Simple JFET-Blend Circuit distorts

[lars-musik]

Hello everybody,

I recently built a EQD Sea Machine for a Bass player and added a simple Jfet-Blender to the circuit. I only tested the pedal with my guitar and everything worked fine. Now the bass player wrote me, that the circuit distorts at high "wet" settings and even more so if he puts his Bass Synth Wah in the chain before it.

Are there blend circuits better suited for high-output or active basses? Do you have any other ideas what could help?

This is the blend circuit in question:



and here for completeness the Sea Machine:







Thanks!

[Gus]

Where did you add the blend in and return?

Are you sure the distortion is only from the blend circuit? Did you try the circuit without the blend with a bass?

R14 is part of the clean blend in the pedal. The depth control mixes into the clean at the - input summing node of IC4A

You might want to adjust the IC4A section for the clean to wet mix and not use the jfet circuit

If you do want a blend use an opamp buffer that has a good output swing.

[anotherjim]

As with any Jfet circuit, headroom can be an issue. I'm starting to stop believing in fixed value bias for them. They're getting too rare/expensive to select from a bag everytime. That circuit may have been drawn up with J201 in mind, which tend to be a little more consistent than others in 9v circuits.

Anyway, is the source voltage anywhere near half supply (4.5v)? You may have to change R3 value. Getting the source somewhere between 3 to 6v ought to be good enough unless you expect very hot signal levels - but if it still distorts at near 4.5v, I don't know anything active that won't.

Puzzled as to how you blend as the delay has fixed amount of dry? If you blend the input with the output of delay, you will be boosting the dry x2 as there is always x1 dry out of the delay.
Dry could be cut for wet only send/return use by removing R14. Made variable with a blend pot at R14/R15 instead of depth pot.

Sea Machine itself won't like hot input to the PT2399. It has no soft limiting to tolerate overload. Something like a 20k-ish trimmer in series with R3 should let you back the input level off into the delay chip. Or you might try the green LED hack on the PT2399 - I think ( could be remembering wrong) - pin7?

[lars-musik]

Thanks for your input. As the pedal works just fine on passive bass and guitar I wonder if some kind of attenuation at the input would do the trick.

Gesendet von meinem GT-I9505G mit Tapatalk

[blackieNYC]

I was very determined to make a jfet blender, it distorted and I gave in and went with op amps.
I think the J201 will clip long before the MPF102, but gain is sacrificed?  That's what I got out of the Fetzer valve article at ROG.  I'd love for some one to 'splain.

[Gus]

For the jfet circuit

You need to think about the Vgs at the drain current you are operating at and what you are driving and the Rdson of the jfet.

IDSS needs to be > 9VDC / 4.7K about 2ma min(max output to 9VDC rail)

You should adjust the jfet circuit bias with the R1, R2 resistor ratio.

Using a scope and signal generator drive the jfet circuit with an input signal driving the jfet almost to clipping and adjust the R1, R2 ratio for equal clipping ALSO Have an output load equal to the next stage or the next stage attached when adjusting the ratio

I would guess you will need to decrease R1s value or increase R2s value

I have posted about adjusting BJT emitter followers offset from 1/2 the power supply voltage biasing in the past

Also you should use noiseless biasing like R.G. and others post about I would use a 1meg from gate to center of the voltage divider you can then use low value resistor for the center node maybe 10K +9VDC to center node. and 15k center node to ground with a cap at the center node to ground etc to start.

[PRR]

The Sea Machine's Depth pot and IC4A is a "blend".

Not sure where you have gone.

Bass can be much bigger than guitar. I suspect the PT chip is overloading in the wet path.

[ElectricDruid]

Seconded. Bass whacks out a much bigger signal than geetar. Headroom and assumptions have to be adjusted accordingly.
As a quick fix, padding the input down as suggested above sounds good to me.

T.

[lars-musik]

The Sea Machine's Depth pot and IC4A is a "blend".

Hi Paul,

Yes, we discussed that with the schematic at hand. But when I tried the pedal I realized (imagined?) that that depth depth pot determines the amount of dry signal mixed to the delayed one, and thus creating the chorus sound. So it is an integral part of sound forming and not really setting the dry-wet-ratio of the final sound.

As a quick fix, padding the input down as suggested above sounds good to me.

Thanks, Tom!

That would be a serial pot at the input? And if I take this road, would I need a gain stage at the end to get to unity again?


I thought about using the Moosapotamus Paralooper instead of the jfet-blender. Firstly, it is opamp driven, so I might avoid a lot of the troubles Gus, Jim and Blackie wrote about (thank you for your help!) and it already has some gain stage at the end.

