Mosfet amp design i just dreamt up

[Tekko]

I just got this working in simulation:


But i wonder, is this my own design or did i just memorize someone elses schematic ?

Note that i did not look at any schematics while doing this one, i just tried randing things until it worked and similation showed reasonable currents through stuff.

[lineup]

Try to get equal current in Q8-Q9 pair.
Put AM-meter in their collectors and adjust R13 until balans.

Q11-Q12, BD139, BD140 is a good choice.
They take ~10 mA in this circuit.
And this is probably a good current level to drive those HEXFET.

Overall you have done a good job
May only be some small adjustments needed.
And should not be difficult to build, as it is very straight forward design.

[djoffe]

It's a good start. I'd recommend a few things:
1. current mirror load in the top of the Q8/Q9 diff pair would make automatic good balance and lower distortion.
2. is it oscillating, or is it a picture artifact? I can't quite tell...
3. You're leaving a fair amount of output swing unrealized by running both the drivers and the output devices on the same rails. The MOSFETS Vgs diminishes the available output swing considerably.
4. Emitter resistors in the input stage can sometimes be useful to increasing the linearity of the input stage...this can help lower distortion (you might have to increase the tail current a bit to compensate)
5. C3 can help stability, but usually costs you a little output distortion...still, the balance of that with Cdom (not shown, but typically from collector to base of Q2), gives nice trade off of stability and distortion.

Experiments:
1. Replace R10 by a short...I don't think that much would change...
2. What is your output offset voltage with no signal at the input?

Keep your investigations going...you are on a good track.

[Tekko]

Here it is with suggested changes:


And heres the frequency response:


I may try to build this one when the parts i've ordred arrives.

[Tekko]

This design have been built and apart from a misplaced resistor on the board, it worked on the first try. Unfortunately the misplaced resistor caused damage to the fets which died half a day later but after replacement to IRF540 and IRF9540 the amp was up and running again.

Heres a few pics:

[nigelwright7557]

There is no DC offset adjustment. Add preset into one of current mirror legs and put a resistor in the other leg.

Bias needs a preset to adjust it accurately and safely.

No need for driver transistors with 2 piars of output transistors. They will only add temperature drift to the bias level.

[Tekko]

A dc offset adjustor is not needed, theres 0.00mV on the output, i've measured it, it was the first thing i did before applying a real speaker.

[Tekko]

I was bored today and did this:

[sregor]

Originally Posted by Tekko

I just got this working in simulation:

But i wonder, is this my own design or did i just memorize someone elses schematic ?

Note that i did not look at any schematics while doing this one, i just tried randing things until it worked and similation showed reasonable currents through stuff.

Looks a sort of like Hafler 220 circuit. Hafler used better outputs (Hitachi's) and mirror LTP inputs. But lots of 2N5401s and 2N5550s my 2cents

[Tekko]

I pulled this design outta my head, tried random things in a simulator until it worked.

Though it does have a slight prollem, near clipping it oscillates slightly as it gets near the negative rail, perhaps a small cap between B and E of the lower driver transistor coulud remedy that.