John Curl's Blowtorch preamplifier part II

[Salas]

Originally posted by syn08
It might be semantical

Like in this one. But the jfet is the load in this case. Has a cascode and a buffer, and it does not oscillate, its a working example.

[Salas]

And here is its simulated Zout. We are lazy DIY hobbyists Iko and me, not engineers by far, with sad oscopes, so no lab tests.

[Salas]

But we believe in Janneman's local current loop benefit and why shunts can do better bcs of that if near the load thing. So when far, remote sensing! Better than nothing.

*A remote sensing example of the simpler non cascoded and non buffered version.

[KSTR]

Salas,

your one qualifies as global NFB IMHO, as we now have a "real" error amp: (cascoded) VAS plus bjt-buffered source follower.

EDIT: BTW, only one Vbe accross J5+R21?

- klaus

[syn08]

Originally posted by KSTR
But then again (we are refering to this, neg side, are we?) the JFET drain is not involved in the circuit at all as could be hooked to any potential to make the FET work with no change in circuit behavior (EDIT: and same thing for the collector, except for the return current path we want in the first place)? But we can leave it at that, it's not worth any hair-splitting IMHO.

Yes; I assumed there are "dots" along the ground wire.

I agree, this is somehow hair splitting. To me, even the notion of "regulator" assumes somehow there's something to return the error information and doing some corrective action to minimize the error, therefore feedback.

[Salas]

Originally posted by KSTR
Salas,

your one qualifies as global NFB IMHO, as we now have a "real" error amp: (cascoded) VAS plus bjt-buffered source follower.

- klaus

Yes indeed. I have shown it after syn08's loop explanations rang a bell.

[janneman]

Originally posted by KSTR
No, check the polarities of the MOSFETs and the way they are hooked up (ignoring the source resistors for a moment). Best seen on the neg side, the source follows the drop on that R to gnd, impessed by the JFET ccs.

- Klaus

Ahh I see, constant Vgd instead of Vgs. Hmmm.

Jan Didden

[Salas]

Originally posted by KSTR


EDIT: BTW, only one Vbe accross J5+R21?

- klaus

Yes just because J5 is a very low pinch off part. Vbe can be lifted higher with a resistor under the lower cascode bjt to ground if needed. It works as it is with 2SK170. (We had many of them left out from matching in an open loop Riaa stage).

[syn08]

Originally posted by Salas


Yes just because J5 is a very low pinch off part. Vbe can be lifted higher with a resistor under the lower cascode bjt to ground if needed. It works as it is with 2SK170. (We had many of them left out from matching in an open loop Riaa stage).

You are learning fast N channel JFETs have negative pinch voltages; depending on transconductance as well, they can reach incipient saturation at very low Vds. The rule of thumb is: two pinch voltages away from the (negative) pinch voltage is good enough. That is, for 2SK170 parts, Vt=-0.5V => Vds=0.5V is already good enough.

[metalman]

Originally posted by john curl
I like common mode chokes, hi H, low R, with external R damping.

I get the high H, low R, but could you expand on what you mean by external R damping.

Thanks,
Terry