P3A customisations

[jaycee]

This being my first dabble into the foray of amplifiers (well, beyond chip amps!) I've studied numerous other peoples successes and failiures with the P3A before proceeding myself, and I've discovered a lot of people chose to twiddle the design somewhat.

So far I've made two twiddles myself, adding an extra pair of output devices (4821/4302) and using an MJE340 for the bias device (so that it will couple with a heatsink better).

I'm wondering if anyone else would like to comment on changes and twiddles they've made to their own P3A's, like changes in the input stage, or using a current sink instead of the bootstrap principle, or other things like that, and whether they've improved reliability or sound quality ?

[paulb]

Mine's strictly stock but you may want to post on the ESP discussion board as well, if you haven't already.
http://www.instantboard.com/users/rode/

[millwood]

I have done more than that,

the one I am settling on has little resemblence to p3a: it has pnp differential pairs loaded with resistors, a mosfet vas, bootstrapped, driving a pair of power mosfets.

my biggest "objection" to the p3a is the use of cfp output stages. I have not had great success getting mosfets cfps to work and I have to have mosfets as output devices.

[slowhands]

Originally posted by jaycee
.....I'm wondering if anyone else would like to comment on changes and twiddles they've made to their own P3A's, like changes in the input stage, or using a current sink instead of the bootstrap principle, or other things like that, and whether they've improved reliability or sound quality ?

1. BC 550C for the input pair (lower offset and noise). (At low voltage <45V, the MPSA18 is even better, but watch the pinout).
2. Current mirror collector loads for the input pair collectors (lower distortion).
3. Current sink for the VAS (no turn on thump, better bass).
4. Possibly lighten the lag compensation and add feedforward.
5. Baker clamps on VAS and CS to reduce "sticking" when overdriven, which will lead to cross conduction (fatal, especially if you use MOSFETs in CF configuration).

All it takes is a Dremel tool and 1 mm carbide bits to put a few more holes in this board to accept the added components. If you are neat, you can't even tell it was modded.

[slowhands]

Originally posted by jaycee
.....I'm wondering if anyone else would like to comment on changes and twiddles they've made to their own P3A's, like changes in the input stage, or using a current sink instead of the bootstrap principle, or other things like that, and whether they've improved reliability or sound quality ?

1. BC 550C for the input pair (lower offset and noise). (At low voltage <45V, the MPSA18 is even better, but watch the pinout).
2. Current mirror collector loads for the input pair collectors (lower distortion).
3. Current sink for the VAS (no turn on thump, better bass).
4. Possibly lighten the lag compensation and add feedforward.
5. Baker clamps on VAS and CS to reduce "sticking" when overdriven, which will lead to cross conduction (fatal, especially if you use MOSFETs in CF configuration).

All it takes is a Dremel tool and 1 mm carbide bits to put a few more holes in this board to accept the added components. If you are neat, you can't even tell it was modded.

[jaycee]

slowhands, take it you've done most of these things yourself I'd be very interested to see exactly what changes you made.

I've been looking at the schematic Kilowatt pasted at some point, and am debating adding the current mirror in the LTP. I've also been looking at the Ampslab designs. Basically I'm still learning my way around amplifier design (without books - not got the cash )

Pinouts and such don't matter as I'm cooking my own pcb. This is mainly for the experience of doing it, as it'd work out cheaper for me to buy Rod's board, but i felt the experience of doing it "by hand" was too good to miss out on.