ad-3208 mod for feedback

[The Landlord is Dead]

I finished building this pedal and I'm pretty happy with it. The only aspect I'd like to change, is the way it feeds back. With the repeats up full it moves into some fairly hi pitched and piercing self-oscillation.

Anyone have any suggestions on how to change the sonic characteristics of the feedback to something more deep and dirty?

[mattv]

That's a can of worms.

Put a low-pass filter in the feedback path. If this degrades the quality of single repeats, you'll want the filter to be adjustable. This could be a separate control, or a dual-gang 'Repeats' pot that increases the filter as you increase the repeats. Or you could have some sort of follower circuit (check the Device articles at hammer.ampage.org for 'pluck follower' maybe) that automagically adjusts the filter as the repeats get louder or more frequent.

[Mark Hammer]

Here's what you do.

You see that 100k resistor in the feedback path that goes to pin 12 on the compandor chip?  Turn that into two 47k resistors in series.  Now wire up two .0047uf caps in series, and tie one end of that pair to the junction between those two resistors.

Get yourself a 3-position, on-off-on SPDT toggle switch,  Connect the common lug of the switch to ground.  Connect one of the outside lugs to the "free" end of the second cap in the pair.  Connect the other outside lug to the  junction between those two caps.

Ta-da!!!  You now have a regeneration treble rolloff switch to tame everything after the first repeat.  In the centre position of the switch, there will be no rolloff.  In one of the side positions, the rolloff will start around 720hz, which is pretty muted and muffled.  In the other switch position, the rolloff will start around 1440hz, which will be "rounded" but not muffled.  If you don't like the sound of those, feel free to experiment with other cap values.  The caps do not HAVE to be equal values.  I just suggested them so you wouldn't have to buy two different ones.

Note the switching from the centre to side positions will produce some popping.  Also note that attenuating the treble by that much will also reduce the perceived loudness of the regen signal, so you may need to turnthe regen up a bit to compensate.

[The Landlord is Dead]

Thanks Mark. Thats exactly the response i was looking for.

I'll give it a shot.

[mattv]

Here's what you do...

maybe my suggestions are too convoluted or not specific enough?

[Mark Hammer]

Your suggestions are great, actually, and clear enough for many.  I try to write for the guy/gal who is lurking and thinks that what you've said is interesting, if only they knew what it translated to in terms of concrete actions.  It's quite possible what I've written is still too exotic for some of them, but anything that extends the community and puts more possibilities in the hands of more people is fine by me.

If there is one thing I've learned from teachng, it is that for every person who asks a question, there are usually many, many more who have a vague sense they need to know more about "that stuff" but can't quite get the question out, and can't quite follow what you told the person who asked the question.

Consider what I write after your posts to be a little like those little video signing insets for the hard of hearing.

The pluck-follower idea is also good, and worth pursuing, but does require a few more things to be useable.  Remember that it sums trigger pulses, so first you have to have a trigger extractor that generates a pulse at the start of each strum.  The pluck follower acts as a kind of lag generator circuit so that each incoming trigger pulse voltage is added on top of the preceding pulse that is still fading away slowly.  In a sense, it works a little like those 4049-based voltage multipliers, where a voltage keeps getting added to itself.

On a related note, one of the things we've danced around here, but never come out and dealt with head-on is the use of envelope signals AS a voltage to be summed.  There are some interesting possibilities for envelope-controlled filters (Tim, get your pen ready.  You can do this.).

Imagine you have a standard Mutron or Dr. Q circuit, or whatever, with a standard half-wave rectifier envelope follower.  That rectifier puts out a slightly rippley, positive-going, signal which sweeps some control element determining the centre-frequency of the filter.  But what if we sent that envelope signal to another stage as well?  That stage would be a pluck-follower, which would sum and lag those brief envelope voltages, yielding a slowly shifting output voltage that could be used for other things.  

For example, the pluck follower could control an LED/LDR combination, with the LDR in parallel with the resistor that sets attack time.  The faster you play, the faster the attack time gets - an adaptive envelope filter.  Alternatively, since the resonance/emphasis of many different ECF's is set by a single fixed or variable resistor, what about placing the LDSR in parallel with that resonance-setting component?  Depending on whether the pluck follower ends up inverting the summed envelope voltages or not, that could result in *reducing* the resonance as you play faster, or increasing it..  Having something that both increased resonance and decreased attack time as you play faster would be terrific, a really emotionally expressive device.

Of course, the modulation-control circuitry ends up being more complex than the filter itself.  But, that's what you get when you want something that responds to your immediate playing style.

[mattv]

Consider what I write after your posts to be a little like those little video signing insets for the hard of hearing.

Which reminds me of the old SNL skit with Garrett Morris, "News for the Hard of Hearing."  

Thanks though. I understand what you're saying. I have so much more time to think about this stuff than to actually do anything, that I wonder if my ideas go a little too far out. Rarely do I have time to try them.

Having something that both increased resonance and decreased attack time as you play faster would be terrific, a really emotionally expressive device.

In my head this sounds fantastic!

[Mark Hammer]

""Which reminds me of the old SNL skit with Garrett Morris, "News for the Hard of Hearing."  

That was the precise reference.

"Thanks though. I understand what you're saying. I have so much more time to think about this stuff than to actually do anything, that I wonder if my ideas go a little too far out. Rarely do I have time to try them."

If I could solder as fast as I think or type, the world would be a different place.

In my head this sounds fantastic!

In mine too, bro', in mine too.

[swt]

mine does the same...i could never use it for gigs. i?ll definetely try that mod...

