Ringmodulator based on AD633

[kothoma]

Inspired by this (thanks to Cortexturizer)

Something like this

Some classic circuits tweaked to perfection. The pedal is insanely pricey, but I get it. There's a lot going on in there. If I had enough time maybe just maaaybe I would attempt a build like this but I doubt I could pull it off. Not saying this is the most complicated build out there, but it would be an ultimate pedal for me, hells one could start a band around this pedal alone hahah.

I remembered the fun you can have with ringmodulators on synthesizers.
I also have a far memory of John Scofield using one on several occasions (probably from a Line 6 MM4 Modulation Modeler).

So that made me look for an easy solution to analogue multiplication (without using transformers and rectifiers like the old).
Here are some schematics: http://experimentalistsanonymous.com/diy/Schematics/Ring%20Modulators%20and%20Frequency%20Shifters/
Older designs seem to rely on the (obsolete?) 1496 ic. But there's also the AD633:

That chip does all the hard work and you just need to add an oscillator to have a carrier:

(based on http://www.sowa.synth.net/modular/rm.html)

Has anyone here already tried this or a similar design?

Edit: I wonder why the op amps (at pins 1 and 3) are biased with a trimmer. Shouldn't ground be fine?

Edit 2: Of course, there should also be the option to use the input signal also on the second input (instead of the carrier wave) to get a nice octave up effect.

[RobA]

Sounds cool. Aren't the trimmers for fine adjustment so that you can knock out everything but the output signal (remove all traces of the carrier and the input).

Ouch the AD633 is costy at Mouser. The LM13700 spec sheet has an example circuit for doing four-quadrant multiplication too. I haven't tried it, but you could probably get a both the multiplier and the carrier generation out of a single LM13700. It might be worth looking into. I wonder if it's been done in the synth world.

Edit: Didn't find anything on using the LM13700 as a multiplier yet, but here is a synth project using the AD633 and the LM13700 to get a sine carrier signal.

[kothoma]

Aren't the trimmers for fine adjustment so that you can knock out everything but the output signal (remove all traces of the carrier and the input).

Yeah, that's quite possible. Thanks for this hint.

Ouch the AD633 is costy at Mouser.

I will try my luck with ebay, that is, China.

BTW There's one US seller selling a 5 piece lot of the SMD version for $10, but what he asks for shipping to Europe is ridiculous...

The LM13700 spec sheet has an example circuit for doing four-quadrant multiplication too. I haven't tried it, but you could probably get a both the multiplier and the carrier generation out of a single LM13700. It might be worth looking into. I wonder if it's been done in the synth world.

Right, I need to look into that!

[RobA]

Forgot the link in my edit. Here it is anyway,
http://www.musicfromouterspace.com/analogsynth_new/RINGMODULATOROCT2007/RINGMODULATOROCT2007.html

Edit: That one is also using the AD633.

Here are a couple of other projects that do use the LM13700
http://experimentalistsanonymous.com/diy/Schematics/Ring%20Modulators%20and%20Frequency%20Shifters/LM13700%20ring%20mod.pdf
http://lushprojects.com/lushone/contour/build/

They might be good starting spots for using the LM13700 as a multiplier.

[kothoma]

Yes, using an OTA looks quite straight forward. But I can't find data up to what frequency this would work.

[RobA]

I've used the LM13700 in a tremolo circuit and run it up to around 400Hz and it works fine. In the Lush project he's running the ring modulator from an audio oscillator as the carrier, so I think it would be OK in that regard.

[kothoma]

Yes, reading up on these links, this OTA idea should really work.
Can you really get a suitable oscillator from half an LM13700 alone? [Edit: you could get a triangle wave, right?]
Or would it be more fun to have two ring modulators going using both OTAs?

[RobA]

I've seen schematics where the LM13700 is used to waveshape an input waveform to get an a good sine wave. I doubt if you could do it with the OTA alone. This link shows an example of shaping a triangle to a sine. http://www.musicfromouterspace.com/analogsynth_new/OLDIESBUTGOODIES/LFO/lfo-3.html

I'm sure you could come up with something else to do with the other half. A simple alternative for getting a sine wave would be to use an XR2206. I played with using these as function generators and they work well. But, I think they are hard to find now too.

Edit: Thinking about it some more, it would be cool to do a two voice version. You could us an ATTiny84 to generate two carrier waves to drive it too.

[kothoma]

Edit: Thinking about it some more, it would be cool to do a two voice version.

That's what I wanted to hear

You could us an ATTiny84 to generate two carrier waves to drive it too.

Using a microcontroller would also allow for many different waveforms.

[RobA]

 . Yeah you can get some pretty good flexibility in waveform output using the microcontroller. You do have to be a bit careful though because the easy ways to get output oscillators require some pretty aggressive post filtering. Still, you can do quite a bit that way.

One other thing about filtering, on my AU ring modulator, I found it really helpful to run the guitar input signal through an LP filter first and then do the ring modulation. If you use a good musical LP, a Moog type say, the results can be very cool. The main point though is to pull down the upper harmonics so that there are fewer cross terms out of the ring modulator.

[aballen]

Anyone verified this schematic?  I'm waiting on my mood ring, though this could work too.

[kothoma]

You do have to be a bit careful though because the easy ways to get output oscillators require some pretty aggressive post filtering. Still, you can do quite a bit that way.

Sure, you need to be careful with aliasing and bit noise.

One other thing about filtering, on my AU ring modulator, I found it really helpful to run the guitar input signal through an LP filter first and then do the ring modulation. If you use a good musical LP, a Moog type say, the results can be very cool. The main point though is to pull down the upper harmonics so that there are fewer cross terms out of the ring modulator.

Interesting idea! It woudn't need to have resonance and maybe the corner frequency could be fixed too? So would two Sallen-Key in a row do?

[billstein]

Aren't the trimmers for fine adjustment so that you can knock out everything but the output signal (remove all traces of the carrier and the input).

Yeah, that's quite possible. Thanks for this hint.

Ouch the AD633 is costy at Mouser.

I will try my luck with ebay, that is, China.

BTW There's one US seller selling a 5 piece lot of the SMD version for $10, but what he asks for shipping to Europe is ridiculous...

The LM13700 spec sheet has an example circuit for doing four-quadrant multiplication too. I haven't tried it, but you could probably get a both the multiplier and the carrier generation out of a single LM13700. It might be worth looking into. I wonder if it's been done in the synth world.

Right, I need to look into that!

Edit: N/M

[RobA]

...
Interesting idea! It woudn't need to have resonance and maybe the corner frequency could be fixed too? So would two Sallen-Key in a row do?

Yeah, I'd think so. I've used various different filters to do it. In the AU version, I use a driven multi-stage ladder-like filter with a non-linearity between the stages, but it isn't needed for just reducing the harmonics. It had the frequency as a parameter, but it was fixed to a value -- it didn't move with the input note. A couple of Sallen-Key filters with the right frequencies would get you a pretty big range of useful input frequency. 

Edit: I should explain that the reason I went with that filter was that if you back off the drive, you cut the upper harmonics and does what I was talking about above. On the other hand, adding a lot of drive is like putting a fuzz box followed by the filter in the circuit. So, it kinda causes it to act more like a filtered square wave going in to the ringmod. In an analog circuit, you could do better using a real fuzz as an option in the signal path.

[culturejam]

A simple alternative for getting a sine wave would be to use an XR2206. I played with using these as function generators and they work well. But, I think they are hard to find now too.

Tayda to the rescue!
http://www.taydaelectronics.com/xr-2206-xr2206cp-xr2206-monolithic-function-generator-ic.html