FET replacement

[JebemMajke]

I want to make Little green wonder. But I don't have any 2sk170's, and I can't find it. Here is the schematic


Can I use, like, 2sk117?

[Earthscum]

Looking at the datasheets, I'd say, solidly, yes. The VGS is similar. With the 10k Source resistor, you could also use a J201.

A JFET like the MPF102 may have a high enough VGS that it could bias the source at too high of a voltage, inducing unwanted clipping.

[armdnrdy]

The only thing that's kind of "special" about this JFET is that it's low noise.

Low noise in an overdrive unit? 

There are a ton of other JFETs that can be used.

[Earthscum]

the 2SK117 is rated for low noise audio use as well. The only significant difference between the two I can spot right away is the IDSS. With the 10k source resistor, it shouldn't have any significant audible change, or performance change. The J201 is a "low noise" transistor, as well.

[Eric.nail]

Low noise on purpose. The LGW is meant to be a high head room, low drive pedal. Let us know the results! if it sounds good with 201's that's a quick solution to Mammoth not carrying what the schem calls for.

[psychedelicfish]

Looks like it's just a buffer to me... why not just use a BJT?

[ggedamed]

One thing to note is the (significantly) better current capability of 2SK170 versus J201, which makes 2SK170 a much better buffer. In this particular circuit it will be next to no difference, though.

From the biasing point of view, there are multiple flavours of 2SK170, the lowest Idss being 2SK170-GR (2.6 to 6.5mA). This makes 2SK117-GR a direct replacement for 2SK170-GR (and 2SK177-BL for 2SK170BL), having the same parameters and the same pinout. But any version will do, really.

Being a half-supply biased buffer, another thing to note is that 2SK170 (and 2SK117) has low Vp (mine are around 0.5V) which makes the headroom of this configuration better than the headroom of a usual (non-rail-to-rail) op amp buffer. Again, this doesn't matter much in a guitar effect pedal.

And yeah, it's just a buffer, you can replace it with whatever BJT/op amp buffer it's at hand.

The only thing that's kind of "special" about this JFET is that it's low noise.

Low noise in an overdrive unit? 

There are a ton of other JFETs that can be used.

So your statement is that overdrives don't need low noise components? I was under the impression that the higher the gain, the lower noise components are needed. And that lower noise is a good thing, regardless of the application.

[armdnrdy]

Looks like it's just a buffer to me... why not just use a BJT?

JFETs are better for input buffers because they are higher impedance.

I still don't understand the purpose of using a low noise JFET in an overdrive circuit. You are creating distortion correct? It is a high gain circuit correct? I don't believe that the "noise" would be created at the very beginning of the circuit. It seems to me that component "noise" would be lost in the mix of the distortion. Unless we're building audiophile overdrives.   

I see using low noise components in circuits such as delays, compressors, and clean boosts but when I skimmed over this schematic my first thought was Tube Screamer.

There are Flangers that actually add a bit of distortion via a couple of diodes to add harmonics to the effect, but let me get this straight..... we're going to quiet down that distortion unit by using low noise components.

I think that JFET was what was on the shelf come build time!

[Johan]

Any noise added before the distortion will be amplified with the signal, including that generated by the input buffer. If the signal is amplifier, so is the noise
J

[armdnrdy]

Any noise added before the distortion will be amplified with the signal, including that generated by the input buffer. If the signal is amplifier, so is the noise
J

I am aware of that but...I don't believe that a low noise JFET at the input would make much of a difference in the sound quality at the output after the signal goes rolling through the drive section. We are adding distortion!

I used to use metal film resistors in overdrive circuits thinking that I would make the circuit as quiet as possible! Show me the metal film in a production tubescreamer. It's kind of redundant to try to quiet something down that's adding a boatload of gain! Not to mention the 4558! Not exactly the quietest op amp in existence. But then there are the guys that put high performance, low noise op amps in an overdrive circuit thinking that a high quality $8.00 op amp has got to sound better. 

I guess it's all up to individual taste. 

