How to sub out a power transistor?

[Mick Bailey]

Here's some advice on conversion of Ge to Si;

Convert Your Amplifier from Germanium to Silicon

The introduction is also worth reading.

[52 Bill]

OK, I swapped out the 3.3 ohm base resistors, went to 6.8k, no sound. I changed them to 15K, now I have sound. The output is low though, about 2-3 volts peak to peak. As mentioned, I swapped out Q10, Q11, and Q12 to the MJ2955's. There is a bit of oscillation on the output, not audible, it's about .5 volts.

The base-emitters read about 6.4V when measuring across Q11 + Q12. Q10 reads about 6.2V. They are mounted directly to the chassis with mica spacers and fresh heat sink compound. The chassis is getting warm now. Not sure if I leave the 15K's in place. Do I need to up the value of the emitter resistors, or leave them?

Also not sure about Q10 with it's reading of 6.2 volts when measuring across B-E. Is it turning on enough? Any help appreciated.

I'm completely confused here with your numbers. Do you really mean 6.8K and 15K or do you mean 6.8 ohms and 15 ohms?

Are you really getting B-E readings of 6.4 volts or 0.64 volts?

How did you set R54 at the input of Q10? If Q10 doesn't work neither will Q11 or Q12.

[tubeamptech]

Ok, I was crossed eyed by the time I wrote that. Yes, I tried 6.8 ohms, and then 15 ohms. The base-emitter junctions at Q11 and Q12 were reading .64 volts, and .62 volts across Q10 base-emitter.
Once I had some signal, I tried turning R54 left or right of the original setting, the volume would drop turning the pot in either direction.

Also, the chassis is warming up considerably-no signal. I think the outputs may be on too much, ie. running too hot.

Talking to another tech friend, he said I should be monitoring voltage across R63, the 1 ohm tied to ground. I should be reading (with conversion) -50mV, as per schematic. Adjust R61 + R59 (3.3 ohm bias resistors) in order to get -50mV across R63. -50mV/1ohm = 50 mA.

Once the quiescent current is established (correct) through Q11 + Q12, then figure out where the loss is.

I'm going to swap out a few more of the caps just to be sure. C32-100MF, etc. They measured ok initially.

I disconnected the primary and secondaries of the transformer between Q10 and the outputs. No shorts to gnd, or between windings, primary end to end measured about 6 ohms, and secondaries 1-2 ohms.

Any thoughts on setting bias correctly?

[52 Bill]

Talking to another tech friend, he said I should be monitoring voltage across R63, the 1 ohm tied to ground. I should be reading (with conversion) -50mV, as per schematic. Adjust R61 + R59 (3.3 ohm bias resistors) in order to get -50mV across R63. -50mV/1ohm = 50 mA.

50mA at idle seems high to me.

[The Dude]

50mA at idle seems high to me.

Yes, it does to me also. Since you're in redesign mode and there's no actual bias procedure, I would simply monitor the current draw of the amp. Start the bias totally cold and bring it up slowly until you see the amp start to draw more current. Back it off slightly and call it good.

[52 Bill]

Yes, it does to me also. Since you're in redesign mode and there's no actual bias procedure, I would simply monitor the current draw of the amp. Start the bias totally cold and bring it up slowly until you see the amp start to draw more current. Back it off slightly and call it good.

I agree, but unfortunatly the amp has to be biased by resistor changes, no simple bias control.

My real concern would be 50mA per transistor X2 = 100mA at idle. The chassis would get warm for sure.

[tubeamptech]

Wouldn't it just be 50mA total, flowing through Q11 and Q12?

What is a typical quiescent current draw for this config, and these transistors-MJ2955?

[Ted]

Wouldn't it just be 50mA total, flowing through Q11 and Q12?

What is a typical quiescent current draw for this config, and these transistors-MJ2955?

A few milliamps would be fine. You want the Output transistors "just on" at idle. Transistor amps of this era tended to be biased cool to protect the germanium output transistors which were pretty fragile (also they have no DC stabilisation apart from those 1R emitter resistors, so biasing will tend to drift as they get hot). They generate quite a lot of distortion which is then corrected (up to a point) by applying negative feedback. You see C30 and R55? They are your negative feedback (to the signal only, not the O/P stage DC).

