39k tube cathode resistor using Fets?

[william]

Speaking of fets emulating tubes, in the Slo100, Dual rectifier, and 5150 amps there is a 39k cathode resistor on the third tube.  What effect does that value have on the tubes.  And will a fet behave the same way.  

I've tried a Slo100 and DR emulation but at the third fet (the 39k one) the sound goes from a decent gain, to a harsh weak crackle.

[petemoore]

From what I've read, Jfets bias and sound like tubes. I'd follow the schematic R values for bias [after a bit, if no-one rebuts]. Should work.
 I was just typing about using a 47k trimpot where a schem [RM Axis] calls for 39k, A nice one to be able to adjust in the Axis's case.

[R.G.]

From what I've read, Jfets bias and sound like tubes.

That is a common fallacy. Think about it - if JFETs were a really acceptable sub for tubes, we (the music world) would not need to still be using fragile glass and metal constructions that we import. JFETs have been around for well over 40 years, about half the time we've had tubes. If they were good enough replacements, we wouldn't still be trying to emulate tubes.

The fallacy started when it was noticed that like tubes, JFETs are depletion devices - that is, they conduct heavily with no bias on the gate/grid, and you have to pull the gate/grid negative to turn them off.  That led to engineers telling one another that you use them like tubes - that is, by self biasing like a triode. This was a big issue to engineers that had spent their entire careers biasing tubes. So JFETs do bias like tubes - meaning "by the same general technique".

JFETs conduct tube by square law - that is, the current in the drain is proportional to the  square of the voltage between gate and source. Triodes are 3/2 power devices, in that the current is proportional to the 3/2 power of the applied grid/cathode voltage. JFETs clip somewhat softer than bipolars in similar circuits, but not the same as tubes.

Speaking of fets emulating tubes, in the Slo100, Dual rectifier, and 5150 amps there is a 39k cathode resistor on the third tube. What effect does that value have on the tubes. And will a fet behave the same way.

A 12AX7 is turned fully off by a voltage of maybe -2V grid to cathode. The purpose of the cathode resistor is to turn the current going through the plate-to-cathode into a voltage that reverse biases the grid. This is a negative feedback arrangement that settles  into a stable bias point depending on the amount of current that the tube conducts with zero volts grid to cathode - this would be Idss in a JFET - and the amount of voltage needed to turn the tube off - this would be Vgsoff in a JFET - it finds a stable point.

A 12AX7 will always be biased between 0V g-k and -2V g-k. The bigger the cathode resistor, the more back bias generated per milliampere of cathode current, and the smaller the eventual stable current it settles to.

In this case, with a 39K resistor there, you only need 2V/39K = 51uA to turn the tube off, so the eventual current is going to be less than that. So this tube is running on a tiny amount of current.

The current in the cathode is the same as the current in the plate, so if there's 1V across the 39K cathode resistor, there's 100K/39K times as much across the plate resistor, or a little less than three times as much. That means that the gain of this stage is also 100/39.

With the tube biased this near cutoff, it cuts off easily on negative signal peaks.  It also goes into positive grid voltage clippping on positive peaks with only about a volt of signal, so it clips both positive and negative peaks with signals of around a volt peak.

Will a JFET behave the same way? Yes and no. Vgsoff for a JFET varies from as little as -0.5V for some switching JFETs, notably the J201, to as much as -10V or more. Clearly, if you stick a -10V JFET in there, it will not bias the same, and won't sound the same. If you happen to get a JFET with a Vgsoff of about -1.5 to -2V, it might be similar because the voltages across the drain and source resistors are only a few volts in clipping, and the tube's ability to handle hundreds of volts doesn't count.

Will it sound the same? No.

Will it sound similar? Maybe. The Devil is in the details.

That harsh weak crackle is one of the reasonably expectable outcomes. You'd need to tinker with the specific JFET to get the biasing right. Unfortunately, tubes are much more uniform than JFETs.

[electrictabs]

well in my emus of these amps i use a 3k9 or 4k7 resistor instead of 39k
check this out
http://diystompboxes.com/sboxforum/viewtopic.php?p=195300#195300

[lovekraft0]

Unless I got this wrong, the whole point of that huge cathode resistor is to "mis-bias" the triode so that stage hard-clips at very low input levels. Couldn't you simply use the standard trimmer on the drain, but bias it similarly out of whack, or are you going to have to add a diode shunted to ground to simulate the triode's hard clipping?