JFET Update (Wampler Plexi Drive)

[Frank_NH]

Just wanted to update everyone on my experiments with JFET biasing.  From my last post (JFETS!)  I reported that I had a spreadsheet for computing the drain/source resistors for a common source JFET amplifier circuit given Idss and Vp for a JFET (and assuming 9V supply voltage).  I also mentioned that the Wampler Plexi Drive (original circuit) should bias for JFETs with Idss of about 0.65 mA and Vp of -0.93 V.  

Well, I decided to go ahead and build the Wampler Plexi-Drive to check this out, and I'm happy to report that it worked!  The three JFETs bias around 4.5 V at the drains!   I had tested all of my J201s for Idss and Vp and found three with that range.  My only problem with the circuit is some humming I'm getting after Q1 which is unrelated to the biasing (I'll describe this in a separate post), but the overdrive sounds great!

So, my conclusion is that it is possible to determine (at least approximately) the JFET parameters required for certain JFET-based overdrive/distortion circuits to bias correctly.  Even if you use a drain trimmer, you can estimate the size of the trimmer so that you don't choose something that is too large or small, and thus difficult to dial in.  You DO need to test all your JFETs, however, using the JFET tester posted here.  I just got a new batch of J202s that I'll be looking at over the weekend - more on that later, bot those should be near-drop-in replacements for 2N5457s...

[rocket88]

That is awesome. But now you're making me have to go through and test all my JFETS. It's going to take forever.

[Frank_NH]

Zach it'll be worth the trouble.  If you only have one or two J201s in a circuit (like the Peppermill for example), then you can just audition specific ones until you find one that works.  But doing that for circuits with 3 or more can be a pain and then you don't know why this one worked and this other one didn't.  Best to sort them, even if you have 100.  It doesn't take that long.

One interesting point about this is that tubes like the 12AX7 triodes are made fairly precisely so that replacing one in a tube amp preamp circuit with fixed cathode and anode resistors usually works OK.  JFETs however can have a wide range of Idss and Vp - so you need to sort them so that at least you can get the same repeatability in applications as you do with tubes.  Even if you use a drain trimmer, biasing two different JFETs exactly at 4.5V drain voltage will result in two different operating points, and hence a difference in gain and response.  That's why Brian Wampler and others have suggested using fixed resistors in circuits like the Plexi-Drive and handpicking the JFETs, versus using trimmers.

[IvIark]

Great work Frank, I'd be interested in seeing the calculation you used or the spreadsheet.
And I agree it is definitely worth the trouble.  Being able to build a JFET circuit consistently without having to rely on noisy trimmers would be a real plus

[rocket88]

Yea Frank, I do have a lot of plans for a bunch of JFET based circuits, so I know it's worth it. I mean that's what I sorted all of those Ge's I bought to make building easier. But man, I feel like that's all I'm doing now, sorting, sorting, sorting. Lol. I've got to build the tester, which I'll try to do this weekend. Btw, in all you've been researching about JFETs, does this apply primarily to amp-in-a-box pedals/tube like JFET based OD's, or is it all circuits that use JFETs?

I need a good build, having been able to since the germanium giant we all came up with, and I just fixed that wah with a new inductor, after sitting not working since it was first built like 3 months ago. Any chance of getting a copy of the spreadsheet with the formulas you used?

Hey Mark. Do you know if the DCA75 does the same calculations for JFETs? I know you've been looking for one for awhile now, and since you mentioned that it can read JFETs, my interest has been peaked, and now with the info from Frank, I've been thinking that it might be worth upgrading to it.

[Surgeon]

This is really cool work you've done Frank! Kudos!

Now, I've been reading the sources you've linked to (given I'll need more time to really do a thorough read) and the basic math formulas for Rs and RD are quite simple. Rd=0.9* [Vcc - (2*Vp)]/Idss and Rs = 0.83*Vp/Idss

Given that we have a schematic and the resistors already defined, I'm wondering how you solved the problem of having two unknown variables (Idss and Vp) there.

Again, I'll need to read a bit more but that's the first thing that comes to mind with what I've read so far. I'm guessing there's more for me to read that would solve this but if you could just point the way I'd be grateful...

Cheers!

[Silver Blues]

I'd love to have this spreadsheet too. Glad to hear it works.

[IvIark]

rocket88 wrote

Hey Mark. Do you know if the DCA75 does the same calculations for JFETs? I know you've been looking for one for awhile now, and since you mentioned that it can read JFETs, my interest has been peaked, and now with the info from Frank, I've been thinking that it might be worth upgrading to it.

It will give you the VGS off but not Idss as far as I know.

