Borbely Cascode

[Lumba Ogir]

the dynamic impedance of current sources should be high, ideally infinite at all frequencies, opposite to voltage references, ideally zero.
High ripple rejection is achieved by that combination. Low Cob of the cascoding device helps further in this regard at high frequencies.
Zener diodes are nasty noise generators with ugly characteristics causing an unpleasant kind of distortion and should not be used in audio as voltage reference.

 

[Lumba Ogir]

Borbely`s input stage would do a better job as current source providing very high impedance and bandwidth.

 

[Onvinyl]

Hi Lumba,
I'm thrilled by the fact that you propagate the 431 shunt reg instead of a zener. I was pleased (sonically) with it where I tried it, but went away from it to get noise lower. The noise penalty could well come from elsewhere, though.

Hi roender,
it's
here

Note that is is a work in progress and might be far from optimum. I'm not so sure that it really qualifies as a CFP-FC, but anyway, the (simed) results are better than a single fet or bjt as cascode and better then a conventionally (further) cascoded FC fet or bjt.
Not build so far, though.

edit: better means in this case, more approriate square wave (if it says anything in a sim), (slightly) extended bandwith and more gain.

Rüdiger

 

[Lumba Ogir]

Rüdiger,
I am using IC voltage references but please do select the parts you like.

 

[CBS240]

....Zener diodes are nasty noise generators with ugly characteristics causing an unpleasant kind of distortion and should not be used in audio as voltage reference.

I agree, using the cascade CS circuit in post 22 and a resistor||small cap creates a much more quiet reference voltage. Maybe a few more parts, but worth it IMO.

 

[MikeBettinger]

I prefer fets, but I use bipolars as well in my less expensive designs. I don't think that it makes much difference.

Sonically I have experienced excellent results with bipolars. Years go, after reading Borbely's FET suggestion, I wondered if I was missing another nth degree of resolution. I ordered his suggested FETs to give it a try (not from him), they sent me GRs and Vs as complements... They're still sitting in a drawer and I'm still waiting for the wheel to come back around. Too many other things to experiment with since then.

I'm not reading any "you gotta hear this" tone to this thread.

Good discussion though. Mike