Driving JFET - what's going on?

[lion]

I'm building sort of a preamp - this is a part of it:



Variable drive into the J201 (to clip the JFET - plan was also to try a diode ladder in front of the JFET). Very basic, but I'm experiencing something odd: With increasing drive levels the resulting response out of the circuit is "tilting" - see plots!!!



The lower plot = minimum drive, and shows the response I'd expect,
and upper = pot at maximum - clearly tilting.

When bypassing the JFET (drive pot wiper >> 68k >> 150k) the response stays flat at all levels.

I've measured the JFET's Vp/Idss and calculated Rd/Rs with ROG's calculator.
Supply is 9V (actually 8.85V) and JFET voltages in circuit are:
Drain = 6.18V
Source = 0.287V.

The signal level into the JFET is aprox 150mV/1.3V at min/max drive levels. Tried other J201's, but no changes in the response behavior.

At max drive levels the JFET is clipping asym, but I wouldn't think that alone would cause the odd response. What am I missing? 

Erik

[blackieNYC]

The drive pot and the 10uf cap make a big pass filter, but they would have to be off by a factor of 100 or so to get up into the audio range. I doubt you've made such a mistake. Is the cap ok?

[merlinb]

Source = 0.287V.
The signal level into the JFET is aprox 1.3V at max drive levels.

Looks like you're pobably overdriving the FET, and the clipping may be playing havoc with your measurement system.

[anotherjim]

If you're driving it into clipping, you will get increased high frequency content from distortion products. If that's what your seeing, aren't you getting what you want?

[PRR]

"Frequency response" is fairly meaningless when OVER-driven.

1+V on Gate with only 0.3V on Source sure is overdriven.

[lion]

Everything's probably alright so far then 

Also I think part of the results I'm gettings is due to lack of experience/operator error - and the plotting tools I have. I'll give it another go and try to be more accurate with levels etc.

With a sine wave the JFET starts clipping around 900mV input, according to my ancient valve croc. The J201 has a Vp of 0.846V - seem to remember reading that Vp also indicates the aprox possible signal before the FET starts clipping. Obviously I could swap in another FET with a higher Vp if needed.

Going to experiment with clipping diodes to ground in front of the JFET - 2-3-4 pairs of 1N34A's and a 1n914 pair string seemed to work in a quick prelim test - and try to scale the clipping points of the diode string vs JFET to get the "tape/head saturation" type of distortion I'm looking for.

Tried "pure" symmetrical clipping for mostly odd order harmonics, but it doesn't give the results I'm after. Now aiming to try a combination of sym clipping (diode ladder) with some asym clipping (JFET) for more even order harmonics mixed in.

Erik

[ashcat_lt]

The source is white noise?  You've got SPAN's slope set to 3db/oct.  I think that's the default, and it's about right for analyzing music, but for what you're trying to do, I think 0db is more appropriate.  If the source is actually white noise, I think the upper trace looks more correct.  Seems to indicate more that you're actually losing treble at lower levels.

[lion]

I used a 20-20.000Hz sweep, but I had some difficulties in the process - hence the 3dB/oct slope you noticed

To start from scratch I have sourced new test tones - sweeps and noise. Not sure if that is the whole answer, but the plots I get now seems much more logical, and - I can't replicate the odd results I got the other day from varying drive settings. New plots are close to flat - and pretty consistent at all drive level.

I must also correct myself re the JFET clipping point - actually it starts clipping much sooner, at around 500mV.

Erik