Paia Gator attack-delay mods?

[Rodgre]

So all this talk about ADSRs, and my eternal lust for an EH Attack Decay or a Pigtronix Attack Sustain pedal has gotten me to dust off my old Paia Gator. While I prefer the Gator to my Slow Gear, I still have some wishes that might be fulfilled with a little modding.

I'd like to be able to control the decay as well as the attack time of the circuit. I would also like to tailor the curve of the swell up, if possible.

I've seen some attack/decay controls on ADSR circuits which seem to be not much more than complimentary pots with diodes in opposing polarities. Is this something that I could do to the Gator to control the decay?

It truly is a sleeper of a pedal and it keeps me from spending more money than I should to buy the above-mentioned pedals. Someday on that Pigrtronix though...  I need that one of these days.

Roger

[lars-musik]

Funny, I just typed the message below but before opening a new thread I had the good sense to do one more search on the topic so I'll revive this ancient one:

I finally got a Paia Gator working following the well-known Gaussmarkov (here) schematic and layout. To overcome my previous problems of insufficient sensitivity I used a brute force method by placing a booster (a SHO) in front of the gator.

However, I am still unhappy with the behaviour of the circuit. If I turn up the attack the tone sets in later than with attack turned down, but the ramp-speed seems to remain basically the same. So, the sound simply comes out with a unpleasant delay at longer settings. I'd like the tone to start immediately after plucking the string at all attack settings but with a slower ramp-up at long attack settings and faster ramp-up speeds and shorter attack settings. Do I make myself clear (for a none-native speaker the description comes out a bit bumpy)? Is it my build or is it just the way, the circuit behaves?

Thanks for your ideas!

[Mark Hammer]

It is not the very best idea (though not the absolute very worst) to put a booster in front of the Gator.  It uses a CA3080, which distorts very easily.  So, while the boost WILL provide a stronger signal to the envelope extractor/rectifier part, and make the unit appear more sensitive, that boosted input will also tend to overdrive the 300.  If you like the tone, fine, but if you want to have more tonal choice (clean OR dirty), it would be smarter to improve the sensitivity of the sidechain specifically (i.e., something between C4 and R18).  You can try reducing R12 a bit by putting another resistor in parallel with it (I'm assuming you have it built on a PCB), so that you don't have to remove R12 and reinstall.  A 47k in parallel will give you the equivalent of about 8.2k, which will increase the gain of that first stage.

As for the way the Attack control responds, you may have it wired opposite.

One of the things that might have been useful in the original design is a minimum volume (i.e., initial gain) control, so that it could swell up from that point, instead of from silence.

[lars-musik]

Hi Mark,

Thanks for your analysis. Indeed, I prefer the "backwards  sounds" with a good portion of dirt. I got a Slow Gear boxed up with a Red Lama and got Fuzz always maxed out.

But I guess your suggestions will be of use for the less dirty players among us.

The Attack envelope is more of a problem for me. In the process of rewiring the attack pot (now the slow setting is on the other end of the travel...) I realised that it is named "Attack/Delay" – so I guess I shouldn't be too upset about delaying the notes, now should i?

However what I figure to be a nice workaround would be a clean blend. Do you think the circuit could be modified to provide one or should I go for an offboard solution?

[Mark Hammer]

One of the tricks to getting a decent analog reverse-tape sound (as opposed to using a reverse patch on a digital delay) is to have the harmonic content increase over the swell.  Normally, most of the harmonics produced by a string occur at the very start, after you pick.  As a physical body, the string will damp vibration for the highest frequencies with the least amount of energy, first, and then slowly the fundamental and lowest-order harmonics will decay.  So, if you have some means to introduce more harmonic content as the note swells, you get closer to a realistic simulation.

I'm undecided as to whether it is better to introduce clean blend, or have a control for setting initial gain on the 3080.  Part of what determines that is whether it is easier and less problematic to do the one than the other.  You can see that "cancelling" the effect involves simply connecting the input of IC4B directly to V+, such that max gain is achieved.  I would imagine that something in between +9V and envelope-voltage only would introduce a minimum starting gain, which the envelope adds to.  And that's pretty much what happens with VCA modules in analog modular synths.

The other route would be to take a feed from the output of the LM308, go around the 3080, and treat IC4A as not only an output stage, but a sort of mixer of two signal sources: the gain-varied 3080, and the direct feed from the 308.  What I am unsure of is whether this would create any complications for the envelope following portion of the circuit.  I guess a simple way to tell would be to run a 1uf cap (+ end going to the 308 output) and 15k-22k resistor in series, and connect the resistor to pin 2 on IC4A so that clean and gain-varied are mixed.  You can tack that on to the copper side of the board for experimenting.

