Largest possible drain resistor for a 2SK30?

[midwayfair]

I read somewhere ... somewhat recently ... something about the transductance of a FET determining the largest useable drain resistor. Unfortunately, I didn't bookmark it or put it in my notes, so I can't find anything about it.

Just a basic JFET amplifier a la the FETzer or the Omega, but running off higher voltage:

48V
|
R1
|
drain >>> Output buffer (DC coupled)
*
Gate
*
source
|
R2 (to set gain)
|
R3 || 47uF (extra resistor to drop the current draw bypassed by a cap e.g. Omega.)

The FETzer calculator calls for a 19K for R1 -- but I remembered seeing something about a 2SK30 or something like it not being able to use a drain resistor much above 10K, so I'm not totally sure that I can use R2/R1 for the approximate gain. Was that nonsense? Still applicable with the output buffer? Something else I should consider? Just use the smaller resistors because a few extra mA is peanuts?

[PRR]

Socket the resistor.

[R.G.]

FETs have limits on their current, voltage, and gain. Well, OK, every device does. I suspect that it's the drain voltage that limits the drain resistance size indirectly.

If you're trying for high gain, use a current-source load for very high impedance at low voltages. This runs up the AC gain without needing large voltages to get suitable DC current settings. The voltage swing on the drain is still limited, as it always is by power and voltage considerations, but you get much more gain - as long as you buffer the drain from any loading on the output.

[PRR]

> Was that nonsense?

Works for me.

You can work a JFET down to impossibly low current, so there is no "maximum size" for the drain resistor. (Except that at some point it can't drive your next-stage input well.)

The JFET's "plate resistance" shunts the drain resistor, but this scales with current. If your DC bias is reasonable, the shunting will be about the same for any reasonable current.
_____________

You say you have 48V and want 19K. The ded-short current will be 2.53mA. The happy bias current will be half, maybe 1/3 to 2/3 of that, or 1mA-1.5mA.

You DO need the "GR" sort from 2SK30. The R O Y sorts won't pass 2.5mA (for sure) thus won't pull-down 19K a full 48V.

The grounded-grid plate resistance of a 2SK30 around 1mA is over 500K. So a 19K loading is not much bothered.

The transconductance around 1mA will be near 0.0018 Mho (S) which is like 555 Ohms. Fully-bypassed no-load, voltage gain is near 34. Given about 20V peak output voltage, the peak input voltage is 0.58V or so. THD will be over 10%, all 2nd harmonic. With partial un-bypass, gain is lower and max input is higher.

I still think, if you can't estimate this on matchbook, sockets are your best friend.

[J0K3RX]

What Paul said...

You need the idss classification 2SK30ATM-GR

See "Note" at bottom of data sheet..