One more JFET sensation from Russia

[456Onno456]

Hi folks,

i have a few question,

just for the understanding (i think the transaltors are worse for russian to german then to english).  If you look @ the first too schems in the post, my opinion is, that they show the "evolution" of the jfet biasing. The thing changed is mainly the 0.4-0.5V positive bias @ the gates. This voltage compensates a part of the positive voltage over Rsource (one should measure the get the exact value). This leads to the hard clipping of the positive half wave and the softclipping of the negative one. Am i wrong with the understanding of those circuits?

The things i don't understand at all:

- All of the schems show relatively large sourcecaps, that means the negative feedback applied is equaled out for most of the frequency (smaller caps), or all of the frequency. I thought, that the magic thing about jfet-triode-emulation ist the ^1.5 transfer function that you achieve with a moderate negative feedback (for ref.: see Dimitriy's paper). The main problem i see is that each stage is amplifying the whole frequency range with a transfer function of ^2. Again the magic thing of an tube amplifier is really sensible biasing of each gain stage. Not only amplification-wise, e.g. the cathode resistor (changing the transfer function and the bias) but also frequency wise (with a cathode bypass cap). If you are talking about those high-gain amps with multiple cascaded gain stages it doesn't matter that much, but if you roll back the gain and volume on the guitar not each stage will be driven into full clipping. That results in a contribution of different harmonic content at each stage.

- The original circuits are consisting of just cathode-biased triodes. There are very few guitaramps out there that have a constant bias (doesn't matter if by diode, or voltage reference). To my knowledge there is a difference in the transfer function, if you bias by resistor (cathode-biased), or constant value (refvoltage).

i don't want to insult someone (mind the language-barrier - as you easily will see i'm not a native speaker), it's just for the understanding of this circuit topology.

Max

[Caferacernoc]

 "All of the schems show relatively large sourcecaps, that means the negative feedback applied is equaled out for most of the frequency (smaller caps), or all of the frequency. I thought, that the magic thing about jfet-triode-emulation ist the ^1.5 transfer function that you achieve with a moderate negative feedback (for ref.: see Dimitriy's paper). The main problem i see is that each stage is amplifying the whole frequency range with a transfer function of ^2. Again the magic thing of an tube amplifier is really sensible biasing of each gain stage. Not only amplification-wise, e.g. the cathode resistor (changing the transfer function and the bias) but also frequency wise (with a cathode bypass cap). If you are talking about those high-gain amps with multiple cascaded gain stages it doesn't matter that much, but if you roll back the gain and volume on the guitar not each stage will be driven into full clipping. That results in a contribution of different harmonic content at each stage."

Agreed. I have tried almost every jfet configuration there is. And I like them better without the source cap.

"- The original circuits are consisting of just cathode-biased triodes. There are very few guitaramps out there that have a constant bias (doesn't matter if by diode, or voltage reference). To my knowledge there is a difference in the transfer function, if you bias by resistor (cathode-biased), or constant value (refvoltage)."

I have not tried refvoltage. I'll have to look into that. What I do like to do is vary my fixed bias in a cascaded circuit. If I have 3 jfet gain stages, what I'll do is have the first one at 6.5 volts. The middle at like 5.5. And the last one at 4.5 or 5 volts. (9 volt battery)