[52 Bill]

Wouldn't it just be 50mA total, flowing through Q11 and Q12?

It's no different than a two tube output section with separate 1 ohm cathode resistors. If you measure 50mV across one of the 1 ohm resistors, that equals 50mA current draw for that transistor/tube. If we assume that there is an equal draw from the second transistor, that will add up to 100mA total.

I guess that I would expect to see a total draw at idle of around 5-10mA.

[Ted]

It's no different than a two tube output section with separate 1 ohm cathode resistors. If you measure 50mV across one of the 1 ohm resistors, that equals 50mA current draw for that transistor/tube. If we assume that there is an equal draw from the second transistor, that will add up to 100mA total.

I guess that I would expect to see a total draw at idle of around 5-10mA.

But the idle current flows through both transistors and emitter resistors in series from the -ve supply rail to 0V - i.e the current through each of those 1 ohm resistors is the same 50mA. The current in each transistor is 50mA but the total is still 50mA.

But yes to 5 to 10 mA as a sensible idle current.

[J M Fahey]

With 2955 outputs(Q11/Q12) and 2955 regulator (Q10) :

0) check and post -V power supply voltage.
We expect something over 50V.

1) use 10 ohms resistors for bias (R59 and R61)

2) no signal, all pots on 0 , measure and post voltage across 1 ohm emitter resistors R62/R63.
By the way, did you check they are not open?

If actually 0V , donīt worry, the transistors are "almost" turned on, not a bad setting for a thermally uncompensated amp.

3) set R54, driver bias pot, to whateverīs needed to get rated voltages. in this case -21V at the emitter.

4) to take the preamp out of the map, lift/open whateverīs connecting (A)>(8) , power amp input and inject signal there (left end of C27)
Some 300/400 mV AC there should get you some 12/15VAC at the output at just clipping.

Post results.

Also post some amp pictures, inside and out.

[Ted]

Just looked at the schematic again. If the voltages drawn on to the O/P stage are correct (and thats a big if) 0.05V at top of lower emitter resistor 25.8 at the top of the upper one, that would suggest that 50mA bias current is actually correct.

[R.G.]

Not getting output for small signals, but it being there for large ones is a classic description of biasing issues.

The power amp works like this: Q9 is there for the voltage gain. It amplifies the signal from tab A and puts it out on C29 to drive Q10. It also takes feedback from the output through R55 and C30 at its emitter. The amplified signal drives the base of Q10, which is a follower used to get enough current for the driver transformer. Q10 is set with its emitter approximately in the middle of the available power supply so it can give maximum voltage swing to the primary winding of the driver transformer.

The critical voltages are for Q9 to have its collector at about 10V, as shown, whether it's silicon or germanium. This is adjusted by tinkering the values of R50 and R51; raising R51's value is the best design move for this, paralleling R50 is the simplest manually.

For Q10, you need to get the emitter to half the power supply. Notice that the schemo cannot be correct. It shows the collector of Q10 at -41.5 and the power rail at -39.5, and they are connected by a wire. No matter. For proper bias, set R54 so that Q10 emitter is about half whatever is on the collector. This forces the current in the 500 ohm emitter resistor to be about 20V/500 = 40ma, and the power in both Q10 and R56 to be about 20*0.04 = 0.8W. They'll be toasty warm. If C32 is leaky, it will let DC into the driver transformer and saturate it. C32 must NOT be leaky for good operation.

For the output stage, you want about 50ma (0.05V across R63) and at the same time for the voltage at the collector of Q12 to be almost exactly half of the voltage from Q13 emitter. Tinker with R61 and R59 to make this so, preferably with C33 disconnected. This gets the DC balance correct for the two outputs. When you reconnect C33, the DC voltage at Q12 collector should not change more than 0.1V. Otherwise, C33 is leaking too much.

When you get to here, the power amp will probably be working correctly. Test it at full power with a sine wave just barely hinting at clipping on the output and also with sine waves of about 0.5V peak. The 0.5V peak sine wave should not be visibly distorted. Crossover distortion indicates too little current in Q11 and/or Q12. Other distortions indicate problems with R55 and C30.

Once this is checked out, start working your way back into the preamp.