Hybrid dreams

[hidnplayr]

I've been thinkering about a hybrid amplifier with mosfets in the output stage..
I didnt know where to post this at first, but since the total amp will be hybrid, and the schematic looks a lot like a tube push-pull amp, i decided to post here.

I dont know if this has been done before but the idea is to make a push-pull version of the parafeed.
Simply put, take 2x the same parafeed and feed it a differential signal.
Now, I believe the 2 chokes can be wound as a push-pull output transformer, on the same core. (For the benefit of having 1 small core instead of 2 big ones, and maybe other things too!)
You could now simply add a secondary winding to the core, and connect the speaker to that, but i'd prefer not to!
(simpler to wind the core, no additional losses,..)

Only thing i want to do now is get rid of the output capacitor (A servo?)

What do you guys think? Am i insane?

 

[smoking-amp]

Looks like the output cap could be deleted if the two windings have equal and low DC resistance, and the bias and balance pots are adjusted correctly. Best to try with a dummy load 1st. The speaker will be common mode floating at +V though.

 

[hidnplayr]

So, perhaps I could add a circuit that disconnects the speaker when there is DC at the output terminals, instead of using a servo. That sound reasonable.
The speaker floating at +V, do you see it as a bad thing? Should I 'move' the ground ? (and use all negative supplys)

Thanks for your input.

 

[smoking-amp]

If someone else were to use the amplifier, they might get into problems with the speaker floating at +V. Say they try connecting the output to the input of something else. Or accidentally short (or intentionally ground) a speaker lead. Could put fuses in I guess.

 

[Wavebourn]

Thermal stability issue has to be addressed properly.

 

[hidnplayr]

Thermal stability issue has to be addressed properly.

Do you mean I should prevent 'runaway' by placing a temperature dependent resistor on the coolblock, and use it in the bias circuit?

 

[Wavebourn]

Do you mean I should prevent 'runaway' by placing a temperature dependent resistor on the coolblock, and use it in the bias circuit?

I don't know. Ideal compensation can't be achieved, but to get satisfactory compensation there are many ways. One of them is temperature dependent resistor. The best AFAIK is feedback by idle current.

 

[hidnplayr]

Sounds like the best sollution is to use an opamp in the bias circuit then.
Since most of them have 2 in a single chip, I could use one per output transistor, and tackle the DC problem too!

This is going to be a bit more complicated then I 'dreamed', but I'm going to build it anyway, it'd better sound good!