MOSFET/MOSFET amplifier

[jarek]

Hi all,
I have in my mind little project with minimal number of active elements. I need some help and advise – I am not a professional person in this matter. I know that this forum concerns Pass Lab products mostly, but what I want to build is very close to design philosophy here. I started from simulations in Spice and got a schematic, which I would to introduce. I have problem with THD figures, are very depended on the level of input voltage. All is very good for single mV and much worst for hundreds of mV. This is before clipping at the output, of course. Is the design has a sense and is worth to try?

[jam]

Jarek,

Nice idea. I would replace R3 and R20 ( I hope Grey is not listening.... ) with current sources. The input fets cound be replaced with smaller devices (lower capacitance) . Looks like something I have been working on.

Regards,
Jam

[Nelson Pass]

Assuming you want some voltage gain, you need some
resistance to ground on your feedback loop, and you can
get it from R3 and R20 if you want. I would seriously consider
some caps across the Zeners, tho.

[markp]

What is your R17 value? In addition to Nelson's advice, try a lower R17 for more feedback.

[Upupa Epops]

Jarek, this circuit will be VERY temperature unstabil !

[Nelson Pass]

Maybe not so much depending on R3,4,7, and 20. Lots of
output stage sinking and high bias would improve that also.

[Upupa Epops]

I don't like amp, which auto - set up procedure have end after several hours . Imagine it by Space Shuttle !

[Upupa Epops]

P.S. : Zero temperature coefficient is by 240 on cca 15 A, I think .

[tschrama]

Hi,

I like this kind of topology and I've been fiddeling with it some time, but that was more than a year ago. I is kind of similar to the ' Simple MJ amplifier' you can find on the net and on this forum. I remember posting a schematic very similar to yours. I got a good comment about the problem of inducing rimple via feedback network: you have done it better than me with the zeners. I also suggest decoupling the zeners with a parralel cap to exclude the non-liniear contribution of the zeners and limit the noise from the zeners.

As for the Iq and Voffset: you could use higher values source resistors.. 0R5 ? Or use a simple single transistor servo around the source resistors.

Liniearity should be pretty good, atleast in class A ( 1.5 Ampere minimum). 2Vrms watt THD should be less than 0.5%, but this is directly related to the load impedance, due to the common source output stage.

Goodluck, and keep us posted.

[jarek]

Thanks for your interest and help.
After reading your comments I understood there were same mistakes in my conception. I tried mix different ways of implementation in one circuit. Well, computer is very helpful but it will not replace a human thinking
I was thinking it over and did two different versions of the schematic. Values of elements are only to start.
This is the first: