Input cap toggle on Jack's MOSFET Boost... How can I do this an not have a pop?

[dap9]

Hey all.  I've built several MOSFET boost pedals, but the ones that I try to add a toggle that goes between two different input caps always pop.  I've tried several different ways of wiring up the toggle, but can't figure it out.  Someone suggested making it so one cap is always in the circuit and then adding another cap in parallel.  I tried figuring this out w/ a DPDT toggle, but after 2 hours, got nothing.

So far, the closest thing has been with a SPDT.  I ran a wire from the footswitch to the center lug.  Each outer lug had one leg of each cap.  The other legs were joined and  I ran a wire from there to the perfboard.  When I did that, I'd get one initial pop when I first turned on the pedal, but that would be it.  There's gotta be a way, right?

If I did the parallel thing, how do I figure out which caps to use so I have the stock value of .001uf and one of .047uf?

Thanks for anyone taking the time!  And FWIW, when wiring up the footswitch, I kept getting pops when I used the diagram on Dano's (beavis) site, but the diagram from JD's site (GGG) took care of this - that and a 1m resistor at the output.

[aziltz]

If I did the parallel thing, how do I figure out which caps to use so I have the stock value of .001uf and one of .047uf?

well if .001uf is stock, and you add 0.047uF in parallel, you have 0.048uF which isn't far off enough from the intended .047uF to be noticeable.  That's what I've seen, at least.  If you want to double the Cap value, add an identical one in parallel, if you want to raise it by 10x, add a .01uF to a 0.001uF.


I've done this in a mod on the OCD circuit, but it was a cap mixer instead of switcher, so I didn't experience any pops.  Hope this helped.

[R O Tiree]

You are using pull-down resistors? If not...

It's very common practice to put a fairly large resistor ahead of the input cap. It presents a very large input impedance to the preceding pedal/guitar (a good thing for max signal transfer), it doesn't bleed more than a few nA of signal away to ground and it stops the "free" end of the cap adopting the bias voltage of the input stage of the pedal. This is what happens when you don't have an input resistor and then engage the pedal:



You can see that I used the exact same component numbers as on the schem at AMZ. I added R7 to simulate the high input impedance of the next pedal/amp. See the sudden jump up to about 2V as I engaged the stomp-switch and a wait of about 7ms before it settled down and gave me a signal at the output (the red trace... blue trace is input signal). That's the POP you get with no pull-down resistor.

OK, here's what happens when you have 2 input caps and try to switch between them. The green and pink traces are the voltages at the left hand ends of the input caps:



See how the output trace (red) jumps about by about 1.5V every time I flipped the switch? More POPS, then.

Lastly, I put a 1M resistor to ground ahead of each cap.



See how much better-behaved the pedal is when I flip the switch? I changed it about 5 times and the pull-down resistors kept the two caps' free ends within about 10mV of where they should have been at all times, unlike in the previous picture, where they floated around about 1V away from datum (0V).

Hope this helps.

[aziltz]

that's great info.

do you think you can put the switch after a single pull down and get the same effect/solution?

[R O Tiree]

Like this?



You can see how the right hand ends of the caps try to adopt the bias voltage and then hold it. More popping.

And a single pull-down ahead of the switch would achieve nothing except to ground the switch - the left hand ends of the caps would still float, so popping again.

[aziltz]

i see.  i was thinking the pull down resistor, then switch, then input caps, but looking again, i think we need the resister to pull the cap to ground keeping it from floating.


i think i've read that the popping can be avoided if the connection isn't broken.  Instead of toggling between, use an off-on switch to add a 2nd cap in parallel to a fixed/permanent cap.  I believe this is the same as a cap mixer (see input bass trims or "range" controls) but with a switch instead of a pot.  Not that its any better than the extra pull down, but it might use less parts.

[R O Tiree]

Like this...



Not quite as bad as before, but still pops quite badly, especially as you first engage.

[GibsonGM]

What about a make-before-break switch? That should allow any bias voltage floating around in there to equalize before the contact is "let go", maybe?
Personally, I prefer the pulldown resistor idea ;o)

[R O Tiree]

Ahh, I see. Like this, then...



