My HV opamp - Working suspiciously well

[bigwill]

I'm very new to solid state designs, and I'm attempting to design a high voltage amplifier to drive the grids of output tubes.

I whipped this up (see attachment) in LTSpice and it works, suspiciously well in fact, I'm getting readings of 0.009% distortion at 20Khz while swinging 100V P-P into the 10k load resistor, with the gain set to 100 (well, near enough)

Now, I'm not 100% trustworthy of this result, seeing as it seems pretty good, and I basically happened upon this design by accident / fiddling. For example, I'm not sure if my current mirror is actually doing anything useful, as it's driving the low impedance input of the VAS transistor.

Basically it would be really great if someone could look over this design for me and suggest improvements where possible

Thanks for any help

[Andrew Eckhardt]

What tube are you driving?

[bigwill]

Originally Posted by Andrew Eckhardt

What tube are you driving?

I don't know yet - most likely pentode mode EL34s or KT88s, I'm just trying to develop a circuit that can easily drive pretty much anything. It is probably overkill but it's an interesting excercise in itself

[Bigun]

what is the advantage in this approach, over say, an all-tube design ? - is it simply that you have lower distortion given all the nfb ?

It would be interesting to hear this circuit, if the distortion is really low it would allow the harmonic signature of the output tube to come through.

[DF96]

Don't forget that any circuit which relies on transistor balance (e.g. LTP, current mirror) will always simulate with lower distortion than in real life.

[bigwill]

Originally Posted by Bigun

what is the advantage in this approach, over say, an all-tube design ? - is it simply that you have lower distortion given all the nfb

Convenience mainly - it would be cool to have a simple universal circuit you can make up on a small PCB and have it drive nearly any valve!


Does anyone have any comment on the circuit itself? Am I doing anything stupid? Are there faster / better transistors I could use?

[Samuel Groner]

Does anyone have any comment on the circuit itself?

One could probably write a book on it--but you're not saying what your goals are. Distortion? Noise? Cost/complexity? Quiescent current? Bandwidth?

As a starter: a current mirror usually benefits greatly from emitter degeneration, e.g. to reduce noise and sensitivity to transistor mismatch.

Samuel

[bigwill]

Originally Posted by Samuel Groner

One could probably write a book on it--but you're not saying what your goals are. Distortion? Noise? Cost/complexity? Quiescent current? Bandwidth?

As a starter: a current mirror usually benefits greatly from emitter degeneration, e.g. to reduce noise and sensitivity to transistor mismatch.

Samuel

My goals would be wide bandwidth / high slew rate and low distortion - I like to fiddle with valve circuits it would be cool to build these as little gain modules you can really "lean" on and know they will perform well.

I added 100 ohm resistors to the current mirror and it increased the overall open loop gain somewhat and of course would have the positive effect you described of being less critical about transistor mismatch

[Markku]

forgive my ignorance but what are the parts that look like filter caps marked a V1 through V6 ? Are these d.c. batteries you are using in an active circuit ?

[hahfran]

Tubes are voltage amplifiers ( rather "charge change amplifiers) and BJTs are current amplifiers. The respective physics of amplification are entirely different.
It makes no sense to drive a voltage amp with a current amp you could easily replace the BJT high V with one or two tubes and stay so to speak in one family of amplification.