A few weeks ago, I played around with a discrete JFET bass preamp design I found online (http://www.albertkreuzer.com/preamp.htm).
One thing I realized from analyzing the circuit, running simulations and eventually building the amplifier was that performance, and even actual operation in some cases, relied heavily on the characteristics of the specific transistors used.
In my research, I found that all JFET manufacturers provide only crude tolerances for specs like VgsOff or Isat. As an exercise, I tried to come up with an amplifier design that would deliver a specific gain over the whole range of spec, but failed. JFETs just vary too much, and a given circuit could very likely break if a specific transistor had to be replaced. In my case, I ended up having to test several transistors and adapt the design to the resulting specific specs.
A good production design wouldn't require this kind of cherry picking. Have I missed anything ? Is there a good way to design a circuit with transistors that have big tolerance on key specs ?
Best Answer
Welcome to the world of negative feedback.
Negative feedback achieves defined gain by providing vastly more than is needed (notionally infinite) and then feeding back output signal to input in antiphase in such a way that when the required gain is achieved the gain is stable.
Searching for something like
amplifier feedback
of jfet amplifier feedback
will find much quality material (and much rubbish :-) )
A reasonable introduction here - The Common Source JFET Amplifier . Cct below is from that page.
FET small signal analysis 50 page slieshow with perhaps enough detail to be useful coutesy of the dreaed DoCStock. EE dept of KUKTEM (Malaysia) - may be findable elsewhere.
Useful - mainly video amps but some good material.
Lots of starter material here
SPICE simulations
May be useful - Circuit workbook with basic solutions Question 5 relates. Others potentially useful.