I started with 2 x 9v batteries and a charging circuit, but it was more trouble than it was worth. Sound was good but unremarkable. Then I tried Shiny and after ironing out an oscillation problem, it was quite nice. Since then I have almost lost track of all the improvements added, but I can report that each was audible. I have LOTS of BIG caps (CRCRCRC); a choke; an LM317 (with black Gate caps); a "Pink Fish" gyrator; the Shiny reg; some more caps, including paper in oil. I have also made some mods to the basic JBOZ with a CCS on the source, another 2sk147 on the drain and adjustable NFB. I don't think I can do much more to imp[rove on it. It is very clear and quite dynamic, (though not "knock your socks off"). It does sound like solid state, though not in an objectionable way. I am now heading into tube territory for a comparison.
hihopes said:
One of the reasons for considering this circuit is that it has been described as transparent and "live", both parameters that are of the utmost importance for me.
Could you draw your circuit? I think it is one of the most refined among the people that posted in this thread.
A PM/email would be fine as well if you dont want to post it publically.
Which are your source and amp?
Haven't got time or software for that right now. Will try to do something soon.
Source Pioneer DV300-k (modified); Zero DAC (highly modified). Amps:- Geoff McCauley's Hybrid design (highly modified) for mids and tweeters; Hart Linsley Hood (modified) kit amp for bass. Speakers - Scanspeak drivers in homebuilt cabs. Home-built speaker, interconnect and power cables (Except for one pair of Van Den Hul Silver interconnects).
There is no hum or hiss right next to the speakers and the Jfet's do sound cleaner and clearer to me than BJT's. This was also apparent when I built the Pink Fish gyrator (also called "Teddyreg") and the Darlington in it was replaced with a "Fetlington". However, Jfets do seem to emphasize treble and bass a bit at the expense of midrange.
System sounds very good. Neutral, natural and quite emotionally engaging. It is this latter point (+ dynamics) that I am looking to improve upon. Perhaps more difficult to get right than clarity.
Source Pioneer DV300-k (modified); Zero DAC (highly modified). Amps:- Geoff McCauley's Hybrid design (highly modified) for mids and tweeters; Hart Linsley Hood (modified) kit amp for bass. Speakers - Scanspeak drivers in homebuilt cabs. Home-built speaker, interconnect and power cables (Except for one pair of Van Den Hul Silver interconnects).
There is no hum or hiss right next to the speakers and the Jfet's do sound cleaner and clearer to me than BJT's. This was also apparent when I built the Pink Fish gyrator (also called "Teddyreg") and the Darlington in it was replaced with a "Fetlington". However, Jfets do seem to emphasize treble and bass a bit at the expense of midrange.
System sounds very good. Neutral, natural and quite emotionally engaging. It is this latter point (+ dynamics) that I am looking to improve upon. Perhaps more difficult to get right than clarity.
I tried the original circuit. Distortion was as reported in the previous posted graphs. Not exactly a viola! experience - Worse than a 6SN7 tube line stage.
I tried degeneration - not much help.
Then tried the j310 - Eureka, Washington!
It is tweaky, but after many hours pushing the FFT for results. what emerged was a clear picture - this JFET can rock as a naked single 1 trick pony gain device.
The story is simple - name your gain, and name your voltage. The more clean gain needed, the more voltage needed.
What is clean? how about 3x gain, 1 V rms out @ .02% -
no feedback, no current sources, 25V supply. But you want drive? direct couple to a B1, where the gain stage output biases the buffer input at 1/2 the supply voltage. No caps on input or interstage. Details at 11
Bob
I tried degeneration - not much help.
Then tried the j310 - Eureka, Washington!
It is tweaky, but after many hours pushing the FFT for results. what emerged was a clear picture - this JFET can rock as a naked single 1 trick pony gain device.
The story is simple - name your gain, and name your voltage. The more clean gain needed, the more voltage needed.
What is clean? how about 3x gain, 1 V rms out @ .02% -
no feedback, no current sources, 25V supply. But you want drive? direct couple to a B1, where the gain stage output biases the buffer input at 1/2 the supply voltage. No caps on input or interstage. Details at 11
Bob
