Output transistor on - side keeps blowing

This guy knows his Shiznet :smiley:

So I can drop in the 220 Ohm resistors and then safely probe,scope, and adjust the bias pot without too much worry? Keeping an eye on the temp of the driver transistors because there will be no heat sink at this point?

What is a safe way to see if the pre amp is putting out some kind of garbage without hooking it up to the power amp? it has a tiny connector and I do know what the pins are. All I can think if is try to set up all the pins and probe the outputs with and without a signal etc?

1/ Yes. Resistors, and bias to MINIMUM. Leave it there untill the amp is fixed.
Use a variac if you have one.

2/ Cross that bridge when you get there. Fix the main amp first.
Run the main amp without load and without input signal.
Carefully measure all the DC voltages.
Post back if you're not sure about something.
I don't expect problems with the preamp at this stage.
Leo..

I should probably mention at the recommendation of GK I changed about half the op amps in the pre amp due to the gain channel was ok until I turned it up then weird noises happened. Further more before I blew the amp last time I was scoping the output and there was a wierd spike in the second half of the positive alternation. If the Diff amp Q2 (tested good) was working....Must have been on the incoming signal from pre amp?

Also I am trying to really get a strong mental grasp on what is happening with this biasing issue. I get that we are trying to set up the bias of both output transistors so that that they amplify just a tiny bit of the signal that was on the wrong side of the alternation to avoid a dead zone in the middle of the waveform. These close to 0v center parts of the alternations would turn on both transistors at the same time but if biased properly the they cancel partly and leave just the right signal left. Am I right so far?

Just curios, I know .7 is a very popular number for diodes, including the ones in the transistor B - E junction. In your experience, is that the number you will always see? Or are there Audio transistors hanging around that that .6 etc?

I am trying to understand the purpose of D7/11 and D18/19. It seems like D18/19 would never turn on and D7 and 11 I don't get what they do at all.

Also Is Q4 just for phase inversion?

Q1/16 and Q8/19 are just earlier stages of amplification before the drivers and finals yes?

Thanks so much for all the help. I realize these are elementary questions. Most of the video's I find only discuss class A and almost none seem to discuss in depth audio amps of the AB design.

HF oscillations can indeed do damage. Evidence of that can usually be found at the amp's output.
R49/R50/R51 usually release their precious smoke.

I think you're starting to understand the bias circuit.
Give the four B/E diodes (drivers and power transistors) a certain pre-voltage, so they already start to work a bit.
Then a small (music) signal does not have to overcome that B/E diode threshold first.

The voltage that a diode just starts conducting is ~0.5volt, but higher with more current. And temperature dependant.
0.65volt is average, and should be used for calculations.

D18/19 is easy to understand. They have the same function as the kickback diode over a relay.
The speaker/crosover connected to the amp is an inductive load. Spikes generated by the speaker are dumped onto the power rails by these diodes.

D7/11 seem to be "soft clip" diodes. e.g. the base of Q17 can't be driven any higher than the volt-drop over R12.

Q4 is a level shifter. Takes things from +57volt to -57volt.

Q1/16 and Q8/19 are the voltage amplifiers. The power stage is the current amplifier.
Leo..

So the final driver and output stages have no voltage gain? Only current gain? I thought that for a while but eventually convinced myself they had some voltage gain as well. Also there for current gain of course. Is this wrong?

Also for "Soft clipping" is this just for circuit protection when adjusting the bias in case a person went to high?

Thank You

Power amps usually have a voltage stage, and a current stage.
But there are also "compound" designs (rare).

The output (current) stage in this amp has indeed no voltage gain.
The bias block/driver bases are pushed/pulled to almost the supply voltages by the voltage stage.
The drivers/power transistors just follow this with more "balls".

Soft clipping has nothing to do with bias.
It is not possible to have a higher voltage out of the amp than the rail voltage (+60v/-60v).
If you try, it "clips" the tops from your signal.
When you turn up the voume more, your sine wave will turn into a square wave.
Clipping has sharp edges.
That distortion adds lots of high frequencies that are not pleasant to hear and kill tweeters.
Soft clip is a way to stop the amp from reaching 100%.
Sort of braking before you hit the brick wall.
Leo..