[Moonsun]

Hi! I built Gator few months ago and make D3 switchable - when it's on (as in original) I have short release with quick drop, but when it's off, release is long and smooth. Only thing it's not adjustable. Don't know if this mod can be reproductable or I was just lucky - still need much to learn...

[Mark Hammer]

What do you mean by "off"?  Does that mean the connection is lifted, or that it is bypassed in some way?  I wonder if a person could use several Schottky types in series, and select how many to bypass in order to get different decay forms.

[lars-musik]

So many things to try and do! Thanks a lot again, Mark. Unfortunately I am going to rehearse with my band tonight (gosh, what a hobby that pedal-building became that a rehearsal feels like a nuisance!) and won't be able to test the things you suggested in the next few days. I remember that you wrote someplace else about a crazy backward-fx-setup back in the days that somehow created that harmonic increase. Somehow I guess I'll leave this to my computer and be content with a volume swell to my liking for the rare live-occasions.

Both your suggestions (divert the 9V "bypass" and the LM308 – IC4A link) look feasible - even for me - and next time at my workbench I'll surely try my luck with both of them. There has been a little FET-Clean Blend circuit somewhere around here (or over at fsb) that I uncovered on my hard drive, so that would be my third option.

And finally there's moonsun's diode thingy. Why not? If only I'd already knew so much more about electrons and their mysterious ways...

[Mark Hammer]

So many things to try and do! Thanks a lot again, Mark. Unfortunately I am going to rehearse with my band tonight (gosh, what a hobby that pedal-building became that a rehearsal feels like a nuisance!) and won't be able to test the things you suggested in the next few days. I remember that you wrote someplace else about a crazy backward-fx-setup back in the days that somehow created that harmonic increase.

Maybe that simply needs to increase the gain of IC4a (so, 220k in the feedback loop instead of 30k, and a back-to-back diode pair in parallel, with some volume attenuation on the output to compensate for the added gain.  As the swell increases, it eventually hits the clipping point of the diodes.

[samhay]

The other route would be to take a feed from the output of the LM308, go around the 3080, and treat IC4A as not only an output stage, but a sort of mixer of two signal sources: the gain-varied 3080, and the direct feed from the 308.  What I am unsure of is whether this would create any complications for the envelope following portion of the circuit.  I guess a simple way to tell would be to run a 1uf cap (+ end going to the 308 output) and 15k-22k resistor in series, and connect the resistor to pin 2 on IC4A so that clean and gain-varied are mixed.  You can tack that on to the copper side of the board for experimenting.

The 3080 is inverting in this design, so the dry signal will have to go to the non-inverting input of IC4B to prevent phase issues. You could tap the output of IC1 with a 30k(ish) resistor and 1u(ish) cap to lug 3 of an e.g. 10k dry mix pot. Lug 1 of the pot goes to VB (which is no longer connected to pin 5 of IC4A) and lug 2 now goes to pin 5 of IC4A.
The envelope detector is feed-forward, so I don't think the dry mix will have any effect on the way the envelope behaves.

I am not sure that a dry mix is going to be useful either, but I guess it is easy enough to find out.

[Moonsun]

What do you mean by "off"?  Does that mean the connection is lifted, or that it is bypassed in some way?  I wonder if a person could use several Schottky types in series, and select how many to bypass in order to get different decay forms.

I mean "disconnected".

[chemosis]

looks like im about a decade late. I am also wondering about control over the decay and will try omitting d3

[Mark Hammer]

"Off" is simply an infinite resistance.  So consider wiring up a 3-position on-off-on SPDT toggle switch to either connect D3 to C8 directly, connect it to C8 through a higher-value resistor (e.g., 1M) or disconnect it, giving normal decay, somewhat shorter decay, and very quick decay.

I would suggest use of a pot, but I'm guessing that it may need to be >1M, which will be hard to find.  Of course, you could always try a 1M variable resistance between D3 and C8 and see if that provides enough of a resistance range to produce useful decay-time differences.

[PRR]

Let's see this gator:


In true PaIa style: hyper-thrifty design. Which leaves little scope for change. Attack is R17, recovery is R18, but gain ICd3 is R19/R18. You can't change just one factor, everything interacts.