cool reverb apart from switch spdt 'pop' issue

[deadastronaut]

@sam: 2 fets?..

1 to switch bypass etc..

1 for led?..is that what you mean?.

[samhay]

nope - One for each side of the SPDT you are using at the moment (they both go to ground, obviously) - as you turn one on, you turn the other off.

[deadastronaut]

sorry man, i'm lost.. diagram?.

[samhay]

Sorry - numbering is not right, and I don't think I have drawn the FET switch correctly, but the DPDT wiring should give you the idea.
I have drawn it with the FET shunting to ground like the switch does at the moment. It would probably make more sense to put the switches in series like you did in the old bread board version.

[R.G.]

The JFET switches approach is one I've suggested privately to DA.

I would do it a bit differently, as the exact source of the pop when signal is present is not clear, and as the JFET switching he did before did not kill the reverb signal.

I would use a combination of series/shunt switching. I'd use N-channel JFETs for the series switches and P-channels for the shunt switches, and float the whole thing on half of the 9V supply. That way, you can use one SPDT switch to operate the switching. The N-channels are followed by a P-channel. When you want the channel on, the N-channel needs a high input signal, which turns the matching P-channel shunt off and lets signal through. When you close a channel, a low signal turns the N channel off and the P-channel on, and gives better isolation than a single series switch. Unfortunately, you need two of these.

I would also buffer the signal after the switch before feeding the reverb. It is possible that this one step would fix things, as I have some questions about what happens to the signal to the reverb when there is signal present and the switch changes.

[samhay]

RG - I thought of using one N- and one P-FET in this configuration, but figured Rob probably doesn't have any P-FETs in his stash. If I'm wrong, then you can save yourself half a DPDT. The series/shunting switching is a good idea if this doesn't do a hard enough mute.

The signal is buffered before the reverb by IC2c. However, Rob - does the pop change depending on where the mix pot is set?

This is another way to do the FET switching. This time, the FETs are in series like the old version. Again, one turns off when the other turns on.

[deadastronaut]

hi sam .

pop is pretty much the same whether dry of half wet etc ..(just tried it through sound forge.)

and no i don't have any P FETS..

i'll pop that 2 fet circuit on a seperate breadboard and link it over, cheers man.

so i wire up the dpdt as above still?

nice to see R.G join in the headache too..

[samhay]

yeah - footswitch is the same. The LED on the botton switch is your indicator, so isn't necessary.
Just check the FET switch is drawn correctly - it should be the same as the version you tried with slacker's 'slow switch'.

[deadastronaut]

yep, looks the same as ians 'slow switch' ...

[merlinb]

RG - I thought of using one N- and one P-FET in this configuration, but figured Rob probably doesn't have any P-FETs in his stash. If I'm wrong, then you can save yourself half a DPDT. The series/shunting switching is a good idea if this doesn't do a hard enough mute.

The signal is buffered before the reverb by IC2c. However, Rob - does the pop change depending on where the mix pot is set?

This is another way to do the FET switching. This time, the FETs are in series like the old version. Again, one turns off when the other turns on.

Change R10  to 10k and move it to the other side of R21 (do the same things for R12). This will improve the offness, with negligible change in on-ness.

                  R21, 1k
JFET----o----/\/\/\---
           |
           |
          |  | 
          |  |  R10, 10k
          |  |
           |
           |
          Vb

[deadastronaut]

right, soldered up my switch, got it on breadboard wired etc..as per schemo  (will try merlins suggestion in a bit )

the good news.. very , very  silent switching back n forth whilst hammering a chord etc...excellent. ...even with the led so i could tell it was on/off..

but, the weird news is, when bypassed i get a few seconds of clean then the reverb oscillates and rises through..to mass oscillation

this is how i wired it up..




i will now try merlins suggestion too.

edit: oops missed the note about R22..

EDIT: yep still osc on bypass..

[samhay]

Well that's mostly good news I guess.
Merlin's suggestion might help as it will keep the wet side of the switching relatively low-impedance. If not, you might want to try the other FET switching idea.

[deadastronaut]

yeah the switch is nice , very promising....

so i just take vb from between R10/R21...when i put them in series..then 10uf >?


edit: tried that but get massive oscillation on both on/bypass..

if i change R21 /1K to anything else its wild osc...

heres how it behaves as per schemo..no pop..but osc on bypass.
https://dl.dropboxusercontent.com/u/7464107/tester.mp3

[R.G.]

edit: tried that but get massive oscillation on both on/bypass..

if i change R21 /1K to anything else its wild osc...

That's one of the reasons I suggested that you buffer after any switching. A low impedance source to both the reverb input and to the bottom of the reverb mix will effectively keep the reverb from feeding back into itself. That's why a 1K works there - it's just low enough to keep the oscillations out.

If you can get enough isolation from a simple series JFET switch, you may be OK with just buffering after the switch into the reverb section.

[deadastronaut]

^ cheers R.G.

for my own reference, and you guys too,  i'll stick this up (how i have it at the moment)



sorry its small, i had to adjust things to A3 in my editor.. ..i can blow up (bad choice of words ) the image if needed.

now this buffer lark?.. ..from C4/C9 ?..

[samhay]

You've drawn the footswitch incorrectly - only one FET gate should be connected to V+ at a time.

[deadastronaut]

oh yes..of course.

i'll correct it. cheers sam.

edit: corrected: 

[deadastronaut]

so which buffers..and where to put them. straight from C4 .  and straight from C9?

Bottom buffer here?..(but using 072) wil this then make the overall output non inverting too? as well as sort feedback bleed over on bypass...questions , questions..
http://www.muzique.com/lab/buffers.htm

btw i tried the switching out today again, no noise whatsoever as it is....excellent'e

[samhay]

If you want to use a discrete transister, then a common drain MOSFET buffer could go between R11 and C4 and will only cost you a FET + 1 resistor.

Oh - you almost certainly don't need to buffer the dry side of your switch.

[deadastronaut]

cheers man,  i have a 2n7000, but no 9v zener diode ...hmmmm..(looking at the link mosfet buffer)

i also have fets j201'5458/5457's..