JFET input/output buffers for a more "tube like" feel?

[MikeH]

What do you think about adding JFET input and/or output buffers to opamp circuits (ie, Rat style, CMOS style, TS style) to give them a more "tube like" feel.  Is it effective?  Is it worth it?  Could you basically attenuate the output of a JFET input buffer to unity gain so as not to overdrive the opamp circuit anymore than it would on it's own, and would doing so just add JFET charachteristics to the circuit without making too overdriven?  And the same for an output buffer, just to add some JFET feel without adding any extra clipping?

[JimRayden]

When not overdriven (as in a buffer), JFETS sound as good as any other device in linear mode (clean). JFETs do not give mojo. JFETs distort in a pleasant manner.

--------
Jimbo

[John Lyons]

Taging on a Jfet as a clipping stage or at the end would add some "tubeness" because of the way it clips but not as a buffer. usually buffers are not gain stages as much as stages that act like book ends in a way. Keep the impedance, high end or gain from falling of the shelf.

John

[brett]

Hi
typically, if you run a JFET flat out for AC gain (e.g. 10k on the drain and 2.4k on the source, bypassed with a 22uF cap), there's enough distortion added to give a nice, but mild,  flavour. 
The highly linear region for JFETs isn't as big as people sometimes imagine.  But then the amount of distortion needed to add a significant flavour (such as tubiness) is much bigger than people imagine, too (ie. several %).
I say go for it.  You'll need max JFET gain followed by a log pot or something to pull the gain back down.  Maybe do that a few times? Twelve times up, ten down (voltage divider), twelve up, ten down, twelve up, ten down, then a log pot to finish ?
cheers

[petemoore]

You'll need max JFET gain followed by a log pot or something to pull the gain back down.  Maybe do that a few times? Twelve times up, ten down (voltage divider), twelve up, ten down, twelve up, ten down, then a log pot to finish ?
  take a look at jfet amp simulators..that^ would look ~alot like ____ amp simulator if some voicing options were added.

[Steben]

Take a look at "Fetzer valves" too at www.runoffgroove.com

Triode biased they offer distortion like triodes: not much, but tasty. At max gain biasing w/o feedback they offer more distortion, especially second and even order harmonics. So I think there's a difference between "colouring" and "tube sound", since tubes can go very clean indeed on their own.

[PaulC]

I like to use a common source jfet dc coupled into an emitter follower as a type of input buffer into op amps or other things.  This way you can have a high imp input with a little nudge in gain from the jfet, and then have a low imp output from the follower.  Also I'm a fan of dropping caps when I can.  Using the Jfet you don't need an input cap.  You see this type of thing a lot in hi end audio.




Later, PaulC
Heritage amps/Tim & timmy pedals

[JimRayden]

Oh boy, a schematic for a Tim pedal. goodie goodie goodie



---------
Jimbo

[Jay Doyle]

Just because a stage doesn't have any voltage gain, doesn't mean that it can't distort.

Thought experiment:

Two JFET buffer stages, 1 Meg input impedance, 10k source resistance. One stage has a J201 with it's maximum Vgs(off) of -1.5V and the other stage is an MPF102 with it's maximum Vgs(off) of -8V. Assume correct biasing.

Now drive each with a 4Vp-p input signal with a nominal 10k output impedance, what happens at the output of each buffer stage? See where I'm going? Think about the input...

Then there is the output impedance of a buffer. Assumed to be well below that of the 'standard' 10k source resistor but that is true for only the part of the signal. As the signal swings towards ground, the effective output impedance increases the closer the signal gets to ground because the JFET is now a much higher resistance than the 10k and the 10k takes over the impedance equation. So in effect, the output impedance is constantly changing. We can use this to create distortion. Follow a buffer stage with a 10k source resistor with a 10k input impedance. Roughly, as the signal is swinging towards V+, the output impedance of the buffer circuit is low so no distortion is caused, but as the signal swings towards ground and the effective output impedance of the buffer stage approaches 10k, almost 1/2 of the signal is lost in the divider action when the signal is near ground. And because the amount of signal lost changes with signal strength (because the divider equation is always changing as the signal pushes and pulls the effective JFET resistance while the 10k remains constant) the output is non-linear and therefore distorted. Because we normally follow a buffer stage with one having a much higher input impedance than the output impedance of the buffer, even at 10k, the divider action change is small as to essentially not have any effect on the signal. Now think about lowering the threshold by lowering the following stage input impedance to 1k...

Distortion is more of an art than a science because it requires you to read between the lines of 'proper' circuit design and find out what and where one can abuse a circuit.

[PaulC]

Oh boy, a schematic for a Tim pedal.  goodie goodie goodie

Ha!  Nope - Tim's way more basic than that.  I'd like to put it in there, but there's enough of the pedals out there that I don't want to make changes.  It is what it is.  I am using this in another pedal called the "hairball" though that I've been sitting on for years. 

Jay makes some great points about buffers, and how you can distort them.  It's been said before that the sound of an 808 vs TS9 is due to how the pedal shows a lower source imp to the amp helping the amp deal with large signals.  But imho what it really is is the fact that the 808 loads down the buffer causing it to clip more than the ts9.  You can hear it, and see it on a scope.

Later, PaulC
Heritage amps/Tim & timmy pedals

[Steben]

I like to use a common source jfet dc coupled into an emitter follower as a type of input buffer into op amps or other things.  This way you can have a high imp input with a little nudge in gain from the jfet, and then have a low imp output from the follower.  Also I'm a fan of dropping caps when I can.  Using the Jfet you don't need an input cap.  You see this type of thing a lot in hi end audio.




Later, PaulC
Heritage amps/Tim & timmy pedals

If you bias it like a Triode Fetzer, you have a very nice input stage, exactly for all the reasons you mention. Make sure you have a high gain transistor in the emitter follower. What comes in for free is that a buffer like the follower, unloads the FET too, giving more gain.

@petemoore: Looks like a NICE discrete MOSfet driver, as long as you choose the right emitter resistor.