True Bypass with LED schemes

[Steve Mavronis]

I want to discuss these two true bypass with LED methods. The top schematic circuit is based on an JD Sleep example. The bottom circuit is based on Gaussmarkov's symetrical 3PDT wiring. In his excellent article (http://gaussmarkov.net/wordpress/thoughts/wiring-up-a-1590b) about wiring a 1590B, he uses a beautiful symmetrical  3PDT wiring pattern (input and output opposite each other and circuit and bypass halves separated) with an added extra (green) wire in the middle that "feeds the stompbox circuit a quiet input signal when the circuit is by-passed."



I tried to draw a schematic of what he is doing with the 3PDT switch connections labeled like a Excel spreadsheet grid with rows and colums:



Compared to the JD Sleep wiring, is it really doing anything different other than a really pretty and organized switch wiring job? But I do really like it a lot for its symmetry. Don't both schematics do exactly the same thing? Also, how does the circuit maintain a quiet input signal during bypass switching when for a moment in time it is not grounded but an open circuit instead?

Also there is an issue is with schematic consistency where normally the 3PDT poles should be drawn in the same order. Gaussmarkov in effect turns the SW1A pinouts (A1, A2, A3) around if drawn as a schematic:

[amptramp]

I see what you mean.  If the board does not see a grounded input as is the case in the bottom picture, it will propagate noise and if the effect contains a delay such as a flanger or an echo, this noise is not cleared out when the effect is switched on.  You may have around a second of noise propagated through the effect and if it is configured to repeat such as an echo, it may be with you for quite a while.  If there is no delay in the effect, there is still enough of an effect from input capacitor charging or discharging that you will get an unpleasant thump when switching.  This may not be a problem if the switching only occurs between songs in a set but if the switching occurs during the song (say, a song starting with a clean guitar than changing to the shimmering sound of a flanger), you will have trouble.  The top schematic does not have this problem although it can exhibit switching noise.

Maybe this is the reason commercial pedals like the Tube Screamer and most of the Boss pedals use electronic flip-flop switching with slow-down RC time constants in the gates of the switching FET's so there is a gradual change from one output to another with no thump.  (Or it may be just the expense of a 3-pole stompswitch.)  The only problem with these pedals is they do not necessarily always come up in the same configuration, either on or off.  This can be fixed easily with an initialize circuit as shown here:

http://www.diystompboxes.com/smfforum/index.php?topic=83980.msg698817#msg698817

This is much simpler than the Boss / TS flip-flop.  In spite of the support for true bypass from a lot of builders, it is only a benefit if you have a high-impedance pickup and do not need to switch anything during a song or if you have so many pedals in the chain that the signal would be degraded by buffers.

[darron]

I've got a write up that I've done on methodology here if you can be bothered reading it: http://www.dazatronyx.com/support/what-is-true-bypass/

scroll down to "let's get technical" to avoid the super boring parts.

[Brymus]

In the Gausamov version 3A and 3C are connected and 1A is conected to ground in bypass mode.
This is because the power is still on to the effect,and it keeps it from effecting (amplifying) whatever noise it picks up at its input while bypassed,by grounding the input (same as unplugging from an amp)
I dont think you drew that right,but I can say for certain without seeing an actual schematic of the switch.

[darron]

In the Gausamov version 3A and 3C are connected and 1A is conected to ground in bypass mode.
This is because the power is still on to the effect,and it keeps it from effecting (amplifying) whatever noise it picks up at its input while bypassed,by grounding the input (same as unplugging from an amp)
I dont think you drew that right,but I can say for certain without seeing an actual schematic of the switch.

the input of the effect in the Gaussmarkov image above does actually ground the signal input to the effect (blue) while in bypass mode. floating effects into opamps and such can also cause self-osscilations especially in high gain circuits.

the diagram can be improved by tying the jumper that goes across the bottom to also connect to the final output which goes to the jack

[Steve Mavronis]

But during the transition of the actual bypass switching nothing is connected for a short momemnt in between states. Switching is not instantaneous. I'll double check to make sure the schematic matches the pinout pattern in case I made any mistake drawing it because I want to depict it right.

It is obvious he's using pole column A for inputs, pole column B for grounds, and pole column C for outputs; plus row 1 for the circuit engaged, row 2 neutral switching point, and row 3 for true bypass. That's the beauty of it really how organized it is to setup the wiring. It just isn't the standard way of drawing the 3 separate poles of a 3PDT switch on a schematic.

I'm actually fine with it since it works. My father takes issue with it being an electrician and electronics technician. He's really partucular about the way you should do things to be uniform, etc. He even has a workplace patent for a circuit he designed:

"Flame detection monitoring system for detecting blockages in blast furnace injection paths"
(includes schematic images) http://www.freepatentsonline.com/6080223.pdf

[anchovie]

That moment in time is far too small to worry about. There are ultra-high-gain metal pedals that use these break-before-make switches for true bypass.

It's not a case of denying that the moment of no-connection happens, it's just whether it's actually noticed or not.

[Brymus]

Am I missing something ?
It looks like the way Steve drew the Guasamov version it wont work when bypassed.
3A needs to be jumpered to 3C not 3B,he is shorting the input when its in bypass,it wont make a sound in bypass mode this way.

[slacker]

Yeah the Gaussmarkov version in the lower picture is drawn incorrectly. SW1A3 not SW1A2 should be connected to SW1C3, which is the same as what Brymus said. Then SW1B3 connects to SW1A1 and there should be no connection between SW1B3 and battery negative.

[Steve Mavronis]

It looks like the way Steve drew the Guasamov version it wont work when bypassed. 3A needs to be jumpered to 3C not 3B,he is shorting the input when its in bypass,it wont make a sound in bypass mode this way.

Yeah I do have A3 and A2 messed up in my drawing. Thanks for corrections so I can fix the schematic drawing later. But (you have me converting row/column to column/row in my head) A2 connects to C3 when bypassed. In Gaussmarkov's illustration he is also grounding the PCB input with that short green wire (when in bypass mode) to the ground lug of the output jack:

So let's double check his pinout using this as a labeling legend so we are talking the same language:

A1 goes to PCB input pad. A2 goes to input jack signal lug. A3 goes to C3 as bypass jumper. B1 goes to LED. B2 goes to output jack ground lug. B3 goes to ground the PCB input - only when bypassed. C1 goes to PCB output pad. C2 goes to output jack signal lug. C3 goes to A3 as bypass jumper.

[slacker]

Yeah that's it. Just to avoid any confusion B3 just goes to A1.

[Steve Mavronis]

I got it. Thanks for helping me understand my errors. I'll redo this even keeping the orientation of the SW1A segment's 3 pinouts facing the same way as the others and route around it. Depicting it backwards was really bugging my father and my drawing being wrong didn't help my case saying I liked Gaussmarkov symmetrical 3PDT wiring better. No wonder my dad thought "he" was doing it wrong when it was really "me" representing it wrong in the first place due to my schematic errors! I would like to use the same wiring as his illustration shows because visually it is so organized and logical.

[UPDATE] I fixed the 3PDT pinout wiring. Does the bottom schematic in the screenshot below look accurate to Gaussmarkov's?