Amp Schematics and Write-up

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Thanks.

Good luck with your design - I am working on another two designs now but don't expect to have these up for the next few months - got to order boards and assemble and test. V. busy at work right now as well.

I do notice you use the VAS cascode but not a hawksford and with fully complimentary designs you have to manually bias the inverting input for the offset. I am surprised you could not get very low thd from a design such as this like you spoke of.
OS
 
Member
Joined 2003
Paid Member
Yes, this interation is definitely overcompensated and there are opportunities to 'loosen up' a bit up. The loop gain at 1KHz is about 50db and at 20KHz around 25dB. I have written it up as it is.
Re the Hawksford VAS, I did some simulation work a while back and this gives about a 20dB improvement at 20KHz. I think 50ppm at 20KHz is acheivable with resistive loaded LTP. If you want to get to single digit ppm with this type of design (and retain the TIM free/non slewing features), you need to go to mirror loaded LTP's, but then you have to address the CMCL problem - I tried it during the breadboarding stages of this amp, and eventually ditched the mirror load - I did not like the additional complexity at the time.
 
Last edited:
Yes, this interation is definitely overcompensated and there are opportunities to 'loosen up' a bit up. The loop gain at 1KHz is about 50db and at 20KHz around 25dB. I have written it up as it is.
Re the Hawksford VAS, I did some simulation work a while back and this gives about a 20dB improvement at 20KHz. I think 50ppm at 20KHz is acheivable with resistive loaded LTP. If you want to get to single digit ppm with this type of design (and retain the TIM free/non slewing features), you need to go to mirror loaded LTP's, but then you have to address the CMCL problem - I tried it during the breadboarding stages of this amp, and eventually ditched the mirror load - I did not like the additional complexity at the time.

Mirror loaded LTP's ... overkill.. 110db , you have to degenerate everything. Higher Hfe devices , a little beta enhancement (EF) VAS ,good to go. With a hawksford / EF , even with an FET differential ...65db at 1k and 25 at 20khz. Who needs single digit PPM ? Under 500 is good enough for me.:D
OS
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.