How to wire Stereo Tremolo for true bypass?

[john.hostetter]

Hello all. I have finished wiring the GGG stereo tremolo: http://www.generalguitargadgets.com/projects/26-tremolo/143-stereo-tremolo

I can't find any information about how to wire bypass switching for this pedal. Rather than buy a 4pdt switch (12$!!!) I would rather use a 3pdt and sacrifice the LED. Is this possible? Any help is greatly appreciated.   

[daverdave]

I take it you want a switch that's gonna bypass BOTH inputs and outputs at the same time. I can't see how you could use a 3pdt to do that with true bypass. Yoy could probably use some sort of fet switching. ailing that I'd be tempted to use one switch for channel one and one for channel 2, or something similar.

[MikeH]

I did it using a 4pdt with no LED- they're about 12 bucks at smallbear.  Other than that you'd have to get into relays, FETs, etc.

[john.hostetter]

Hmm. Bummer. I'm not ready to start buying $12 switches. What if I abandoned true bypass? Could I simply bypass both channels at once with a 3pdt? Thanks for the incredibly fast response again, this board is incredibly helpful.

[Mark Hammer]

That would depend on what sort of bypass you envision.

For example, do you want or need to have both outputs viable when you bypass?  Do you need to have full/true/straight-wire bypass?  If a person felt no need for an indicator LED and didn't feel full bypass was that critical to sonic integrity, you could do it with a DPDT.

[earthtonesaudio]

If you don't need true bypass, your options become nearly unlimited. 

With a 3PDT, you could leave both inputs connected and only switch the outputs, but that means each input is loaded by about 100k, or 50k if only using one input.  That's somewhat low and you might notice a treble loss.  Another option is just to kill the LFO by tying pin 11 of the 4049 low.  That will stop the effect but the output levels will be dependent on the depth and level controls.

Then there's electronic switching, where practically anything is possible.  My personal preference would be to use electronic switching that both stops the LFO and gives you unity gain buffered outputs.

[Mark Hammer]

I'm inclined to agree with you there about the LFO.  Unless you have a guaranteed tick-free LFO, it probably IS a good idea to disable the LFO if possible when bypassing a tremolo....or any other modulated effect for that matter.  Of course, the critical question is whether the enabling of the LFO when the effect is switched in again might result in an even more audible tick during the switching itself.  Either way, an electronic switch is probably the easiest way to do this in stereo.

[Processaurus]

Alex is right, the input impedance is too low for the "tone sucking" type bypass (50K worst case scenario, with one input plugged in), which would actually cause tone sucking.

I did a stereo output (mono input) trem and used a FET buffered bypass with the GGG discrete FET buffer, which sounds nice. 3PDT, two poles to switch the output jacks between the effect and the buffered input signal, and there was a pole left over for the LED.

The idea of just stopping the LFO is elegant, though.  That could have the advantage of a super smooth transition between trem and no trem.

[Mark Hammer]

I should have looked at the schematic linked to before I opened my big mouth.

The simplest thing to do is to just reroute R11 and R12 to V+, so that both LEDs are full brightness (or close to it) and the LDR is at minimum resistance (or close to it).  Do that and you'd have two inverting stages in series which would get you an in-phase output.  That, plus a status indicator LED, is feasible with a 3PDT.

The only real challenge here is that, since the R7/R9 and R8/R10 networks each provide a gain of just under 5 for U1a and U2b, how to achieve an appropriate output level for the "bypass" (effect cancellation) mode.

Seems to me that you could use a suitable series resistance between R11 and V+ and R12 and V+ to adjust the resulting LDR resistance such that the bypass-state gain of U1a and U2b are each what you want.

The other approach is to stick the drain and source of a JFET in parallel with the LDR in each channel, and use a good ol' Boss-type flip-flop to switch 'em, using a momentary.  There would be three parallel resistances: the JFET, the LDR, and the P1/R5 (P2/R6) network.  When the effect is on and the LDR's value is high (LED is "off"), the gain of each op-amp gain stage is set by the value of the pot+resistor network.  When the LED in each channel lights up, the LDR value goes lower than the pot+resistor network and exerts a greater influence on the gain.  When the JFET gets turned on and achieves a low drain-source resistance, it supercedes the other two and the gain is governed by the JFET resistance.

The downside here is that, while it is simpler in some ways, a) it involves more parts, and b) it is loudest in bypass mode.

Could we fix that?  Perhaps.  Let's say the LDRs are each bypassed by JFETs activated by the "flip" of the flip-flop.  Can we do something useful with the "flop"?  Perhaps a pair of JFETs actuated by the complementary flop part of the flip-flop could be used to modify the gain of the effect-cancellation mode so that effect mode could be louder OR softer than bypass.

Lets call the JFETs in parallel with the LDRs Q1 and Q2.  Imagine that R7 and R8 are each replaced with two fixed resistors in series R7a/R7b (2k2/1k)and R8a/R8b (2k2/1k), and a JFET drain-source path is wired up in parallel with each one of the added resistors (Q3, Q4).  In "effect" mode, Q1/Q2 are "off" and Q3/Q4 are "on".  This would mean that the added 2k2 resistor in each path is effectively disabled by being shunted by the much lower "on" resistance of Q3/Q4.  So, we have the op-amp gain of 5x being modulated by the value of the LDR and the pot in parallel with it.

Press the momentary switch and the JFET state changes.  Now the whole LDR+pot+resistor network is superceded by Q1/Q2.  And when Q3/Q4 are turned off, the gain of the op-amp is essentially [4k7/(2k2+1k)] = 1.47.  So, we have a bit of oomph to the bypass output if we want, or we can use the volume pot to adjust accordingly.

Does that work for you?

