Jawari - any mods ?

[solarplexus]

Hi,

I like the jawari as an effect but sometimes I'm getting a little tired of the distortion at full gain.  I was wondering if changing the J201 with another jfet, like the mpf102 would do any difference in the sound and remove some of the distortion.  I know that the J201 is a high gain transistor, so I thought it could be the reason for the distortion at full gain.  What do you think?

Emondos

[petemoore]

Socket Q and swap, the Jfet looks to be a gain stage, MPF102 is lower gain.

[petemoore]

You could try making the source resistor [10k] variable. ala 4k7 stop resistor and a 10k pot. This would allow resistance range from about 4k7 to 14k7.
 Or clip a 50k pot to across the 10k to see what lowering the source resistor value does.
 Ive gone as far as to pull the source lead from the socket, bend it slightly so it rests on the plastic outside of the socket, wrap a series resistor to it and plug the other R lead into the socket..
 Q swaps or Sresustance changes will mod it. to what I don't know...
 Whenever I see Jfet gain stage, I see socket as allowing the ability to find what 'fits' best there as Jfets are quite inconsistant,,,I don't know whif that matters in the Jawari.

[Jaicen_solo]

It might be an idea to replace the Ge diodes with some very low voltage schottky diodes as they probably square the signal to some degree, especially when combined with the silicon diode to ground which follows.
Try adding a second LED in series with the 1N4148 diode and see if that alters the distortion.
An MPF102 will be a lower gain device, but remember that you need a fair amount of gain to drive the rectifier stage.

[solarplexus]

Would it be possible to put LED's after or before the germ. diodes to remove some of the clipping?

[Mark Hammer]

The clipping/distortion is the basis of the Jawari.  The sitar-like tone is a product of the way in which the extra harmonic content systematically "falls away".  What you hear is not the filtering of treble content that already exists, but rather the "unfuzzing" of what you feed it.  I'm not saying one can't produce the effect UNLESS you have heavy distortion, but you can't produce it in that circuit in the complete ABSENCE of some distortion.

Although I suppose it would be in direct contravention of Tim Escobedo's design principles/ethic - i.e., the most effect/impact with the fewest parts - there is no law which says you can't combine the Jawari with a clean version of the same signal so that the distortion is a much less obvious part of the sound.  Here, I'm thinking of a Sparkle Drive sort of thing where the signal is split and clean+effect signals are proportionally blended into a final mix.

[solarplexus]

Thanks for the help everyone.  

Mark : I am not saying that I don't want any distortion in the effect, sometimes I really like what I hear, but am looking for a signal with less distortion, and more of a clean processed signal.  Have you ever heard "What the hell have I?" from Alice in Chains?  THAT'S what I would love to have as an effect (and probably it is not possible, I don't know what Cantrell used for this).  Maybe I'm totally out of the game here, but still, I like the Jawari as it is now, but I would like to remove some of the gain because I find the distortion in it too fuzzy.  I read that the circuit won't work without the Ge diodes (for example, replacing it with silicon diodes), is that true?

But as far as mixing the clean and dirty signal together, how could it be possible to do so?

[Mark Hammer]

First off, there are (at least) three versions of the Jawari, two with a gain control and one without.  I made one of them for a friends kid, I think the one without a gain control.  Perhaps the gain control version is more suitable for your needs.

The earlier gain-trimpot version does not have a DC-blocking cap on the output (but it does have a level control).  The fixed gain version does.  In the fixed gain version, the .1uf cap and 100k resistor on the output form a highpass filter with a 16hz corner frequency.

If one used a circuit like Sean McLennan's B Blender, or any other circuit that splits an input signal (one version with a send/return loop) and provides a means for blending them back together, then the Jawari and clean signal couold be blended to perfection.  Since you wouldn't need the fundamental from the Jawari (since you have it from the clean copy), you could afford to shave off some of the low end of the output from the Jawari.  Changing the .1uf cap on the output to, say, 4700pf would raise the low end rolloff to around 340hz, which would make the Jawari side sort of the icing to be added to the fundamental "cake".

Others would probably have more useful advice too, but my hunch is that matching the 1N34 diodes *might* help, but not necessarily.  It assumes the two half-coils of the transformer are also matched, which they probably aren't (we're not exactly talking high-priced audiophile transformer here).  Perhaps the optimal strategy is to check the output of the transformer leads where the diodes are tied to, and select diodes that complement those outputs so as to produce some degree of matching where the three diodes meet.

The resistance to ground for the silicon diode is also something to be explored.  Perhaps a 100R resistor in series with a 1k trimpot to ground might be a good idea for fine tuning.  The objective is to be able to get the optimal harmonic sweep without having to push the front end quite so much with the gain trimpot on the FET.

At this point, I think we need some words from Tim himself.

[solarplexus]

Thanks Mark, I'll try that and see what happens.