Fun with JFETs and servos

jubal81

Well-known member
Chuck's thread on the Brownout OD finally motivated me to do some breadboarding with JEFTs and opamp servos for biasing.

So far, things are looking pretty neat. I put together a couple stages of the Wampler Black 65 for testing and sure enough, I'm able to swap in and out JFETs willy-nilly and have it bias up perfectly. I don't have any J201s, so I've been using 2SK170, 2Sk117 and the good old 2N5457. I've got it set up so I'm getting 5.8V on the drains with a 9V supply. When they begin to break up, the distortion looks identical on the scope, the only difference being that the higher |vp| transistors have less gain/more headroom.

As far as the Black 65 goes, I've read that they're set up with 4V on the drains and I've not been able to dial it that low with this setup and my JFETs. (Using a rail-rail JFET opamp). I've always done 5.8 or so on a 9V supply with my projects where top and bottom of the signal start distorting at the same time.

Anyway, here's the circuit snippet I've got worked up and some pics of scope readings with different JFETs.

Screen Shot 2021-07-23 at 9.32.30 AM.png
2SK170:

Screen-Shot-2021-07-23-at-9-50-21-AM.png


2N5457:

Screen-Shot-2021-07-23-at-9-52-04-AM.png
 
Spent a little more time playing around with this today. Unfortunately, an LM358 doesn't perform well. Think I'll look around for a cheaper JFET-input rail-to-rail opamp.

Also confirmed it's easy to raise or lower the value of the drain resistor to dial in the amount of gain while maintaining proper bias.

Another thing I'm noticing again is the clipping character of a JFET amplifier stage is different when the source is bypassed with a capacitor. With the cap, the clipping transition is rounder. Without the cap, you get a really sharp style of clipping. I think one interesting thing to explore with the Black 65 would be to use a 'lower gain' JFET on the clipping stage, but with a bypass cap to see if you get a rounder clipping with the same amount of gain. Or, with the servo bias, you might be able to switch from one style to another by switching the passives on the source of the JFET.

Schematic update: Using 18K and 24K resistors for the Vref into the servos. This is giving about 5.3V on the drains on a 9V supply.
 
Thanks for sharing this.

Would a LM258 or LM324 work? I think I've got some. I may give this a try.
 
Thanks for sharing this.

Would a LM258 or LM324 work? I think I've got some. I may give this a try.
Nope. those are pretty much the 358.

Only other rail-to-rail I've got on hand is a 2272 and it works fine (mosfet-based). It's through-hole, but they're $5 each these days. OUCH!

If something comes of this, I'd like to make it through-hole capable so you guys can build one if you want, and I might have found the ticket. The TL052 looks like a rail-to-rail version of the good ol' TL072 and it's only 84 cents for the dual opamp at Mouser. Can't believe I never knew about that one before now.
 
Very cool stuff indeed. I have been pondering building another Black 65 since the one I made on an etched MBP layout went astray staying on a friends board. Would be cool to get an updated version of this and very neat to dive deeper in these ideas. Cheers!
 
  • Like
Reactions: fig
Very cool stuff indeed. I have been pondering building another Black 65 since the one I made on an etched MBP layout went astray staying on a friends board. Would be cool to get an updated version of this and very neat to dive deeper in these ideas. Cheers!
It's a good one. I was heavily into JFET ODs for years and the Black 65 is easily in the top 3 for me in that category.
If I build this out, I'm most likely going to leave off the bonus boost stage. There are a lot better boosts to hit it with, especially a good ol' Klone. Might look into emitter bypass caps for a low/high gain switch, though.
 
Your post a couple above is the first I've ever heard of a TL052 !

Of course, I'm still but a padawan in the grand scale of things, and there's much I've yet to discover that's probably old hat to the sage JFETi knights that inhabit this forum.

Fascinating thread.
 
  • Like
Reactions: fig
Oh I see now. I missed the giant IC1_A/B 😖. Duh fig. I've got some TL062s, so I'll give it go with one of those...unless there's another glaring detail I'm missing? Apologies. I am applying the learn as I go technique, so patience is greatly appreciated.
 
Oh I see now. I missed the giant IC1_A/B 😖. Duh fig. I've got some TL062s, so I'll give it go with one of those...unless there's another glaring detail I'm missing? Apologies. I am applying the learn as I go technique, so patience is greatly appreciated.
No sweat! I love this stuff.
The TL062 won't give the best results. I tried an 072 today and it couldn't keep up. Looks like the trick is a Fet-input, rail-to-rail opamp for this.
 
Awesome and I think leaving the boost off in the Black 65 is a good idea from my point as well. Maybe do the emitter bypass cap on a DPDT on/off/on for three different positions switching two jfets?
 
I was looking at the TL052 sheet and it says "minimum supply voltage: 10v". I guess a dual rail 9v would work or a charge pump to get 18v (15v regulated) also would work.
 
I was looking at the TL052 sheet and it says "minimum supply voltage: 10v". I guess a dual rail 9v would work or a charge pump to get 18v (15v regulated) also would work.
Good catch. Totally whiffed on that. I've already got a few coming, so we'll see if they start up with 9V. I think there's a pretty good chance.
 
I was looking at the TL052 sheet and it says "minimum supply voltage: 10v". I guess a dual rail 9v would work or a charge pump to get 18v (15v regulated) also would work.
With the max @ 30vdc, I'll bet 18vdc is the sweet spot, even though I'd be swapping a few caps. I can dial in 10v to test it anyway.
 
24-TL052s arrived today. I'll at least get started this evening, maybe finish it, if I can pay attention long enough. If anyone wants a couple of these, let me know.
 
24-TL052s arrived today. I'll at least get started this evening, maybe finish it, if I can pay attention long enough. If anyone wants a couple of these, let me know.
Wow! I don't think I even have a shipping notice yet on mine. Ordered from Digikey.

I'm also planning on trying a non-inverting servo bias on the JFET gates and possibly just a simple voltage divider on the gates. You can bias up a Black 65 with trimmers that way without changing the Drain or Source resistor values.

EDIT: Looks like my orders are on track to be here Friday.
 
  • Like
Reactions: fig
Yeah, Mouser knocked this one out. I don't normally use them but they had these and the MPSA14s that I needed for someone. Well played Mouser.
 
I'll try the J201s and some 2N5485s. If I can get this probe working, I'll also try to get some waveforms posted. I think I've spent more time looking at replacement probes than scopes. :oops:
 
I'll try the J201s and some 2N5485s. If I can get this probe working, I'll also try to get some waveforms posted. I think I've spent more time looking at replacement probes than scopes. :oops:
If you're going by the schematic above, use 18K for R8 and 24K for R9. That should give you about 5.3V on the drains with a 9V supply. Also, I used 2u2 for C3 ad C7 because I had those laying around. Those caps could probably drop to 100n without running into trouble.
 
Back
Top