[Frank_NH]

Surgeon:

My spreadsheet doesn't use the Fetzer formulas but rather the general relations for the generic common source jfet amp.  The main equation relates the drain current to Idss and Vgs,off (Vgs,off is essentially the same as the pinch off voltage, where Id = 0) as follows:

Id = Idss*(1-Vgs/Vgs,(off))^2

where Id is the drain current and Vgs is gate-source voltage (which is negative, since Vg < Vs for the JFET to be "on").  Note that for Vgs = 0, Id = Idss, which is the max current that can flow through the JFET (the so-called saturation current), and if Vgs = Vgs,off, the drain current goes to zero.

This equation along with some simple relations for the voltage and currents through the drain and source resistors can yield the solution for the bias.  Note that this is a DC bias (i.e. gate voltage is zero), but that is what you set by selecting the source and drain resistors.

I should note that you have a lot of choices for the bias level.  For example you may choose to bias such that Id = Idss/2 and Vd = Vdd/2, where Vdd = supply voltage (9V), and Vd is the drain voltage we're all familiar with.  Or you can choose the source resistor according to the Fetzer formula, choose 4.5V for Vd and let the other parameters fall out of the analysis.

I'll plan to make my spreadsheet available after making it a bit more tidy.  Again, it's nothing earth shattering, and only applies to stages which have the common source configuration.  But for me, it's been helpful in understanding how knowing Idss and Vp  can be helpful in selecting JFETs for specific circuits.

[Frank_NH]

Zach - just make the JFET tester.  It's pretty simple to build and use with an existing DVM.  Let me know if folks would like to see a tutorial.

Once you have the tester, you can go through and measure a pack of JFETs pretty quickly.  No need to invest in any new gear - unless you like gear

[rocket88]

I figured I would build the tester anyway, just you never know what some tools can do to make your life just a little easier. Plus I was figuring that if the dca75 would do it for you, I would sell my dca55 to someone here that could use it. Especially since it doesn't seem as available in the US for some reason. And to be honest I've had the gas to get a new toy and need an excuse. Lol. Maybe I'll act sully spend the money on a soldering station instead of my little weller, but it's been good to me and does everything I ask and more.

It's a bummer that the DCA75 doesn't do more with FETs.

[motterpaul]

I just wanted you to know that I also built the Wampler Plexi drive (with trimmers) and I think it is one of the best overdrives here. I love it. I didn't test my J201s first (because I built this a couple months ago) but I had no problem biasing them by ear to make it sound great. It also makes almost no noise at all, surpriingly quiet for how loud it is.

Meanwhile I tried to build the THOR with some 2N5457s I just got and I get bad oscillation and a very trebly throughput.

I seem to be having a problem with my JFET tester (non-IC version) in that I tend to get very low readings. 001 if I am lucky on the lowest setting.

I could swear I built it correctly but you never know. I noticed many other people saying they also had problems with the non-IC.

I just got two different batches of 2n5457s from different sources, and they look very different from one another.

(1) has very large lettering on the flat side and says
M2N
5457
then on the back side it has three lines fully across the curve; white white and blue.

(2) the other looks like this on flat side
F D512
2N
5457
and the curve side is blank.

I have not had much luck with either batch o far, does anyone know a simple circuit to "test" 2n5457s?

[alltrax74]

I built about 10-15 Plexi Drives, all without trimmers.
What I do is use sockets for the drain resistor, put the right value in then try the JFETs that will bias the best (4.5 - 5V).
If I don't exactly nail it, I swap the drain resistor for the closest value (which generally is within 10%), and that does the trick pretty well ;-)

[Frank_NH]

Alltrax74 - what were your final drain resistor values?  Curious to see what worked for your JFETs.  To be honest, I'm not a big fan of trimmers any more.  They are very noisy when adjusting and may contribute some noise to the build overall.  Unless you plan to regularly swap out your JFETs like tubes  I think it's better to have a fixed resistor.

Brian Wampler mentioned in some posts FSB that the fixed source/drain resistor values in the Plexi-Drive were key in getting the "right" sound from the effect.  Hence, he strongly advocated testing JFETs and using ones that biased correctly versus using trimmers.  There is a similar situation with the Fairfield Barbershop (which also uses fixed drain resistors).

[alltrax74]

Something between 13 and 17k - so yes that's more like 15/20% - and even if BW is right (I also read its post on FSB), close values like the one I use don't really make a difference

[Frank_NH]

Well, those values aren't too way off.  And the quality of the sound will always be subjective, so if it sounds good as it is, then your values are right!

[alltrax74]

And it does sound damn good my friend (and saves so much time!)