FET input stage?

[slowferrari]

Looking at options for a FET input gain stage. It looks like there are a couple of ways to approach it. I've seen discussion of using high voltage DMOS to replace a triode, but it seems like the Dumble approach is to derive a low voltage from the preamp plate supply, so as to use what would eventually be called a "fetzer valve".

I'm wondering about the relative advantages and disadvantages of these two options.

Also wondering whether it's possible to use something like an MPF960 (max Vds is about 60v) to use something like a ZVEX SHO as the front end. It's a very different approach than either DMOS or JFETs, but if I can derive a low voltage to run it off of, why not?

The reason being I have a bunch of similar MOSFETs around, and the THT versions of both the JFET for the dumble input and the LND150s typically recommended for triode replacement are getting a little hard to find, unless I'm just looking in the wrong places.

Either way, curious if anyone's got advice on using a FET front end in general.

[Keppy]

Pedal guy here.


Both JFETs and MOSFETs are capable of very high input impedances, so you'll have no problem getting the usual 1MΩ input impedance. JFETs are more commonly substituted for triodes due to some similarities regarding device gain, biasing, and harmonic distortion. In many cases, you can build a triode stage as a low voltage JFET stage with no other substitutions.


If you use a low voltage supply, you might get a bit of distortion/color in the input stage. With a high voltage supply I think you're less likely to notice the differences at a similar level of gain.


The suitability of the SHO stage depends on how you go about it. The gain control on that effect causes crackling, so I personally wouldn't use it in an amp without replacing that control with a fixed resistor.


In the pedal world, MOSFET boosters (especially the SHO) tend to be capable of more gain than JFET boosters. Obviously this depends on a lot more than just the device, but if you want to drive the following stage hard, a MOSFET is probably the way to go. To simulate a normal triode input stage, either has enough gain.


Note the protection diodes on the MOSFETs. They are static sensitive, and an input stage will subject them to stress when plugging in. That doesn't mean don't use them, it just means don't skip the diode.


The high voltage MOSFET schem seems suspect to me. I see no biasing for the MOSFET gate. Common MOSFETs need the gate biased a few volts higher than the source to operate linearly, which is different than JFETs and triodes.


When using a MOSFET you may experience issues when bypassing the source resistor with a cap like in the high voltage schem, because the frequencies passed by that capacitor are boosted as much as the device can boost them, and a MOSFET is capable of more gain than your average triode/JFET. A source-bypassed MOSFET stage is probably not what you want IMO.


Basically, if you just want it to work, use one of the existing JFET designs. If you want to experiment, read this: http://www.geofex.com/Article_Folders/mosboost/mosboost.htm