Soundclips! Schem! AD633 Trem and Ring Mod

[varialbender]

Thanks so much for all your help.
I breadboarded my AD633 Trem/Ring Mod, and it worked!
The oscillator is a modified "Super Simple Dual LFO" from eaced.com.
The AD633 stuff is pretty much straight from the data sheet.
The power supply is thanks to Paul Perry.
Right now, it's got pots for triangle/square, rate, and mix.
I need to up the square pot to 10k and filter the square wave a bit, cause it's popping.
I also need to add circuitry to the X2 and Y2 inputs of the AD633 to null the bleedthrough.
I'm preceding it with a phaser soon, and adding stereo outs and a bit of other circuitry.
This is just the first section of the prototype, but I'm just really happy that my first 'design' ever functioned.
It's only my second ever pedal built, after the electra distortion, so I'm proud.
Below are the soundclips and the schematic for power supply and oscillator+multiplier.
Sorry for the bad sound quality, it's an acoustic guitar into a Dean Markley amp.
Sorry for the large wave file and bad bmp quality.
Enjoy, and thanks again for all your help.

http://media.putfile.com/ProtoTrem01

[Transmogrifox]

That is a really nice clean sounding clip.  There's some pretty incredible potential here by using the simple circuit you have in its current level of design, even without the phaser.  It sounds like you were getting some really neat swells.  Colored the sound just right.

[varialbender]

Thanks a lot Transmogrifox, that means a lot.
I'm considering having it separate from the phaser, but I guess I'll have to add the phaser circuit and see if I prefer them together or apart.
Thanks again, I know you've helped me personally on numerous occasions with various questions. I know you helped me understand phasers, and helped troubleshoot some of my breadboarding problems. Always a great help!

[Transmogrifox]

Thanks a lot Transmogrifox, that means a lot.
I'm considering having it separate from the phaser, but I guess I'll have to add the phaser circuit and see if I prefer them together or apart.
Thanks again, I know you've helped me personally on numerous occasions with various questions. I know you helped me understand phasers, and helped troubleshoot some of my breadboarding problems. Always a great help!

Thanks for sharing the clip and putting time into developing yet another fine DIY project!

[goodrevdoc]

Thats way cool VB. Does the effect go from trem to ring mod just by turnibg the 1M pot?
-justin

[John Lyons]

Was the ringmod stuff in the second half of the clip? I only was able to hear little bits of the initial clip...dial up hear putfile is a little weird...Sheesh. From what I heard the trem sounded nice.
I'd like to hear the Ringmod though specifically.
Nice work

john

[varialbender]

Yeah, the ring mod is later in the clip, and there's only a little bit of it.
I'll try to add a file for just ringmod after my next improvement on the circuit.
I'll try to get the square wave sounding as it should, getting rid of the bleedthrough, adding a tilt control, and add a range switch.
Right now, the last bit of the 1M pot gives ring mod sound. After the range switch is added, there'll be more room for high frequency stuff.
Thanks for the comments.

[Paul Perry (Frostwave)]

Getting the bleedthru out of my ring modulator took two years & lots of tears. But the keys are 1. the trimming of the AD633 and 2. decoupling the oscillator section (in my case by 220 ohm resistors in the power supplies to the osc section & 220uF bypass electros) and 3. layout to avoid capacative pickup of loose oscillator noise.
Nothing worse than getting it 'just right' & then closing the box.... and hearing the bleedthru again...

[varialbender]

Scary. I had no idea bleedthrough was that difficult to deal with.
Any advice on how to decide what's best for decoupling the oscillator?
I don't know much about it
Thanks

[brett]

Hi
RE: decoupling the oscillator section (in my case by 220 ohm resistors in the power supplies to the osc section & 220uF bypass electros)

Agree totally.  Decoupling can be a pain.

For a couple of years now I've been using a series 100 ohm resistor and 470uF cap to ground at the 9V in to ALL my projects.  Inside, I use parallel 10uF and 0.1uF bypass caps near all op-amps and oscillators.  And through painful experience I learned to ALWAYS use TLC555s instead of LM555s.  It's like 555s were designed to throw peaks and dips into audio gear.

[varialbender]

Oops, just went over the schem and found an error.
I really mixed the guitar input at Y1 with the output of the AD633.
I didn't mix in the oscillator.
I'll change the schem on my next revision, which I'm working on right now.
I did some research on decoupling, and it seems alright.

Thanks

[Transmogrifox]

Oops, just went over the schem and found an error.
I really mixed the guitar input at Y1 with the output of the AD633.
I didn't mix in the oscillator.
I'll change the schem on my next revision, which I'm working on right now.
I did some research on decoupling, and it seems alright.

Thanks

A mathematical equivalent is to add a DC offset to the oscillator.  You may also use the Z input for mixing the original guitar signal with the output.

