Walco Chord and Note Sustainer build report

[brad]

Oh...it looks like it isn't necessary after all  :oops:

Horrah for weird pedals  

[David]

Oh...it looks like it isn't necessary after all

Ummm...  maybe, maybe not.  Look, these guys are trying to minimize cost and maximize profit, right?  If they specify a dual-gate MOSFET, there must be a reason.  Check Col's thread on the Maestro Sustainer.  There's a reference in there to the NTE equivalent of the 3SK30.

I can't help but wonder if this dual-versus-single gate business accounts for the noise.

[brad]

It does seem odd that they'd use such an odd part.  The NTE222 substitute is one hella expensive transistor too.

This would make a really cool general DIY project with a bit of user-friendly optimisation.

[Doug_H]

That sure has a nice sound and it looks like a simple build. If the 1st stg is supposed to be a mosfet, maybe someone should try a BS170 or 2n7000 in there and see how it works, provided the bias network supplies a high enough threshold voltage to the gate (2v for bs170 IIRC...).

Doug

[David]

That sure has a nice sound and it looks like a simple build. If the 1st stg is supposed to be a mosfet, maybe someone should try a BS170 or 2n7000 in there and see how it works, provided the bias network supplies a high enough threshold voltage to the gate (2v for bs170 IIRC...).

Doug

My humble apologies.  I mistakenly typed MOSFET instead of JFET.  This circuit calls for an odd duck called a dual-gate JFET, a 3SK30 IIRC.  I found a NTE sub for this at about $9.00 US.  If there were a way to implement it with "normal" JFET or MOSFET, I'll bet a lot of people would go running for their soldering irons.  I know I would!

What does a dual-gate JFET do that we couldn't somehow "reverse engineer"?

[David]

That sure has a nice sound and it looks like a simple build. If the 1st stg is supposed to be a mosfet, maybe someone should try a BS170 or 2n7000 in there and see how it works, provided the bias network supplies a high enough threshold voltage to the gate (2v for bs170 IIRC...).

Doug

My humble apologies.  I mistakenly typed MOSFET instead of JFET.  This circuit calls for an odd duck called a dual-gate JFET, a 3SK30 IIRC.  I found a NTE sub for this at about $9.00 US.  If there were a way to implement it with "normal" JFET or MOSFET, I'll bet a lot of people would go running for their soldering irons.  I know I would!

What does a dual-gate JFET do that we couldn't somehow "reverse engineer"?

[Doug_H]

What does a dual-gate JFET do that we couldn't somehow "reverse engineer"?

I don't know if it make much difference if the clone that Rick built with an mfp102 sounds like the clip, which he claims it does. It would be fun to experiment with different device types in this circuit though.

Doug

[cd]

Check out this link:

http://home.planet.nl/~heuve345/electronics/course/lesson18.html

Naturally my theory sucks but it seems to me G2 being grounded is used to set the gain of the rest of the MOSFET.

[brad]

My humble apologies.  I mistakenly typed MOSFET instead of JFET.  This circuit calls for an odd duck called a dual-gate JFET, a 3SK30 IIRC.  I found a NTE sub for this at about $9.00 US.  If there were a way to implement it with "normal" JFET or MOSFET, I'll bet a lot of people would go running for their soldering irons.  I know I would!

What does a dual-gate JFET do that we couldn't somehow "reverse engineer"?

I found this page with a few specs after searching 'round a bit, and it does in fact say the 3SK30 is a dual-gate Jfet:
http://www.americanmicrosemi.com/products/specs/3SK30.phtml

However, the NTE222 is a dual-gate Mosfet!  The page cd linked to says that a dual gate mosfet is simply two regular mosfets in series (and has a nice little diagram to show how they're connected).  So perhaps a dual gate jfet could be made in a similar way by connecting the drain and source of two single jfets.

[brad]

The confusing thing is that the second gate is just connected to ground in the Walco.  According to the page cd posted, dual gate mosfets are usually used for AM or TV applications where the second gate is connected to either a positive voltage or a second signal that modulates the gain of the first.

So I can think of two circumstances for the 3SK30:

a) the design called for a regular jfet, but the manufacturer used a dual-gate jfet with the 2nd gate tied to ground instead because they had a big pile of 'em lying around the factory (perhaps they usually made TVs or radios?)  I don't see how tying the 2nd gate to ground can be taking advantage of the transistor's automatic gain control function!  :?

b) the 2SK30 is actually a dual gate mosfet and the americanmicrosemi website just listed it as a jfet by mistake.  I only say this because I can't find any evidence of other dual gate jfets online.  Again, if it really is a mosfet...see above.

DAMN OBSCURE COMPONENTS  :x

[wui223]

Yes i agree with u , this is a norm in industry just use whatever availalbe with minimum cost. But i wonder how could we improve the noise quality? Being playing with hiss noise is driving me crazy

[brad]

I doubt you can do much about the noise beyond using good low-noise components.  The more the signal is compressed, the more audible the noise floor.

[Nasse]

What I understand the bipolar transistor is the gain control thing, that fet looks like a high impedance for it

Quess you could scale resistors for lower impedance but how low and does that help much dunno

[Ge_Whiz]

Seems to me that if the front-end is a dual-gate FET (of whatever variety), it would make more sense to have the rectified envelope from the 2SC644 going to the second gate. Are we sure that the schematic is correct? Did someone say the second gate is earthed?

Could be that the intention was to use the dual-gate device as a modulator, they ordered the parts, and then found that it worked better running both signals to the same gate.

[Steben]

I can't believe the circuit won't work with well-chosen cheaper parts. A JFET input is possible no? I don't see any reasons not to use that. The whole idea is about a variable voltage divider (input resistor and 10M to ground parallel to FET). 3k3 as input would work as good with less noise. the 10M could be swapped with 100k or 10k, miuch depends on the resistance of the choosen FET.

This looks like the ELEKTOR compressor/overdrive from 1986 (look at Aron's schematics). Also diode filter to FET, but in feedback loop of TLO71.
Anyone remembers? Those were lower resistor values.

[col]

Have a look at my Maestro Sustainer thread. I added to it in the past couple of days so it shouldn't be too hard to find. I have listed the parts I got the best success with.

Col

[Nasse]

If someone has a working circuit and wants to experiment I believe if you connect a small cap across that 10 megaohms resistor (from collector of 2SC644 to ground) it might cut noise at minimum attenuation, and it might work like a noise gate, filter opens up with larger signals. Dont know what value the cap should be, maybe it could be around 10 nF for start

[wui223]

U mean parallel with it or series ?

[Nasse]

Parallel

I´m off for few days

[Eb7+9]

If you want to experiment with lesser input resistance to reduce noise you'll have to ratio down the dynamic resistance of the shunting BJT by paralleling another device or two to the first one ...

try this: stick one more device in parallel (Base2Base, etc) and lower the input series resistance so that you have the same range response to signal envelope you had before (use a 1meg pot at first) ... the 10Meg resistor is there to provide DC reference to the FET's gate only and can stay as is, or lowered to 1Meg once you've got it working with releatively less series resistance ... then add another sunthing device and lower input resistance some more till you get good balance again ... the S/N should improve somewhat ...

This paralleling of shunting devices will also reduce the Release time through an increased sum in Base currents - yielding faster relax-slewing ... this faster decaying can be compensated for to a degree by increasing the averaging/filter cap value ...

~jc