Simple(ish) discrete op-amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
amplifierguru said:
...You should maybe build a Q=2 bandpass filter to clean up that soundcard output - it works wonders on the noise too!..
A bandpass filter - I should have thought of that. It would still leave noise picked up by the input, but it would help improve the resolution of my measurements, which is always good.


amplifierguru said:
...Didn't mean to be over the top with earlier comments. I get a little passionate...
Don't worry about that, I know what forums make people do. I am curious however: Is there a reason you think PSRR would be poor? As I see it, current in the input stage is determined by the input FETs and the voltage across them is pretty much independent of supply voltage; the current in the Vas depends on the current mirror, which again sees a constant voltage, so I can't see anywhere in particular that noise could be injected; any insight is always welcome.


Workhorse said:
Hi D-Evil,

How it sounds.......well or unwell

I think its very good sounding piece of devilfier:D ;)
mlloyd1 said:
...so, have you listened yet? how does it sound?
Yes I have listened to it. I did a few (sighted) comparisons to various op-amps. To be honest I really can't tell the difference, and I generally can't with circuits that are 'good enough'. The only detectable difference is the noise level, which differs both in level and character between op-amps (my circuit is not the best for noise, but not the worst either).


mlloyd1 said:
...based on a hint from a post in "the other place" by andy_c, i have been looking at some other jfets to use in those areas when i need higher idss than provided by the 2SK170 and 2SJ74. i think the part numbers are J175 (p) and J111 (n), but i have to check at home...
I have noticed that there are a few close-enough pairs among the many Jxxx JFETs, but they aren't cheap enough compared to the Toshiba parts in the UK, plus they aren't binned by Idss, so it's probably more hit-and-miss getting suitable ones for a given circuit. Very high Idss though... maybe high enough for small power-amp output stages?
 
Hi Evil,

Re the soundcard - I don't have an image of the output alone, but I do have one of my earlier amp. They're the same to all intents and purposes. Before I built the filter it had -86dB HD2 and -96dB floor. See pic.


regarding the high transconductance FET's - their drain impedance is quite low, evidenced by their rel poor current source ability c.f. bipolars. So the impedance divider this makes to the supply line determines the PSRR. For preamps good low Z low noise regs along the lines of Jan Didden's (TAA 1/87) render it insignificant. But ideally, or for power amplifiers, a bipolar cascode is my preference.

My folded cascode comp JFET preamp design for Magnet used no cascode but good chip regs and low AC load draw.


Cheers,
Greg
 

Attachments

  • eidetic gb2b thd+n residual.png
    eidetic gb2b thd+n residual.png
    7.3 KB · Views: 309
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.