[varialbender]

If you added enough of a DC offset that the bottom of the wave was at 0V, that'd give you the equivalent of what some call amplitude modulation. This is like ring modulation with a half and half blend of wet and dry. In low frequencies, the trem would go from full volume to 0 and back to full. Adding more DC offset would just make the whole thing louder eh? Concidering it's a VCA, anyway... So, I guess up to that point it would function a bit in the same way, but not exactly, and only at voltages low enough under Vcc. Mine is around 11V to -11V, so that wouldn't work. Unless you're saying that adding DC to that would max out on the top part of the wave, and continue on lower parts, nearing a straight line if it's entirely maxed out... though I'm not sure if that's right and if that's what you mean...

As for the Z input, that would only allow for half and half, and if I wanted closer to dry, I'd need to add resistance at the AD633's output.

I understand you're not suggesting that this is the method that will work for me, but rather a method that might work in some situation, and that you're just getting me thinking, so thanks again. After all, the best part of being creative and innovative is doing things differently.

[Transmogrifox]

Unless you're saying that adding DC to that would max out on the top part of the wave, and continue on lower parts, nearing a straight line if it's entirely maxed out... though I'm not sure if that's right and if that's what you mean...

I think you're onto it.  There's one pot you'd need to add:  Carrier amplitude.   If you add some DC, and dial the carrier to amplitude to 0, then you just have a pot-controlled VCA.  As you start to dial in more carrier, it becomes more dominant in its effect on the guitar signal.

Anyway, by controlling the carrier amplitude, you can adjust from 100% clean to 100% modulation.  You could even make this voltage controlled so you could modulate it even more so you could pan from amplitude modulation into the suppressed carrier modulation to 100%, then at inverted phase back to AM on the other side.

I had my PWM-controlled ring mod set up this way.  I really liked it for tremolo.  At full suppressed carrier modulation (like you're doing) the carrier is essentially rectified.  You can put in, say, a tri and add a bit of DC shift to it.  The first peak comes up a little bit, and the second peak comes up much louder so you get this kind of 1-2-1-2-1-2 type sound.  Kinda neat, I think.  It also allows you to do square wave tremolo.  If you put a low frequency square wave into this thing you'll just hear a "click, click click" sound every time it quickly inverts the phase on your input signal.  As you begin to offset the square wave, you essentially are adjusting its depth.

As you have yours configured right now, the triangle wave sounds twice as fast for tremolo because your ear can't discern the difference between the inverted and noninverted guitar. 

In the end, all I'm saying is this:
Datasheet says the output, Y = (x1-x2)*(y1-y2)/10

=(x1y1 - x1y2 -x2y1 + x2y2)/10

x1=guitar
y1=carrier
x2=0
y2=constant (k)

= [(guitar * carrier) - (guitar*constant) + 0 + 0 ]/10

Let's make the constant negative (less than 0), so we have

[(guitar * carrier) + (guitar*constant)]/10

When you mix on the output, you're doing this:
x2=0
y2=0
so the output is this
(guitar * carrier) /10

When you mix it at the output, you give the dry guitar signal a certain gain, "constant":
guitar* constant,

and you add it to the output:

[(guitar * carrier) + (guitar*constant)]/10

No difference, EXCEPT...
how this "constant" is controlled.

In either case, it all comes down to what you want to do.  The Z input is a convenient place to add the clean guitar because the mixer is integrated with the chip and the output is buffered...though again you couldn't mix to full-clean unless you were to also adjust the carrier input amplitude.

There are doubtlessly multiple ways to skin this cat and it all depends on how many knobs you want and/or what you want them to control.  Mixing the output with the input separately from the IC could make it a 2-knobber : 1 for frequency and 1 for mix.

Adding a 3rd for DC offset is something I would do.  I need to order some of these AD633's.  I never did get as nice of results as you have with the AD633 from my PWM experiment (though it was interesting to play with it).

[varialbender]

Wow, fox, you never fail to go in depth while still being coherent. Great help, thanks.
It's crazy how many different ways there are to do things. I've thought about a few of them, but it's great to hear new takes on things all the time. For example, I would have automatically put the DC shift before the chip, or maybe even some magic with the oscillator opamps, but hadn't thought to use the Y2 input.
I'm definitely a fan of the sound of two different amplitude peaks on tremolo. Anything to keep the wave sounding interesting. As I'm planning now, that part of the mix knob will keep the oscillator sounding interesting, as well as square/triangle, tilt, and rate knobs (and rate jack, hopefully for both CV and expression pedal).
When I first plugged it in, I had the wave set to square and mix to full wet, and I definitely experienced the pops and clicks, nice and loud.
So far, I'd definitely recommend getting the AD633. I've heard nothing but great things about this chip, as I'm sure you've heard, and my extremely limited experience with them already has me smiling, so it can only get better when I fix the few issues I havn't worked on yet.

Thanks again

[Jaicen_solo]

The ad633 is a brilliant chip!
I'm designing a ring mod based on it at the moment as well. There are a lot of synth DIY projects that use this chip so have a look around (ie, at CGS synth).
It really does sound good when you feed it a clean high frequency sine into the carrier. Getting it to cancel is less easy if your carrier has lots of harmonics.

[tcobretti]

Any more progress on this circuit?  Is this the complete circuit?  Anybody built this?

[tcobretti]

Also, in an AC circuit, if something goes to gnd what do you connect it to?