JFET biasing options and gain?

[ncusack]

Hello,

I've been looking into JFET based circuits lately and have been wondering about different biasing choices and JFET selection. I started with the Fetzer and have basically concluded that you require a low IDSS and Vgs to squeeze reasonable gain out of the normal Fetzer stage. Recently I came across Brian Marshall's Liquid Sunshine circuit. This design seems to stray from the fetzer idea but seems to achieve similar smooth sounding results. The Liquid Sunshine V2 schematic can be found online so i'll refrain from posting it. What I'm wondering about is the use of a 10k source and ~22k drain resistor with the 2n5458. My calculations for that Jfet show in a fetzer setup you'd need much smaller source and drain resistors to get the tube like sound and the gain would only be about 3dB where the LS Larger resistor setup with a bypass cap can hit ~20dB. Now I guess what I'm wondering is if anyone could elaborate on the differences in Jfet biasing vs tonal quality and possibly explain why Brian may have chosen to do things the way he did with the LS.

Regards,

Neill

[Groovenut]

Since the 10k source resistors are bypassed by capacitors, the AC gain is full from the 3dB (~1.1kHz first gain stage, ~50hz second gain stage) knee point of the source RC combo. However the DC bias is set by the 10k and is pretty close to cutoff. Any input signal over ~75mV will clip the amplifier stages. He's using both jfets in this design as cold clippers. The first clips the lower half of the waveform, causing a decent amount of even order harmonics, then inverts the output, the second clips the lower half of the waveform (unclipped by the first jfet) and inverts the output causing the final wave form to be clipped on both sides. It also appears he's using a negative feedback loop to restore some of the bass lost in the pre stage filtering. Cool design 

[ncusack]

Ok that explains the LS circuit. As far as biasing a JFET warm or cold goes is there anything that can be done to overcome the wide variation in part tolerance? I found someones stab at Brians MKIII version of the LS using J201 jfets and that seems to use a set voltage reference instead of the usual ground reference. Would something like that help or would you still need to hand pick your Jfets for each circuit?

[blackieNYC]

There's some helpful stuff in this recent thread maybe:
http://www.diystompboxes.com/smfforum/index.php?topic=111724.0
I think, and I could use some backup on this, the fetzer should not be thought of as a maximum performance FET boost, but an imitation of a fender tube input stage.  The low Vgs will get you earlier clipping in a j201 fetzer.