Could that be a practicable solution?

[anotherjim]

If the input just to the delay is padded down then the wet is going to sound quieter compared to the dry. A proper wet/dry mix either by a blend control or 2 separate volume controls into that mix can obviously restore the balance, but it might be quieter than bypassed.

In this delay, if you drop the input -6dB by making R3=20k, then you can recover that by +6dB by making R13=4k7. R13 forms a high pass filter with C11, so to maintain the original roll-off frequency, C11 increases to 2u2. That's a fixed solution, but you could still make R3 variable. The increased gain at the output will make it seem a bit noisier.

A lot of these circuits are a little "cut down". You wouldn't normally expect the input to a recorder (which the delay chip is) to have no means of setting the recording level. Still, that's what we usually have in order to keep the knob count down. Guitar stompboxes are usually made with the understanding that average passive guitar signal level will be maintained at all the points in the chain. We all know that can't actually be guaranteed.

When 2 signals are mixed together the output can be a little louder. If they were identical they just add and the output gets +6dB boost. With a time delay between them, they will from moment to moment add or subtract somewhat and, on average, sound no louder than the original dry only.
So a lot of chorus/phaser/flanger circuits get away with fixed 50:50 wet/dry mixes with no gain and compare well to bypassed level.

++If you place this 50:50 wet/dry in a send-return loop, then the return is mixed again with the dry and may result in an 50:100 wet/dry ratio. With effects that sound best/deepest with 50:50 mixes, that breaks the balance++

Some amps have 4 jacks for 2 effect loop connections. A Series connect that inserts FX in the pre-amp path and a Parallel connect that is the classic send from pre-amp and return mix before power amp. The series for things like distortion and the parallel for things like reverb. The owners manual for this amp will advise only having 100% wet delay in the Parallel loop.

All delay effects used in a send-return loop work best if there is no dry signal in them. The wet/dry mixing is going to happen anyway where the return gets mixed back in. A simple switch could be used to cut the delays'  built in dry feed for use in a loop and needn't worry about any fancy anti-pop precautions as you are unlikely to change it in performance.

[Kipper4]

Although it's a small parts nice little add on I've found that the jfet blender does not play well with all circuits Lars.
There are other (op amp) alternatives. IIRC even RG did one.
You may find it more tweekable and the input output impedances better suited to your needs.

I keep one handy (made on a perfboard cutoff) as a breadboard add on. I think I put alligator clips on the ins and outs to hook it up.

I have several other such cutoff circuit snippets too. Bax tonestack, active tonestack, etc.

Rich

[lars-musik]

If the input just to the delay is padded down then the wet is going to sound quieter compared to the dry. A proper wet/dry mix either by a blend control or 2 separate volume controls into that mix can obviously restore the balance, but it might be quieter than bypassed.

Thanks for the very thorough explanation of dry/wet ratios in delay-based modulations. I had to read it more than once and hope that I understood some of it.

However (to fit my more simplistic mind): If I just pad down the whole she-bang (I'll attach a schematic of the circuit as I built it minus LFO) - meaning a voltage divider or a serial resistor (still not sure which is the right way to do do it) - at the input and then make up for the gain loss at the end (before the output) by some means of boosting circuit? I know it doesn't sound super elegant, but would it be wrong?

At the moment:



Better?

There are other (op amp) alternatives. IIRC even RG did one.

Hi Rich,

I only came up with this moosapotamus schematic. I am sure, RG has all the building blocks for a parallel looper on his site - but at the moment I fear I'd cause more problems if I start assembling them on my own.

If you happen to find some other opamp-based blender on your hard drive I'd be glad to see if it does the trick.

[anotherjim]

Yes to that use of the Paralooper.
Having a pot right at the input will, of course, work to cut the level as required. It has up to x11 make up gain at the output - that should be more than enough.
The input pot at 100k will load some top end out of passive pickups.

I'm going to put out a question myself.
Could the input bias resistor (R4) be a 1M pot (wiper to U1 pin3) without causing crackles?
For sure I've seen it done with dual supply circuits where the bias is from 0v.

Is modifying that delay out of the question? Although I can see the use of having a pedal that can master any other pedal.

[lars-musik]

Yes to that use of the Paralooper.

That's very good to hear!

The input pot at 100k will load some top end out of passive pickups.

I had only that active low-z pickup in mind, but of course you are right - the pedal should work with "normal" PUs as well.

A 1M potentiometer it shall be then.

Is modifying that delay out of the question? Although I can see the use of having a pedal that can master any other pedal.

For that particular built I'd really like to leave it as it is. But for future builds I'll keep your deliberations in mind!

Thanks again!