[george]

Here's what you do.

that's what I love about this guy

[The Landlord is Dead]

O.k. I'm getting ready to do this mod however I'm thinking I want to make this a pot not a switch. That way I have more tonal control and can even play with it in real time dub style. I'm still fairly new to all this so I wouldn't mind some help and confirmation.

First I'm gonna swap out the 100k resistor to a 47k to get more repeats. I've already swapped the 22k resistor to 10k at pin 2 of the tl072 to give more delay level.

So wiring a simple low pass filter: tie the 3 leg of the pot to the 47k resistor. Then tie the wiper to leg 3. Wire the 1 leg to the cap and then cap to ground. Using 100k for the pot and a .02uf or higher for the cap.

This could be a separate control, or a dual-gang 'Repeats' pot that increases the filter as you increase the repeats.

this also sounds interesting. Can anyone elaborate on how to wire this up.

[Mark Hammer]

First I'm gonna swap out the 100k resistor to a 47k to get more repeats. I've already swapped the 22k resistor to 10k at pin 2 of the tl072 to give more delay level.

So wiring a simple low pass filter: tie the 3 leg of the pot to the 47k resistor. Then tie the wiper to leg 3. Wire the 1 leg to the cap and then cap to ground. Using 100k for the pot and a .02uf or higher for the cap.

This could be a separate control, or a dual-gang 'Repeats' pot that increases the filter as you increase the repeats.

this also sounds interesting. Can anyone elaborate on how to wire this up.

The best solution to this that I ever saw was the tracking filter on the MXR and Ross analog delays.  There, a switched resistor filter was used and the same clocked that stepped the BBD also served to set the clock rate (and cutoff frequency as a result) of the lowpass anti-aliasing filters.  As the delay time increased, the filter rolled off more treble.  Of course, this only works when a clock can be used to control things, and in the case of balancing out bandwidth and regeneration, clocks aren't a factor, so another solution is needed.

The goal here is to recognize that artifacts and artifact-related audio grime accumulates the more you recycle the signal.  So, what you want to do is have treble-cut and regeneration-amount controls that track each other.  As regen amount goes up, treble-amount goes down.

Step 1: Get yourself a dual-ganged audio-taper 50k pot.  I don't know how ideal audio taper is for the treble-cut function, but it is necessary for the regen-amount function and in this case regen-amount trumps treble-cut.

Step 2: Wire up the regen control as per usual, using one half of the dual pot.

Step 3: Any RC combination where the resistor is in series with the signal path and the cap goes to ground will act as a lowpass filter.  The cutoff is set by Freq= 1/(2*pi*R*C) and rolls off at 6db/octave.  If you have an RC filter involving a 50k pot, then you start from the end and work backwards.  Assuming a 50k series resistance, and a maximum cut starting at, say, 500hz (pretty muffled, but that's our designated max treble cut), the cap to ground should be between 6n8 (468hz) and 5n6 (568hz).   At minimum resistance, the filter cutoff is well above the passband of the delay signal so no worries there. The second pot should be wired up using the lugs you would normally use for ground and wiper if it was a volume control.  That way as you increase the level of the regen, the series resistance imposed by the second pot increases and the amount of treble cut becomes more severe.

Step 4: Fine tuning.  The amount of treble-cut produced may or may not be appropriate for the regen amount selected or the person using it,  Moreover, the 2nd pot also contributes to the series resistance going to the 571 chip, and as that goes lower, there is, in effect more regeneration.  As a result, I would suggest replacing the 100k resistor with a 56-62k resistor plus the 50k pot section.  That will be more than 100k but recognize that it will only be so when the regen amount is cranked.  The other question is whether the taper of the 2nd pot is okay.  Here, you may want to take a look at the "Secret Life of Pots" document over at GEOFEX to learn a bit about use of parallel resistors to custom-tune pot tapers.  You may be able to fix things using the strategies outlined there.

Ultimately, I think that separate treble-cut and regen-amount pots provide a better solution, but I understand that not every chassis can accommodate an additional pot.  Hopefully what I've outlined here will accomplish what you want and need it to do.  Bear in mind that you can always tinker with the cap value.

[swt]

i've tried that mod and it didn't work. The tone of the repeats is dirtier, but the high pitch oscillation is still there. Any ideas what can be causing this?.

[Mark Hammer]

Does the high pitch oscillation change tone when you adjust the delay time?  If so, what you are hearing is clock bleedthrough.  That becomes a layout/shielding problem rather than a problem with the mod per se.

[swt]

Hi Mark!!. no, turning the delay pot doesn't change it, it's some sort of feedback in the delay signal. The first tap is ok, then it becomes feedbacking, and turning into the high pitch oscillation thing. It's strange to explain it. i?ve made two of them using different layouts, including the ggg one, and both did the same. Some guys suggested that maybe the comp/exp was oscillating, i used ne570/571, and both did the same, it's supposed to be a sa571. What do you think?. Thanks for your attention and replies!

[Mark Hammer]

I'm bummed out from the Flames losing the Stanley Cup this evening so my thinking is clouded, to say the least.  What I *can* confirm is that it should not be the chips referred to.  An NE571 is a 570 is an SA571.  The 570 version supposedly has better noise specs but not like day and night.

[The Landlord is Dead]

Mark I just wanted to say thanks again for all your help. Your explantions are always thorough and extremely helpfull.

My house is up for sale this week so all my projects our on hold while it's showing. Hopefully, by the time I start up again, you and swt will have figured it out :wink:

P.S. I'm bummed about the flames to. That city deserved it more then tampa. That last 6 minutes broke my heart.