FET Vgs Question (source to gate-->cathode to grid voltage)

[mac]

Typically the cathode of a preamp tube is at 1v - 1.5v.
In a high voltage FET simulation like those at ROG, is it better to use a FET with a Vgs value close to 1v to 1.5v?
In other words, does Vgs has any impact on the sound?

mac

[defaced]

Assuming cathode bias, it affects the bias point of the device.  It is device dependent and is not the same for all devices.  Yes, Vgs affects the sound.  This is how cold clipping stages in high gain tube amps do their dirty work. 

[PRR]

> Typically the cathode of a preamp tube is at 1v - 1.5v.

Because the turn-off voltage of a vacuum triode is about Vpk/Mu, we use tubes with Mu near 100 and plate voltages around 200V. Turn-ON is zero Vgk, turn-off is about 2Vgk, the "in-between" is around 1V.

Why do we do this? Input overload. Hot hard-strummed pickups can exceed 0.5V peak. In conventional e-guitar, preamp distortion should be small (output overload is used for emphasis). A 0.5V signal on a tube biased near 1V gives about 2%-3% THD mostly pure 2nd, which is inoffensive and hardly audible (but may play a role in guitar sound).

> use a FET with a Vgs value close to 1v to 1.5v?

A JFET's turn-off voltage is not usefully affected by the channel voltage. Vto is "built in". Datasheets often call this "Vgs(off)". Production variations lead to a wide range, such as 0.5V to 10V. JFETs are generally available "sorted" into wide Vto bins:

VGS(off) Gate-Source Cutoff Voltage
VDS = 20V, ID = 10nA
J201   -0.3    -1.5
J202   -0.8    -4   
J203   -2      -10 

A simple theory would suggest picking a Vto somewhat larger than peak-to-peak pickup voltage, similar to a tube preamp, for the same reason: making full use of device capability.

However:

The I/V curve of JFETs is NOT the same as a tube, square-root instead of three-half power. Working a JFET to the same % of cutoff voltage will not give identical distortion.

A vacuum triode has significant (hidden) negative feedback plate to grid; negligible in a JFET.

"Plate voltage" of common JFETs is VERY limited compared to vacuum triodes. "High Voltage" is 24V-40V, a tenth of the 200V-400V supply used on tube amps. JFET amps typically clip their outputs before their inputs.

JFET parameter variation is FAR greater than 12AX7. The AX holds Mu within 10%, Gm within 20%, a 1:1.2 or 1:1.4 spread, for 95% of tubes made over 60 years in dozens of places. Sorted JFETs cover 1:5 spread. Moreover of the zillion part-numbers ever made, your favorite supplier has a limited selection and often not quite what you would prefer.

JFET design is a lot about adjusting (preferably self-adjusting) the circuit to the JFETs you actually get.

JFETs are very much cheaper than 12AX7. A wise design does not ask for the very-most from each device (as we do in tubes), because a "slack" design can always use an added stage, often cheaper than hand-sorting or trimming loose-spec devices. 

But sure. Sort some JFETs as ~~0.8V, ~~1.5V, and ~~3V lots. Try in various circuits with various pickups, players, styles.

[mac]

When I was experimenting with a fet version of the Valve Jr preamp running at 45V I noticed differences between 2sk117 (0.4v), 2sk246 (1.2v), mpf102 (3v) and bf245a (0.9v).
The 117s have more gain than the others, they seemed to clip harder than the rest. Not bad for a metal pedal. The 246s were more realistic.

Now I want to try 2sk373, 100v fets.

A vacuum triode has significant (hidden) negative feedback plate to grid; negligible in a JFET.

Looks like the gate is almost isolated from the drain and source.

mac