Fine apart from the slight clipping from 285ms to 300ms or so. Much better behaved, though. It's hard-clipping, so it would "blat" for about 100th of a second (more, I guess if the gain was set higher than 1.5 as I have simulated it here).

Having said that, how much is a 1M resistor to give you no popping and no other weird behaviour, like clipping as the bias shifts slightly? 2 cents?

[aziltz]

well that sets the ballot straight!

the individual pulldowns seems to be the way to go, i'd never thought of that before.  For me though, if its worth having a toggle, its worth having a pot there.

carry on!

[Vitrolin]

...and what would happend if SW1 shortens C1 instead of leaving it with one side disconected?

[R O Tiree]

There'd be no DC-blocking with the switch closed (which is the function of the input cap, after all, isn't it?). So, if there was any residual DC from the previous pedal, then it would pop for Britain.

Edit - Turns out, when I ran the sim, that the gate of the MOSFET was hooked to ground by the shorted switch through the pull-down resistor, so that messed up the bias to the extent that no signal was present at the output at all!

[R O Tiree]

DC-blocking works both ways, then. Strips off residual DC from the previous pedal/stage and preserves bias looking at things from the other side of the cap.

[dap9]

Holy canoli!  Thanks RO!!  Some of that explanation is still a little over my head, but I think after several reads and re-reads, it'll make more sense.  Excellent description and visuals!

Azlitz,  The pot idea intrigues me, what value would you recommend?  I'm assuming a Linear pot would be preferable, right?

EDIT:  And that was the ticket!!!  Just made those tweaks and sure enough, that pop is gone.  Thanks so much!

[R.G.]

It's worth noting that *anything* which prevents the capacitors from "relaxing" into a different DC voltage works.

You can achieve the same result with resistors in series with each capacitor to the signal source, and using the cap selection switch to short out the series resistor for the selected cap. In this case, the resistor holds the outboard end of the cap at the signal source DC level.

Thou shalt not switch DC levels on thy signal line, lest there be pops.

[aziltz]

Azlitz,  The pot idea intrigues me, what value would you recommend?  I'm assuming a Linear pot would be preferable, right?

here's a schematic similar to a range master,

http://www.generalguitargadgets.com/diagrams/joam_sc_fb.gif

same idea as the "Bass" Control.  I've used 50k before with good, smooth results.  If its a linear pot, you can flip the orientation to get it to sweep backwards or not.  Use whatever cap values you like, but the larger cap has to go with the pot.

[Mark Hammer]

Jeez Louise!

Just stick two caps in series and use a toggle to shunt one or the other.  Each cap will have every opportunity to drain and pops will be gone.

[R O Tiree]

Jeez Louise!

Just stick two caps in series and use a toggle to shunt one or the other.  Each cap will have every opportunity to drain and pops will be gone.

Errr... clearly not, then.

When C6 is shunted and you switch over to shunt C1 instead, the voltage at the junction of C1 and C6 goes instantly from Vbias+signal (5V +/-) to 0V +/-. BANG!. Vbias drops to 0V and takes a while to re-establish - until it does, the MOSFET is mis-biased. So, that's a 3V BANG at the output and a 45ms hiatus while Vbias is re-established (as C6 charges) and the circuit functions properly again.

When C1 is shunted and you switch over to shunt C6 instead, the junction of C1 and C6 is at 0V, which is then connected suddenly (and drastically) to Vbias (the potential of the "open" terminal of the switch). Not surprisingly, as C1 accepts this voltage suddenly applied to its right-hand end, Vbias drops to 0V once again and the pedal BANGs once again and it takes a wee while to re-establish Vbias (only about 7ms, as the cap is much smaller and doesn't take as long to charge up).

To re-iterate what R.G. said:

Thou shalt not switch DC levels on thy signal line, lest there be pops.

[aziltz]

i think he meant one or the other, but not both.  i've been wrong before though.

what happens if C1 is permanent and C6 is either shunted or not?

[R.G.]

Capacitors distribute DC initially in the inverse of their capacitance as the initial charge is absorbed, then they settle to the ratio of their DC leakages over the long time.

Mark is correct if the connection of the two caps is grounded through a big resistor. This removes entirely the possibility of the two caps distributing DC among them, unless the resistor is of the order of 1/10 of the leakage resistance of the cap with the lowest resistance.