JFet vs bipolar

[dimitri]

the main virtue is a _true_ galvanic isolation between console and dozens of effects. Inserting the effect in one line can cause hum and crosstalk in completely independent line. I was confronted with such problems. The explanation is on Jensen site:
http://www.reed-electronics.com/ednm...98/12df_03.htm
http://www.reed-electronics.com/ednm...98/12df_03.pdf

[lineup]

Originally posted by Asbjorn

I've been doing some listening tests of jfet vs bipolar transistors, and found some generel sound differences.

Jfets tends to sound softly, while bipolar sounds more dynamic.

Okay.
I think this is a myth.
I did some comparing spice tests with differential inputs.
Lownoise JFETs vs. Lownoise BJT.

Note: My test used ONLY input pair, with collector resistors.
They were tuned to do exactly same AC current job.

for example: 5 uA current AC output across 1 kOhm resistors
or 5 /50 mV ac across same resistors.

This means
- open loop
- one transistor base/gate tied to GND
- the other fead with a few mVolt input

I did tests without Emitter/source resistors.
In this case I adjusted mV input, to do same job across collector load
( BJT had like 4-5 higher gain than JFET. uA AC current / mVolt input )

The THD harmonics ( 2nd + 3rd ) was at very similar level of magnitude.
At same job done.

The only significant difference was that JFET had a bit higher 3rd harm dist, than 2nd harmonics.
While BJT, bipolar had the other way: Like 7-8 dB higher 2nd harmonic distortion, than 3rd

This might come as surprise.
But actually this was so in my test.

I also tested with same input level: 1mV / 5 mV
In this case I adjusted with Emitter/Source resistors
to get equal job done.
Of course the bipolar in this case can afford more resistive degeneration
as they have like 4-5 higher AC current output per milliVolt AC in.

This gave like 12 dB lower level distortion than JFET pair. ( -149 dB vs. -137 dB )
At same level in, same collector job done.

=======================

My Conclusions:

All else equal,
- Lownoise JFET ( 2SK170 ) have higher 3rd harm. distortion, by their nature, than bipolar lownoise

- For lowest 3rd harm. distortion + overall THD, use bipolar input differential
- Or use JFET while letting them perform an easier job
+ use some Source degeneration RESISTORS for them

Remember, not only low THD is a positive amplifier feature.
Also we have low bias current and DC Output offset to consider, many times.

Lineup
----------------------------

Attachment.
2SC2240, The Fourier result.

[richwalters]

You will find a thesis on exactly the same subject Audio Frequency Noise Characteristics of JFets AN74-4 by Siliconix. JFet design catalogue 1983! Also composite Op-Amp for high performance AN81-3.
Unfortunately both no longer in print.

I have no software means for copying/ sending both articles.

richj