views about the circuit

[pcaicedoare]

it's a good day

 

[sheeck]

Have anybody PCB for this amp? http://www.blackwidowaudio.com/

 

[Bonsai]

There are a few comments here about 'old work horses' relating to the 2N3055. This shows a complete lack of appreciation for some of the topological advances made in audio amplifier design over the last 40 years.

 

[steveu]

Palesha,
Comments? It is a good place to start. The circuit is ye olde so it could benefit greatly from a few modifications:

1. Remove R3,C2,R2,D1. Add a constant current diode between TR2/TR3 emitters and +40V.
2. Replace TR2/TR3 with an SSM2220P match pair.
3. Put a 470pF cap in parallel with R4 and a 4k7 in series with C3 (for high frequency stability reasons).
4. Remove C6,R1,R11. Connect a constant current diode (5mA to 10mA, 1W) from the R12,R13 node to +40V. Put a 47uF cap between the R21,R13 node and the R22,D4 node.
5. 2N3055s are very old, slow workhorses. You would be better to use more modern output transistors with much higher Ft, like MJ15022 or similar.
6. Add a 10-ohm resistor in series with a 220nF capacitor across the output to ensure the circuit is loaded at high frequencies (improves stability).

You can pick and choose which to make, they are independent. C7 is a critical component for the stability of the circuit. It should have a value of at least 47pF, perhaps 100pF for initia switch on. Otherwise the circuit may oscillate or oscillate when you connect a speaker to it. For best results use tantalum electrolytics.

1. D1 and C2 perform a turn-off de-thump that a CC diode will not. Constant current diodes are not used much by professional for a number of reasons. C2 also filters supply noise. The best option here is the standard current constant cct using two bipolars but as is, is not bad. Added de-thumping comes from a current mirror instead of R5.
2. The integrated matched pair SSM2220P has a breakdown voltage lower that the shown rail voltage, but +/-40V is too high for 2N3055 anyway, and about 4x too much power for a single pair. A pair of transistors from the same batch is close enough and it is more important to match R4 and R7.
3. Good idea.
4. Again if anything, I would use a pair of BJTs to make a current source. The existing bootstrap cct has advantages. A VBE multiplier is a better bias circuit.
5. MJ15022's are not fast, and not new either. All "power base" transistors tend to suffer from turn-off delays. I think the point was what can be done with 2n3055s, and the correct answer is a classic quasi-complimentary amp up to 40W max, ie +/-20VDC rails max.
6. Correct, and kind or important.
More:
a. D5 and R17 are a bad idea, but the bias has to be adjusted if you remove them.
b. R10 and R20 are unnecessary if you pick the right values for R15 and R16, ie about 0.1 to 0.22 Ohms.
c. C1, C4, and the output RC network need a separate wire to the power supply ground.
d. An output isolation inductor + resistor in series with the output prevents the speaker wires from messing with the feedback and causing RF and instabilities.

nuff for now
Steve