Soft clipping has nothing to do with bias.
It is not possible to have a higher voltage out of the amp than the rail voltage (+60v/-60v).
If you try, it "clips" the tops from your signal.
When you turn up the volume more, your sine wave will turn into a square wave.
Clipping has sharp edges.
That distortion adds lots of high frequencies that are not pleasant to hear and kill tweeters.
Soft clip is a way to stop the amp from reaching 100%.
Sort of braking before you hit the brick wall.
Leo..

I guess I can't understand how you would get a situation where you had a higher voltage then the rails because there is no place where it is available to get it form? Is it so that the voltage from the pre amp signal doesn't add on top of the rail voltage and give you rail + signal voltage?

Also...Does this 220 ohm resistor trick work on power mosfets (Gate to source) as well? I have a finicky switch mode that keeps blowing it's outputs while I'm trying to adjust the feedback pots etc.

Thank You

The output from a preamp is normally a few volts top/top.
The voltage stage of the main amp has a fixed gain (R17:R19/20), so if you feed more than a few volts into the main amp, the voltage stage will clip (at about +55v/-55v).

Mosfets and SMPSs is a different kettle of fish.
Amplifier tricks don't apply there.
Post a picture/schematic, and I will see if I can help.
Leo..

Ipod Camera is shit. I have a different thread for that issue. I get back to that problem. This one is more important. Thank you for all your explanations and tricks. If there is something I can do for you please let me know.

Thank You

The Saga Continues....

Replaced every single transistor on the board with the exact parts right from GK except for Q2 which was a sub they shipped me. I have them installed right. I am getting exactly the same readings as before so I don't believe it was the cause of the problem anyways.

So here is the testing scenario:

No tone, No load, no Final output Transistors (Q21,Q22). I have 220 ohm resistors installed B - E in there place.

My rail voltages for + and - 60vdc are actualy +58 and -58.1 on my multimeter which is not NIST calibrated. + and -15vdc is reading +13.8vdc and -14vdc.

Is this going to be a problem? I have a nice digital Tek O-Scope that is probably way more accurate but for the life of me I don;t dare hook it to anywhere but cold ground which I can't find a good spot on this board. All I see is the earth ground going straight to the chassis. I see no place to clamp onto the neutral with this board. Thought about putting something on somewhere. Last time i tried to use Hot ground and that was a mistake :frowning: Thought it would work like a volt meter but I forgot the probe is earth grounded. Should I just lift the ground on the scope with a wall adapter and use hot ground? Isolation transformers are expensive and I can't afford it right now.

Any ways...Moving on...

With R21 adjusted all the way down I am supposed to be able to set up +1.0vdc on Q17 and -1.2vdc on Q20.

However....Q17 is already at +1.030vdc and that is already to high (But probably ok?)
Q20 is -1.103vdc.
If I adjust R21 to make Q20 -1.2ish then Q17 becomes +1.1ish etc.
What should I do here?
Thought I would see if the fuse R42 is blown causing the emitters to float. Haven't found it yet. Measuring between the 2 emitters reads 472 ohms. That is not right but not open either.

Update: Yep, found it. It's wide open. Going to replace and hopefully Bias will be right.

So I need FR25-100. It is 100 ohm 1/4 watt @ 1% tolerance.
Not finding it at Digikey or mouser.

How about this one?

Man those are expensive.

Can I just use a regular 1/4 watt 100 ohm precision resistor or is that a bad idea?

Thank You

Hi again.
Seems you might have found the problem. R42.
Tolerance is not that important there, but fuseable is.
If you can't find an affordable fuseable resistor, use a standard cheap 1/4watt metalfilm resistor.
Mount it a few milimeters off the board, so when it burns it can't do any harm.
Keep that bias DOWN untill you're sure everything is ok.
Leo..

OK so I had an Idea that I could buy 1 fusible one and then get some cheaper metal film ones just for setting it up. Put the real one in at the end?

Can this Fuse being open cause my bias range issue? I have 100 ohm resistors but I don't know what kind. They just the plain tan ones. Also tolerance is probably gold.

The voltages from the supply being ever so slightly low is not a problem either right?

BTW it hides under a heat sing so I can't give it too much height off the board.

How about this one? It's 5% instead of 1% but Fusible 100 Ohm 1/4 Watt.

Thank You.

I woud certainly put it in there to try.
And buy the proper one before you close the case.

Without that resistor, with the same bias setting, bias current in the output stage is higher.
That could have blown the output stage.

