Acoustic Model 450 - Blowing Driver Transistors

[smilan]

Hi, I need help troubleshooting this power amp. (Acoustic 450 Amp Head)
I received it in a non working condition after being plugged into a 230V instead of a 120V mains (with the power transformer wired for 120V).
The main filter cap (C505) and the bridge rectifier were shorted and I replaced them with new ones.
Two of the power transistors (480037 / 2N3773) were faulty and replaced all of the output transistors with new 2N3773G transistors.
The driver transistors Q4 and Q5 were also faulty and replaced with a pair of 2N5320 and 2N5322.
The preamp section seems to work fine and I'm getting a healthy signal at R401.
What is happening now is that as soon as I connect the output transistors (with no speaker load connected) both drivers burn.
I took voltage reading with the amp connected to a 100W bulb limiter and without the output transistors (my reading in green).
The largest differences between my measurements to the schematic was at the collector of Q404 (8V on my unit vs o 0.6V at the schematic)
And on Q1 collector (2.5V on my unit vs. 27V at the schematic).
What can make those large differences?
What else can make the drivers to get burn so fast?

 

[Khron]

The driver transistors Q4 and Q5 were also faulty and replaced with a pair of 2N5320 and 2N5322.

Where are Q4 and Q5 in the schematic?

The readings at R434 / R435 would indicate a significant (idle) current flowing through R430/R431. Are Q404 & Q405 still healthy? If that's where the bigger voltage difference is, why not suspect those (too)?

C409, output DC-blocking capacitor still ok?

Q412/Q413 seem to be V-I limiters, but those should "trigger" only closer to clipping (by reducing the bias voltage between the bases of the drivers).

Q415 and the parts around it would seem to be the bias generator - if that went open-circuit, the bias would indeed slam to max, and stuff gets hot & smokey.

 

[smilan]

Where are Q4 and Q5 in the schematic?

Q4 and Q5 are Q404 and Q405 (the drivers)

The readings at R434 / R435 would indicate a significant (idle) current flowing through R430/R431. Are Q404 & Q405 still healthy? If that's where the bigger voltage difference is, why not suspect those (too)?

Q404 and Q405 failed the last time I put the output transistors, now I installed a new pair.
And yes, those are the main suspects...

C409, output DC-blocking capacitor still ok?

Yes

Q412/Q413 seem to be V-I limiters, but those should "trigger" only closer to clipping (by reducing the bias voltage between the bases of the drivers).

Q415 and the parts around it would seem to be the bias generator - if that went open-circuit, the bias would indeed slam to max, and stuff gets hot & smokey.

Q415 is not shorted, those are it's voltage readings:
E: 34 - 35V (the voltage drops constantly, I'm guessing that with relation to heat?)
B: 34 - 35V (the voltage drops constantly)
C: 44V
The schematic does not have any voltage references for Q415, so I don't know if it is the right voltages...

 

[Khron]

The schematic does not have any voltage references for Q415, so I don't know if it is the right voltages...

https://sound-au.com/amp_design.htm#s31

Sections 3.1 and 3.2 are relevant in this case.

If Q415 was shorted C-E, that would be the best case scenario, and that would PREVENT the drivers from conducting. As it stands, Q415 might not be "open" enough, thus the voltage drop across it is too great, which makes the drivers conduct too much.

Note, i'm no expert, and this might not be the (only) root cause, but the symptoms do start to point in that direction.

On second thought, the differential pair at the input could also have some issues, and that 2.5V at the collector of Q401 is over ten times lower than it's supposed to be (27V).

Are R410 and R411 the values they're supposed to be? Are CR406 / CR407 ok? What about the VAS, Q403? That could also be oscillating actually; is C406 still in one piece? You could also temporarily remove C405 (and check it) and see if anything changes - that's there for AC (slightly positive) feedback, to increase the output impedance of the amp, to offer a more "tube-like" response.

 

[smilan]

If Q415 was shorted C-E, that would be the best case scenario, and that would PREVENT the drivers from conducting. As it stands, Q415 might not be "open" enough, thus the voltage drop across it is too great, which makes the drivers conduct too much.

How is possible to control opening of Q415?