[john.hostetter]

Wow guys. I have been looking for information on all this for a couple days and have found only a fraction of the content you just relayed to me in the span of one afternoon. I can't thank you all enough. I think I am going to attempt this with a buffer because a) I need to learn about them, and b) I think Mark's idea would take some experimenting on my end (I am quite the noob). I'm actually going to build two of these, so maybe for the next one I will try to implement Mark's solution. Either way I will share my results. Thanks again everybody!

[earthtonesaudio]

The more I look at that circuit, the more I want to redesign the whole thing.  

You could do the same job with just a quad op-amp.  Each of the audio paths could be done with one non-inverting amplifier, with LDR in the feedback loop.  Then, the LFO could be done with a single op-amp, and the in/out of phase bit could be done with another op-amp as a unity gain inverting amplifier.  Bypassing the effect would be a simple matter of shorting both LDRs (and killing the LFO probably), which could be done with JFETs or simply by sticking the oscillator high so the LEDs stay bright.

With the LDRs at minimum resistance, each audio path has a gain of approximately unity.  You can ditch the output level control in favor of a gain control (like how it's done in the MXR Micro Amp), and when the effect is "bypassed" the output is unity no matter where the gains are set.


...Sort of like this:

[john.hostetter]

Is this correct? Or do I just buffer after the switch?

[Processaurus]

Note, the switch lugs should be wider than they are tall, to show the orientation of the switch, like so:

_  _  _
_  _  _
_  _  _

[Processaurus]

The more I look at that circuit, the more I want to redesign the whole thing.  

You could do the same job with just a quad op-amp.  Each of the audio paths could be done with one non-inverting amplifier, with LDR in the feedback loop.  Then, the LFO could be done with a single op-amp, and the in/out of phase bit could be done with another op-amp as a unity gain inverting amplifier.  Bypassing the effect would be a simple matter of shorting both LDRs (and killing the LFO probably), which could be done with JFETs or simply by sticking the oscillator high so the LEDs stay bright.

With the LDRs at minimum resistance, each audio path has a gain of approximately unity.  You can ditch the output level control in favor of a gain control (like how it's done in the MXR Micro Amp), and when the effect is "bypassed" the output is unity no matter where the gains are set.


...Sort of like this:

That looks cool, I was thinking it could have been two non-inverting stages as well, but at a quick glance it would seem the inverting amplifier would provide more depth, because the gain with non inverting amplifiers is at least 1, whereas the inverting amplifier can have a gain of less than 1.

There is the trick as well, of making an inverted LED, by hooking the cathode to the LFO opamp's output, and the anode to V+ (through a resistor, naturally), if you like.  Yours looks like it works, though.  Yours looks to be missing a channel's LED.

[earthtonesaudio]

You're right about the difference in maximum depth as well as the missing LED, Ben.  I just forgot the LED but I didn't even consider the depth issue until you mentioned it.  An inverting stage seems to be the only configuration that allows gains both above and below 1 using a single variable resistor.  It is possible I was too quick to discard Anderton's ideas.  Hmm...

[john.hostetter]

Thank you Ben. I've been learning a lot the past couple of days thanks to you guys. It's all coming into focus!

[PRR]

> I want to redesign the whole thing.

Me too.

Anderton was no fool, but sometimes there's other ways to do it. Especially as we demand more and more and more frills, like low-loading and utter bypass, stuff we just didn't worry about back in the day.

> inverting stage seems to be the only configuration that allows gains both above and below 1

This is nearly equivalent but has a very high input and offers a buffered no-FX output as a bonus:



With LDR very light, we have small loss in the LDR+1K load, then 1:5 gain in the output stage (mmmmm... the 4K7 should maybe be closer to 3K9), say 1:4 gain overall.  With LDR dark, we have the ~~20K:1K divider and 1:5 gain for 4:1 loss.

Missing details per original or standard custom. Bipolar +/-9V supply. Input cap may be 0.01uFd. The output volume pot is probably needed to trim the median gain (or trimmer the 4K7 NFB resistor). It is still a full TL074 plus a hunk of CMOS.

A DPDT will select FX or pretty-clean buffered input.

Another pole would be needed to stop the oscillator. This one may not be too thumpy; logical layout (keep the racket away from the nice neighborhood) may leave the no-FX path "dead-clean".

Otherwise: use relays.

[PRR]

If you obsess about hiss... the steep loss in the LDR plus the make-up gain may reveal the TL074's 2uV hiss. Since noise resistance is very low (1K), the make-up amp could be 5532 for about one-third the hiss (in this particular circuit).

You could instead take some gain in the input stage. With 18V total supply, you could probably take the whole 1:5 here, vari-loss with LDR+1K, and take right off the 1K which is low enough to drive any cable. To get a unity gain buffered output, use very low value resistors for the gain-stage NFB (3K9+560) and take buffered output across the lower resistor (560). Now we are down to one pair opamp, which ought to be a TL072 type because it faces into the high impedance of the guitar.

[john.hostetter]

Awesome. I got the thing up and running wonderfully. Many thanks to all those who baby-stepped me through this. BUT! I am still being picky. The only problems I have with it now are that it is not at unity gain with the effect on, and the 3pdt pops. Any suggestions to fix these? I'll post a picture of my cool enclosure ASAP.

[john.hostetter]

UPDATE! I now have a stereo tremolo that sounds and works awesome! The unity gain problem was solved by changing R9 and R10 to 8k2 and R3 and R4 to 220k. This was suggested by Pete W. here:

http://www.diystompboxes.com/smfforum/index.php?topic=57127.0

I now have to turn the volume down to get unity gain (which I like). I added pulldown resistors and the 3pdt has a minimal pop, hardly noticeable.
I hope this information is helpful for others, this pedal is great although it has taken me nearly a year to get it running!