How lazy can I get with this boss switching?

[blackieNYC]

I got a boss switching system working on the breadboard, and started to build it on perf.  It's for a momentary loop insert that is very quiet in the switching.  I'm happy with it.  I am envisioning this at the end of the signal chain, inserting some wackiness in brief spots.
In building it now, I was testing it with a signal generator and a pedal at the input or output. I did a frequency sweep and there is no loss. I realize I had left out the input and output buffers, connecting pedal inputs and outputs Boss-style to a 1 meg to ground, series 1uf cap, and 4.5v bias 100kohm pull up resistor.  Since this will never be the first device in my chain - The input buffer would allow me to do this right after the guitar. I think I can skip the input buffer, and I don't see why I would want an output buffer at all - the output of the input or inserted pedals will be capable of driving a load already, and there won't be long cables involved. I'm really pressed for space in the pedal.
So with a pedal before and after this (2SK30 fet) circuit, will I be happy with skipping the buffers?
http://www.freeinfosociety.com/electronics/schemview.php?id=116

[R.G.]

That's a lot like asking "Will I be happy I picked the lunch special?". The devil is in the details.

If no one but you will ever use this thing, then the note about it never being first in a chain may be true. Frankly, I suspect that even if this is yours and yours alone, never sold or passed to anyone, you're going to one day try sticking a guitar right into the input. It is a dead certainty that if you give or sell it to anyone else, they will try it with a guitar right into it. So that goes into your happiness scoring equation.

Likewise, the bit about not needing an output buffer depends on whatever is plugged into the back of the thing having a much higher impedance than whatever is plugged into the loop. This is likelier to be true than never having a guitar in front.

This same set of things would be true whether you were using JFETs, CMOS switches, or even a hard-contact switch or relay.

And speaking of which, if it's for a momentary punch in, why not use just a DPDT hard contact switch? There has to be a foot on the button when it's active anyway.

[blackieNYC]

"Will I be happy with ...."  I chose this wording with you in mind, RG! Knew that would lure you in!

I tried a momentary 2p2t - you replied to this in another post, and I don't think I made it clear that I want to be able to kick this in mid-note, while playing, and only the FET switching was smooth enough.
By including a toggle switch that connects the stomp switch to one of the flip flop transistors' bases instead of that (Boss) junction of two 470pf caps, I can change from momentary to latching. Who told me this? I'll go back and find that to give proper credit. 
Thanks. I think I'll step up the enclosure size so I can include the buffers.  I was getting very ambitious. 

[Mark Hammer]

So, while Allan and RG have embarked on a thread about FET-switching, here's a brief digression that emerges as a tangent from the thread.

You'll note that the standard arrangement in Boss and similar FET switches, is that the flip-flop output feeds the gates of the switching FETs through a 1M resistor and diode, with a .047uf cap to ground from the resistor/diode junction.  The resistor and cap add some lag time to the transition from one state to the other in the flip-flop, such that the FET isn't turned off or on so suddenly as to produce a pop.

But here's what I'm wondering: Could that lag be slowed down even more, such that a standard Boss-type flip-flop could be used to produce a "morph" from one path to another, that took about, say, 1 second to transpire?  Or is it the case that the FETs demand certain things that preclude that?

[PRR]

> Could that lag be slowed down even more....

Well, try it. 1uFd is 50X slower. 1 Second may mean finding very large film-caps (electrolytics tend to leak too much) or going 10Meg on the resistors. But a few minutes with a hot iron should tell more than talking/typing.

A probable ill-effect: the switch tends to distort as it goes through the zone between on and off. For small signals and a fairly fast transistion, your ear does not notice. A drawn-out fade, or high levels, may be more disturbing.

Also a long fade "exposes" the fact that the 50% points do not coincide. You may have a moment of silence; or excess overlap; I'm not going to work that out now. Again a hot iron tells all.

[Mark Hammer]

I'll try it out tomorrow.  I have some 1uf and 2u2 non-polarized film caps.

[blackieNYC]

Does a FET based tremolo prove the concept, Mark?

[Mark Hammer]

Not sure.  Now that I think of it, doesn't the diode feeding the gate kind of defeat the idea of a smooth transition?  Note that when FETs are used for tremolo, the LFO feeds the FET gate directly.

[amptramp]

Paul has raised a legitimate concern that the signal may go blank or overlap with a long time constant.  You can adjust this by replacing the resistor with two resistors, one of them in parallel with a diode.  In one direction, the resistor across the diode is shorted when the diode is forward biased.  The sum of the resistances is available in the other direction.  You can tune the circuit for best overlap / gap in this way.  The diode can go in either direction depending on the effect you want.

[PRR]

Or you can take "lazy" out of the budget and use good VCAs.

[Digital Larry]

And don't forget that "crossing at 50%" is not the goal if they have a linear gain characteristic - you'll have a volume drop in the middle.  THAT Corp (who manufacture high quality VCAs) have an application note about a high quality pan circuit.  It uses a fairly elaborate circuit which does piecewise approximation of the proper control curves to get close to constant power output.  http://www.thatcorp.com/datashts/dn120.pdf

[blackieNYC]

Ooh cool! Thanks!