On second thought, the differential pair at the input could also have some issues, and that 2.5V at the collector of Q401 is over ten times lower than it's supposed to be (27V).

I agree, it seems wrong, but I don't really understand what can make this difference?

Are R410 and R411 the values they're supposed to be?

R410 - 2.179
R411 - 0.998K
Bothe seem to be with the right values.

Are CR406 / CR407 ok?

Yes, with R411 in place the diodes conducting both ways but when R411 is disconnected both diodes perform well (conducting only in one way).

What about the VAS, Q403? That could also be oscillating actually; is C406 still in one piece?

Q 403 tested good and C406 measured 473pF, so both seems to be good.

You could also temporarily remove C405 (and check it) and see if anything changes - that's there for AC (slightly positive) feedback, to increase the output impedance of the amp, to offer a more "tube-like" response.

I removed C405 and couldn't see any change the not in the collector of Q404 and not in the collector of Q401.

 

[Khron]

Between the bases of Q404 and Q405 you're supposed to have about 2.4V (47.8-45.4).

Currently you're measuring about 9V, so the drivers and final-stage transistors are about as "open" as they can get.

So the issue must be somewhere between emitter of Q407 (R445, C416, Q415, R444) and the collector of the VAS transistor, Q403. I'm guessing some broken trace, open-circuit resistor, or if it's present in your unit, open R416 trimpot or toasted CR408/409 diodes.

I'd also replace the Q401 / Q402 input pair, just to eliminate them as suspects. It's very weird that they have the same, healthy-looking voltages on the base and emitter, but Q401's collector is almost at ground potential, despite the things connected to it seeming ok.

 

[smilan]

Between the bases of Q404 and Q405 you're supposed to have about 2.4V (47.8-45.4).

Currently you're measuring about 9V, so the drivers and final-stage transistors are about as "open" as they can get.

So the issue must be somewhere between emitter of Q407 (R445, C416, Q415, R444) and the collector of the VAS transistor, Q403. I'm guessing some broken trace, open-circuit resistor, or if it's present in your unit, open R416 trimpot or toasted CR408/409 diodes.

R445 - 670R
R444- in my unit there is a 3K9 (orange, white, red) measured 3K88 instead of 2K7 in the schematic, do you think it worth to try it with a 2K7? (the amp used to work with the 3K9)
C416 - 473pF
Q415 - Tested well

I'd also replace the Q401 / Q402 input pair, just to eliminate them as suspects. It's very weird that they have the same, healthy-looking voltages on the base and emitter, but Q401's collector is almost at ground potential, despite the things connected to it seeming ok.

I've done it earlier and replaced Q401 and Q402 no change in the behavior of this stage.

Also R436 and R437 are "Factory selected value resistors". In my unit they are both 2K2, Do you think that different values can work better?

 

[JohnRoberts]

From a quick drive by glance Q 401 and Q402 are clearly not operating normally. There should not be more than a single diode drop voltage between base and emitter. The schematic suggest 0.6V as normal voltage drop... your measurements show 0.9V across Q401 b-e and 0.8V across Q402 b-e. The fact that Q401 collector is measuring low voltage despite strong 0.9V of base Vbe is definitely suspicious. Not expected behavior from healthy devices.

JR

 

[CJ]

Not a high parts count, so you can shotgun the whole circuit with new parts in about 1 hour,

 

[smilan]

From a quick drive by glance Q 401 and Q402 are clearly not operating normally. There should not be more than a single diode drop voltage between base and emitter. The schematic suggest 0.6V as normal voltage drop... your measurements show 0.9V across Q401 b-e and 0.8V across Q402 b-e. The fact that Q401 collector is measuring low voltage despite strong 0.9V of base Vbe is definitely suspicious. Not expected behavior from healthy devices.

JR

I replaced Q401 and Q402 with new transistors and it acts the same, I've double checked all the components around those transistors and all tested well (CR405 and C403 replaced with new components and all the resistors tested pretty close to their values).

Other than that, it seems to me like the difference between the voltage of the base of Q404 and Q405 is too big, but I don't know how to control it?