[ggedamed]

I totally agree that the sound quality (if I correctly understand what you're referring) does not change with a lower noise component. But this was never a point.
If the gain is set at maximum, the noise comes through when you're not playing, even in a Tube Screamer. For an extreme distortion I would still use low noise components, even if just for the peace of mind.
Sure, lots of times the guitar itself is a much bigger noise source and you can turn off your pedals or the guitar volume when you're not playing.

So, it's like you said, all up to individual taste.

BTW, it doesn't make sense to search metal film resistors in a Tube Screamer. It's not like Maxon would choose the components with the best parameters, regardless of price. And metal film resistors have no mojo, anyway.

[duck_arse]

another thing to note is that 2SK170 (and 2SK117) has low Vp (mine are around 0.5V) which makes the headroom of this configuration better than the headroom of a usual (non-rail-to-rail) op amp buffer

from the fetzer2 disection at ROG:

The proposed JFET stage will start clipping when the input voltage exceeds +/- |Vp|. It is interesting to notice that the maximum input voltage without clipping is independent of Vcc and Idss. Then clipping-free input dynamic range is defined by: Vi = +/- |Vp|

it seems to me that using a jfet with low Vp will reduce the headroom rather than increase it. with a low Vp, you start distorting the signal at the buffer itself.

[Earthscum]

Your VP is the gate being half a volt (or whatever the VP is) than the Source. So, in a half-voltage bias, like this, if the VP is .5V, then yout Source voltage is actually going to be 5V instead of the 4.5V. That's what I mentioned earlier about using something like the MPF102. Say the MPF JFET has a VP of 3.5V, then your source would try to settle at 4.5V+3.5V=8V, leaving you with only 1V between Source and Drain, which would cause an asymetrical signal through what was supposed to be a buffer.

One thing I don't get is the resistor before the cap (not the pulldown resistor). The input impedance is so high that the resistor in front of the cap is moot, unless it's possibly for Gate protection? 

[ggedamed]

another thing to note is that 2SK170 (and 2SK117) has low Vp (mine are around 0.5V) which makes the headroom of this configuration better than the headroom of a usual (non-rail-to-rail) op amp buffer

from the fetzer2 disection at ROG:

The proposed JFET stage will start clipping when the input voltage exceeds +/- |Vp|. It is interesting to notice that the maximum input voltage without clipping is independent of Vcc and Idss. Then clipping-free input dynamic range is defined by: Vi = +/- |Vp|

it seems to me that using a jfet with low Vp will reduce the headroom rather than increase it. with a low Vp, you start distorting the signal at the buffer itself.

You chopped away the significant part of what I said.

[...] Being a half-supply biased buffer, another thing to note is that 2SK170 (and 2SK117) has low Vp (mine are around 0.5V) which makes the headroom of this configuration better than the headroom of a usual (non-rail-to-rail) op amp buffer. Again, this doesn't matter much in a guitar effect pedal. [...]

Here's my understanding of the matter:
The Fetzer Valve is self-biased common-source amplifier (the gate is connected to the ground), while the buffer we're talking about is voltage divider biased source follower (the gate is connected to the half-supply point).
In a -0.5V Vp JFET Fetzer Valve the source settles at 0.5V (-Vp higher than the gate) and the signal will be clipped as soon as the input signal goes lower than -0.5V (which makes the source reach the ground voltage), because obviously the source cannot go lower. And the circuit is an amplifier, not a buffer.
With the same JFET in the buffer above, the source settles at 4.5V -(-Vp) = 5V  (again, -Vp higher than the gate) and the signal can go 4Vpp with no clipping. So, in this circuit, the lower Vp is, the better. LATER EDIT: if you want to squeeze every milivolt of headroom you can bias the JFET at 4.25V instead of 4.5V (half-supply) and your signal can go 4.25Vpp. A 8.5V input voltage swing for a buffer with a 9V power supply is not bad at all.
And in a TL072 buffer circuit the input signal can go ~2.5Vpp with no clipping, because of TL072 internal architecture.

[duck_arse]

I love jfets because of their simplicity, but at the same time, just about everything about them baffles me.

so why am I using self-bias and worring about distortion and Vp in my input stages when **all** I need is the second resistor for the voltage divider bias instead?