Don't worry about the supply voltages being lower.
They depend on mains power fluctuations.

!!! Lower that bias all the way BEFORE you put the power transistors in !!!
!!! Leave it on low while testing the amp !!!
Adjust after you have replaced the resistor with the original one.
Leo..

[/!!! Lower that bias all the way BEFORE you put the power transistors in !!!
!!! Leave it on low while testing the amp !!!
Adjust after you have replaced the resistor with the original one.
Leo..quote]

It's been all the way down. I just crept it up .1v for a few seconds to see what effect it would have on the - side bias. I put it back down again. I'm not going to turn it up until i figure out why the Bias is already so high on the positive side without having it turned up.

Bias voltage is measured ACROSS C12. And finally, when the power transistors are in, over the 0.4ohm emiter resistors.
There is NO positive bias or negative bias.

Output OFFSET (on the speaker terminal) is also important. Has to be as low as possible, but 50mv is ok.

If the base of the driver needs 1 or 1.5volt to do that, so be it.
Leo..

Bias voltage is measured OVER C12

I am sorry but I don't understand this statement. Do you mean across C12? Like how you would measure the voltage drop across a resistor etc? If so what should it be?

I replaced R42 which was blown today. It had very little effect on the voltage at the base of Q17. Now 1.045v (should be 1.0) and Q20 is -1.106v (should be -1.2v). If I turn up Q20 to -1.2 I will get 1.145v at Q17 which is already too high as per the drawing. (These measured with bias all the way down and no test tone or load etc.

Upon further investigation I have found that R34 is not fusable resistor anymore but a precision one.
I don't know if that is a big deal or not.

R9 is out of spec. 850 ohms instead of the 1000.

Voltage at R34 should be -1.55 and at R30 -1.0 but both read 2.1v
R29 should be -0.38v but is reading 1.4 volts. Tired now. will try to understand what is happening in the AM.

Changed to across..

The voltage between the two driver bases won't change much when you replace R42.
But the current through the drivers will.

Look upon C12 (and all the bias the parts attached to it) as an elevator.
It can rise up and down, almost to the 60th floor and 60 floors down.
But the height of the elevator always stays the same.

A fusable resistor is just a normal resistor that burns up in a controlled way when something is wrong.
Up to you to use fusables or not.

R9 should be 1k. I assume you measured that disconnected, not in circuit.
A wrong value here could upset the balance of the amp.
Leo..

The voltage between the two driver bases won't change much when you replace R42.
But the current through the drivers will

Understood.

Look upon C12 (and all the bias the parts attached to it) as an elevator.
It can rise up and down, almost to the 60th floor and 60 floors down.
But the height of the elevator always stays the same.

OK, But what voltage should I see across C12? How do i know when I have it set right?

A fusable resistor is just a normal resistor that burns up in a controlled way when something is wrong.
Up to you to use fusables or not.

Understood. Just wanted to know if fusible was important here or just a cost saver.

R9 should be 1k. I assume you measured that disconnected, not in circuit.
A wrong value here could upset the balance of the amp.

I measured it just in circuit for an initial reading. going to pull it and test it right tomorrow.

Somewhere I think I was advised the voltages at the bases of the transistors listed in the drawing needed to be exact and so shouldn't the R29,30, and 34 voltages drops etc. Is that right? They are all high.

Thank You

jarrod0987:
OK, But what voltage should I see across C12? How do i know when I have it set right?

The voltage across C12 should be ~~~2volt.
That voltage is there to put the power stage (Q17,20,21,22) on the edge of conducting.
That threshold point is temperature dependent (power/music).
So the bias circuit measures the temp, and adjust that voltage.
That variation in voltage is not huge, but even the slightest change makes a big difference to the power stage current.
How do you know it's right? Bias current causes a volt drop across the four 0.1ohm resistors.
Set the bias to 20-40mA when the amp is fixed. Explained in previous posts.

jarrod0987:
Somewhere I think I was advised the voltages at the bases of the transistors listed in the drawing needed to be exact and so shouldn't the R29,30, and 34 voltages drops etc. Is that right? They are all high.

The whole voltage stage is a balancing act.
Start measuring voltage across R12/R13 (~1.55volt).
Q4 converts that into a current (0.5mA).
That is converted into a voltage across R34 > R30 > R29.
! the -1.55v listed next to R34 is the voltage across D13 and R34 !
Leo..