 

[JohnRoberts]

I am not sure what exactly the problem is but the 90v rail only measuring 70V is suspicious. The output biased at 1/2 supply looks close to correct at 35v, but the 70v rail is wrong.

is Q407 base 43.6v and emitter 35.5V ? that would be almost 8V Vbe. Am I reading your notes correctly?

and 8V at base Q406 looks wrong with emitter only 94 mV

Is something getting crazy hot inside that amp... ? Lots of invalid voltages.

JR

 

[smilan]

I am not sure what exactly the problem is but the 90v rail only measuring 70V is suspicious. The output biased at 1/2 supply looks close to correct at 35v, but the 70v rail is wrong.

Make a long story short, the problem with this amplifier is that Q404 and Q405 are getting too how and blowing.
While trying to troubleshoot this issue I took those measurements with the amplifier connected to a 100W light bulb limiter and a variac that dropped the 90V rail to 70V. Without the variac and the current limiter there is 92V at the reservoir cap, but Q404 and Q4050 getting burn after few seconds.

is Q407 base 43.6v and emitter 35.5V ? that would be almost 8V Vbe. Am I reading your notes correctly?

and 8V at base Q406 looks wrong with emitter only 94 mV

I took those measurements with the amplifier connected to a 100W light bulb limiter and a variac that dropped the 90V rail to 70V. Without the variac and the current limiter there is 92V at the reservoir cap, but Q404 and Q4050 get burned after a few seconds.

 

[smilan]

Not a high parts count, so you can shotgun the whole circuit with new parts in about 1 hour,

I've already replaced all the transistors, diodes, high wattage resistors and electrolytic caps on the power amp board, until now with no change...

 

[Khron]

Could still be some issue with the PCB traces, solder joints or smaller resistors...

 

[JohnRoberts]

Make a long story short, the problem with this amplifier is that Q404 and Q405 are getting too how and blowing.

OK that makes sense (if I substitute hot for how).

If the output devices have too much bias voltage they will over heat and release their magic smoke.

Taking a small step back that is a quasi-complementary topology. Decent PNP power devices were not widely available back in the day, so a small PNP (Q404) is configured with a power NPN (Q406) to behave like a single power PNP. The class A bias is established by Vbe multiplier Q415. By feeding the base with a simple voltage divider those resistors act like feedback resistors to control the voltage collector-emitter. Typically the two resistors in that Vbe divider connect between the bases of the two output stage power transistors. A variation in this design connects the bottom resistor R444 as expected, but connects the top resistor R445 to the emitter of Q407. This should still work give or take a diode drop here and there.

To suss out what is wrong, you need to measure the voltages around the Vbe multiplier Q415. Note: the fact the value of R445 is not printed on the schematic suggests that it might be tweaked for different sets of transistors. It looks like Q415 collector is 44.5V and emitter is 35.5V. It would be useful to know what the base voltage of Q415 is, and value of R445.

JR

While trying to troubleshoot this issue I took those measurements with the amplifier connected to a 100W light bulb limiter and a variac that dropped the 90V rail to 70V. Without the variac and the current limiter there is 92V at the reservoir cap, but Q404 and Q4050 getting burn after few seconds.

I took those measurements with the amplifier connected to a 100W light bulb limiter and a variac that dropped the 90V rail to 70V. Without the variac and the current limiter there is 92V at the reservoir cap, but Q404 and Q4050 get burned after a few seconds.

 

[smilan]

Thank you all, I found the problem!
While replacing the diodes I made a mistake and installed CR410 backwards... After putting it in the right direction everything seems to work fine in the power amp with all the transistors installed. Now I found that there is a loud noise (50Hz with strong 100Hz harmony) coming from the pre amp (when I disconnect the wire feeding the signal to the power amp the amp is quiet so I think recapping the preamp is the next thing to do...

 

[JohnRoberts]

Thank you all, I found the problem!
While replacing the diodes I made a mistake and installed CR410 backwards...

CR410 is part of the current limiting, if installed backwards it just would not current limit in the negative direction. Perhaps something else was wrong.

JR

After putting it in the right direction everything seems to work fine in the power amp with all the transistors installed. Now I found that there is a loud noise (50Hz with strong 100Hz harmony) coming from the pre amp (when I disconnect the wire feeding the signal to the power amp the amp is quiet so I think recapping the preamp is the next thing to do...