The JFET Bender Preamp

[tarahall]

Attached below is a layout I did and an updated schematic. I took you suggestion and added resistors in series with the pots to limit their span and make them easier to adjust.

Would there be any difference/advantage in inserting the FETs between the resistor(s) R4 and/or R9 and the caps(s) C1 and C3 instead of between the caps and ground ?

[Ironsinthefire]

It shouldn't be that hard to find P-channel JFETs.

(Refer to the schematic JFET_bender_7.gif) Without P-channels, use an N-ch on the first stage. Reverse diodes D1 to D5 except connect D5 to the +15 rail. Keep R3 and R4 the same and use negative bias on the Gate of N-ch Q1. Insert an inverting stage with a gain of one between the first stage and the tone stack. Use the stage after the tone stack as is.

Thanks!
Working on  another project, I was recently burned on a shipment of 2n5484 fets from eBay/China (they shipped mystery NPN bipolars with the Fairchild logo and 2n5484 silkscreened on them!) so I'm hesitant to work with parts I can't get from Digikey/Mouser, etc.  Anyway, I decided to dig deeper and found some 2SJ and 2SK fets in my parts bin which seem to work. I have a few more design theory questions about this circuit (basically wondering about the stage gains, the use of a single "grid clamping simulator diode" on the 2nd stage etc.) but will wait until I've organized my thoughts a bit more before asking to avoid wasting your time.

[Loudthud]

Would there be any difference/advantage in inserting the FETs between the resistor(s) R4 and/or R9 and the caps(s) C1 and C3 instead of between the caps and ground ?

When you overdrive one of these stages, there could be some DC accumulation on C1 or C3 because of the asymmetrical clipping. This DC would interact with the DC bias on the Gate of the JFET. It's' hard to say if it would be audible or if it would sound good or bad. Try it.

FYI: The polarity of C3 might be an issue. The offset of the opamp could go either way. Short out C3. If pin 1 of the opamp goes positive, C3 is OK. If the polarity is negative, reverse C3. If you can find a 4.7uF film cap or a Non-Polarized electrolytic, you can use those without a worry.

[tarahall]

The JFET is used as a variable resistor in a region where it's resistance changes at different current levels. MPF102 JFETs were used because they had a bendable resistance in the right range.

I post this as it may be of some interest to those following this thread.

On researching that presumably "CRATE" overdrive circuit idea in US patent 5647004 filed by St. Louis Music, I came across Scholz's US patent 4627094 which uses an N-Ch JFET (possibly a 2n4340) to vary the gain in the compressor IC (probably a TL072) using the gate's bias voltage being varied by a DBX 2252 rms level detector chip - see fig 2a in the patent. Not exactly the same structure as yours Loudthud I know, but it would obviously lend weight to that JFET variable resistor concept. 

I guess this patent this was the original circuit idea for one of the Rockman models. I know they used TL072's in the little headphone rockmans as I have two of them (a model 1 & a model 2), and the JFET used in those was a 2N4340 although those models don't use the DBX chip.

I have circuit tracings of the 2 models that I did using Express PCB's schematic program if anyone is interested in copies but I have not verified their complete accuracy - they were only used as guide for my curiousity

[Ironsinthefire]

Etched the PCB tonight, so I will probably be up and running soon. (Had a test circuit on a protoboard.)

A thought or two- based on your scope traces, it looks like the gain of the tube circuit you emulated with the bender is about 60X. That must be unloaded? In the tube Fender, the first stage is loaded by the tone stack and the gain is more like 43X. (going by the AC voltages off the Fender DRRI schematic, 37mV in =1.6V out) So, maybe better to adjust things to more closely emulate the original?

I'm guessing that your design process was to note that the tube circuit swings about 200V pk-pk at the limits, the op-amp/jfet is limited to 30V or less pk-pk by the 15 volt rails, so the gain of the op amp stage has to be scaled accordingly to get the right headroom. 200/30= 6.6,  so using your gain of 60X, the op amp's gain should be 60/6.6= 9X . I may try revising the gain of my build down, as the op amp won't get loaded down at all from the tone stack due the OpAmp's low output impedance.

[Loudthud]

The load in the 12AX7 X-Y photo is just the 10X scope probe DC coupled. Any load like a tone stack would cause a loop in the X-Y trace at just about any frequency. And speaking of frequency, the gain (Vout/Vin) will change with frequency and signal level so you really can't put a single number on the gain. My objective for this project was to show how, by simple means, the essence of a tube preamp can be duplicated. It certainly doesn't duplicate everything a tube preamp does to the nth degree like some preamps that use 50 or 100 opamps does. I hope people that build it will notice and train their ears to hear the difference between the typical soild state preamp and the preamp with tube-like low order distortion.

[Ironsinthefire]

Design goals understood, thanks for taking the time to explain. Looks a good "best for least" solution.

I've got the board partially stuffed and the first stage working. I'm using a 2SJ something  for the FET and it seems to work, I've got some asymmetry on the scope,  but I think the FET specs might be a bit off. Found that what appears to be the SMD equivalent of the 2N5462 from Digikey- the MMBF5462 at 48cents each and a nice source of mini SMD to through-hole kludge boards for 25 cents each.

http://mklec.com/components/circuit_boards_pcbs/sot-23-smd-to-through-hole-adapter-pcb

I should probably stop trying to gild the lilly, just build this thing stock and see how it sounds.

It would be cool to take a multi-pole switch and wire it to be able to switch the fets and grid clamping simulation diodes out so you could A/B the difference. (You could have the ground leg of the op amp feedback network switch to a trimpot to ground so you could match the FET in circuit /bypassed levels exactly )

[Ironsinthefire]

Up and running. This is a scope shot showing input and output of the first stage. Jfet is rounding off the bottom of the trace, when you push it harder the diode "grid clamping" starts acting on the top of the trace.

This is using a J103 fet adjusted to the 80% gain mark. Looks OK?

[Loudthud]

That looks good. Where did you find a J103? I don't have any data on that part.

[Ironsinthefire]

I got the 2sj103 from my bag of fets, I think it's obsolete now in through-hole format.  I think it was a leftover from back when I used to repair Yamaha audio products. It's not too twitchy to adjust, I used single turn pots for my build.

Datasheet:
https://www.promelec.ru/pdf/2SJ103.pdf

The pinout requires criss-crossing leads for your PCB layout.

[Loudthud]

OK. I didn't think of the implied 2S prefix because there are parts in the USA market that are "Jxxx".

Many USA JFETs have the Gate on what would be pin 3 of the Toshiba part because they use a transistor lead frame where the Collector goes to pin 3. Collector on a transistor die is the back side contact. Don't worry about getting Drain and Source right. They can be interchanged.