Walco Chord and Note Sustainer build report

[wui223]

i use a 2N5458 for FET to replace 3Sk30A but i doesnt work i try BS170 also it also have no sound at all, somebody pls help me out

[col]

I tried a 2N5457 for the FET and it just generated distortion. A 2N5459 was best and a J201 close second. If you pm me with your e-mail I can send you the stripboard layout I used but I can't access the internet until next Monday or Tuesday. Best Transistors I found for the others were BC184C for the two low gain and a BC549C for the high gain. They still havn't got rid of the hiss completely and it is still too noticable to be used for recording or at a gig. I built the Dynacomp clone from Torchy's vero layout and it was superb so I've not bothered with the Walco since. If I could get it to be quiet I'd box it up as it sounds so good and try it head to head with the Dynacomp.

Col

[Eb7+9]

When experimenting with different FET's you'll want to replace the biasing resistors by trim-pots ... try 5k for source and 20k for drain - 330 ohms in the source ckt is very low for most of these ... that would be a likely source of clipping ...

Then try doubling or tripling the shunt transistor and reducing input resistance to see if you can get the noise down ...

As well, you've got a Germanium diode clamping the output - this will give you only about 200mV of headroom max ... distortion caused at that point would depend on the output of your instrument and how much gain is developped in the chain ... if you want to remedy this potential problem while maintaining the signal path design you can duplicate the last stage and use one output for the feed to the detector and the other as signal output ...

~jc

[David]

As well, you've got a Germanium diode clamping the output - this will give you only about 200mV of headroom max ... distortion caused at that point would depend on the output of your instrument and how much gain is developped in the chain ... if you want to remedy this potential problem while maintaining the signal path design you can duplicate the last stage and use one output for the feed to the detector and the other as signal output ...

~jc

JC:

Would headroom increase if a silicon diode were substituted for the germanium?  How about if R.G.'s trick of using a transistor as a low-leakage diode were used?  The beauty of this thing seems to be in its simplicity as well as in its capabilities.

[Nasse]

My comments were based on Elektor circuit published 1987, I´ll try to post re-drawn schem or something maybe tomorrow. Don´t know if it sounds any good or really help noise with it´s some lower resistances and that cap/noise gate option. But nice sounding clips and easy and cheap circuit too that walco project.

[Eb7+9]

Would headroom increase if a silicon diode were substituted for the germanium?  How about if R.G.'s trick of using a transistor as a low-leakage diode were used?  The beauty of this thing seems to be in its simplicity as well as in its capabilities.

... feedback-loop compressor circuits can exhibit lots of MOJO ... seems especially so if the envelope detector and gain reduction circuitry are both quite non-linear ...

I don't see how R.G.'s idea would fit in here though - maybe he can chime in ... using two diodes like it's done here is a common idea for envelope detection ... you can use a silicon diode at the output but clamping at -.7v instead of at -.2v might mean there's not enough average DC left on the cap to fully turn "on" the shunting transistor ... but try it anyway and see what happens ... I would leave the second "switching" diode as Ge or it will also rob more DC off the cap again ... because the signal levels fed to the detector circuit are not very high you want these diodes to be as ideal as possible - hence the choice of Ge diodes with their smaller turn-on voltage, especially in the switching position ...

I like the idea of duplicating the last stage because it leaves you open to increasing the gain in that duplicated stage or even adding another gain stage allowing the injection of greater amounts of signal to the detector network (signal inversion doesn't matter) ... this in turn would help activate the shunting transistor(s) even more producing lower "average" resistance in the devices and thus making it possible to use an even smaller input devider resistor ... yielding lower noise from the overall circuit ...

[Nasse]

http://img.photobucket.com/albums/v411/Nasse/lousy.jpg

Sorry the pic is very low quality, almost unreadable re draw of Elektor compressor summer 1987

But the transistor attenuator at input is somewhat similar like walco thingie, in principle