Substitute JFET buffers with MOSFET's.

[gjcamann]

So I've seen several pedals where JFET's are used as buffers. Like the Paisley Drive for instance. Since these aren't used for tone shaping, the JFET was probably chosen because of low noise, and better impedance, and maybe cost at the time the pedal was originally designed (and perhaps some barely perceivable impact on tone). But with JFET's being the way they are, I would like to substitute a MOSFET - which also has good noise and impedance characterstics.

I've read the great info here  http://www.muzique.com/lab/buffers.htm about MOSFET buffers, and high frequency roll off is mentioned, but doesn't seem to be an issue for a 2N7000 with 20pF.

Does anyone have any other sage advice on replacing JFET's with MOSFET's?

Am I wrong about the "barely perceivable impact on tone" assumption?

Thanks!

[amptramp]

JFET's are always depletion-mode devices meaning thay need a bias in the direction that tends to turn them off to work.  MOSFET's can be had as depletion- or enhancement-mode devices but the overwhelming majority of them are enhancement-mode which means they need bias in the direction that turns them on.  For n-channel devices, JFET's need negative bias and MOSFET's need positive bias.  A Tillman booster is biased with a source resistor, whereas there is no comparable design with a MOSFET - it needs a divider string from Vcc to Gnd to set the correct bias.  You can design with either one, but you could not just remove a JFET from a circuit and replace it with a MOSFET.  JFET and MOSFET capacitances are comparable for devices with the same current and voltage handling ability and the same gain.  The capacitances are larger than tube capacitances and change with operating point, so you can have some capacitance-induced variations in waveform purity as well as a limitation on resistances you can use in the circuit, but for normal stage gains and drain resistances, this can be kept to a negligible level.  This means lower gate and drain resistances than would be needed to set the DC operating point.

If you have a problem with Miller capacitance (gate to drain capacitance that is reflected at the input as Cdg capacitance x (stage gain + 1)), go with a cascode stage where the drain voltage is held nearly constant and Cdg is just added to the gate to source capacitance Cgs.

[defaced]

Most of my MOSFET experience is in high voltage amp circuits.  Using RG's MOSFET Follies as a guide: http://www.geofex.com/Article_Folders/mosfet_folly/mosfetfolly.htm
He makes the point that the gate/source capacitance is not much of an issue when configured as a follower.  I've used MOSFETs as followers in non-tonal stages of amps with success (IRF820 because I have a good number of them even though other parts may be better suited to the task).  Does this same trend follow in low voltage circuits with typically used MOSFETs?

[gjcamann]

Thanks for the replies.

To be clear, the application of this input/output buffering is as part of a larger effect circuit. Not boosting.

Here is a link to a Paisley Drive schematic with JFET buffers, the buffer topology looks very similar to the MOSFET buffer from AMZ.
http://www.sabrotone.com/?p=3177
http://www.muzique.com/lab/buffers.htm

(I would link right to the images, but want to respect the original owner's rights)

[PRR]

> the Paisley Drive ......... But with JFET's being the way they are

What way? Hard to bias?

Look at the Paisley's JFET. The Gate is tied-up to 4.5V. The device's Gate-Source can be almost anything and it will still work. (This is *different* from voltage-gain stages.)

In fact you can drop a MOSFET or a high-gain BJT in this socket with almost no difference.

Go ahead and try.

MOSFETs are noted for higher noise (hiss) levels. I recently skimmed a paper which computed MOSFET noise from basic physics and showed no increase. I dunno if the problem is practical production methods (good *clean* oxide is tougher to make than a clean junction), or the fact that 99.9% of MOSFETs wind up switching, or what.

[bool]

When biased at Vcc/2 or higher, as followers, you can drop in mosfets in place of jfets. I think you mean a 2N7000 as this "mosfet".

I would use a darlington like a mpsa14 instead, unless you seek that tiny bit of the mosfet "tone" ...

my 2c

[gjcamann]

I would use a darlington like a mpsa14 instead, unless you seek that tiny bit of the mosfet "tone" ...

MODFETs!
Sweet!
I've never heard of these before, but they seem perfect for pedals - lower noise, high gain!

Care to expound more on the benefits of these?

Hmmmmm. I can't wait to drop them into the clipping section and see what happens! Maybe i''l try an SHO with one of these.

[bool]

Imo, there really isn't much to elaborate here

as buffers, each of these types (jfet, mosfet, bjt, darlington) would work if above mentioned biasing conditions would be met

each of these would contribute a tiny bit of its "tone" into the overall tonal equation

however, there will be a difference in "response" - or "vibe" when each of these types will be in the buffer position, a difference perhaps more "felt" than "heard"

so - there isn't much to talk about here - go and test each of these and choose the one that feels "most good" to you when playing

[Gus]

Darlington transistors have two Vbe drops.  Effects are often 9VDC powered so darlingtons are not the best device to use if you want max output swing
An example.   If you use a darlington as an emitter follower with a 1/2 power supply ref voltage for the base 4.5VDC the emitter will be at two Vbe drops below 4.5VDC

You need to be careful with mosfets they can break, they need good gate protection.

Why would you want to replace a jfet with a mosfet?

You can use high beta BJTs or different circuits for the follower with a BJT(s) for a 1meg or more input resistance.  Do a search for some of my posts about buffers and boosts I have shared a number of circuits here

Looking at the paisley drive link I don't see a reason to use J201s.  I see 511K gate resistors you can use a MPSA18 etc with 511K maybe a >511k resistors would make more sense with a 201.

[gjcamann]

So it sounds like it's really as easy as; change a resistor and drop in the MOSFET, BJT, darlington, whatever you have a surplus of. I was wondering if there were some caviates, but doesn't seem so.
Thanks all.

[R.G.]

You'll be back.

[jymaze]

Yep, you can drop anything you want, just don't mind the smoke and smell of exploding components...

[Gus]

It is not that easy.