110-220
B 200-450
C 420-800

The BC546 is a better part, but check the suffix of all of the BC transistors. It matters.

I think there is some A and B parts here: https://www.mouser.com/Search/Refine?Keyword=BC546 Then there is lead spacing/

Check them. Compare with the other channel. Compare with the originals.

if you can;t find out, you may be able to ask for a parts list. https://www.img-stageline.com/products/pa-systems/pa-amplifiers/analogue-pa-amplifiers/sta-500/

The schematics say STA-160 BTW.

 

get me these values:
Q23b to Q28e that's a single differential measurement
Q40b to Q36e
and
Q23b to Q40b

Signal, no signal doesn't matter.
thanks