Squarer

[g3rmanium]

Hi,

I built a working squarer out of six parts. It's not perfect, but it's a start.



For samples and the parts list, see my blog.

Enjoy

[Ben N]

Neat. Schem?

[Meanderthal]

 What's the stainless steel looking thing with the 1/4 in. plug sticking out at an angle? Looks like a modded cattle prod...

[Barcode80]

sounds kick ass! post the schem! i'll add an lpb to the end and hear that monster roar! have you tried chords yet?

[Processaurus]

Sounds pretty interesting, a little like a ring mod that you have tuned perfectly in tune with a note. 

Is the lack of sustain because the dynamics of the guitar notes get squared too (something that came up in other threads about squaring octavias, including designs like JC Maillet's "Harmonic Multiplier", and Geofex's JFET doubler)?  You might try severely compressing the carrier side, so its more of a constant signal to multiply the other, like a regular ring mod.

[g3rmanium]

have you tried chords yet?

Will try to get some more clips up in the evening.

[g3rmanium]

Neat. Schem?

Thanks.

Take +9V through a 470 Ohm resistor to the IA0515S, from there wire + and - of the AD633 accordingly. Wire a 0.1 cap from both supply rails to GND. Send input signal to X1 and Y1 inputs. Wire Z to GND. Take output from W through a cap. Done.

[g3rmanium]

What's the stainless steel looking thing with the 1/4 in. plug sticking out at an angle?

Neutrik plug 

[g3rmanium]

Is the lack of sustain because the dynamics of the guitar notes get squared too (something that came up in other threads about squaring octavias, including designs like JC Maillet's "Harmonic Multiplier", and Geofex's JFET doubler)?

I was thinking this happens because once you get lower than +-1 V, the voltage will decrease rapidly. You know, 0.2 V * 0.2 V = 0.04 V. Maybe adding a constant voltage helps -- one that moves the entire waveform to one axis but is not too high to cause clipping in the signal peaks.

[sfr]

I always wonder about this when I see photos of your breadboard - what's up with the jumper wires with the shrink tubing on the ends?  are you soldering solid core wire onto the end of stranded wire for easier insertion in the breadboard or something?

(sorry to derail things; I'm at work and can't listen to clips.)

[g3rmanium]

I always wonder about this when I see photos of your breadboard - what's up with the jumper wires with the shrink tubing on the ends?

They're called "flexible Steckbrücken" and I buy them from Conrad. 

[Eb7+9]

that's cool Ge, I always wanted to give that AD633 a shot ... it sounds similar to the old-school multiplier circuit for sure ... did you try the second squarer circuit in the app note by any chance ??

one thing to note, the voltage function is input-scaled by a factor of ten in this chip - you get ten volts output at ten volts input ... so the signal has to make it that far before it starts jumping - a nice thing about that is you won't have that problem here, even with an input amplifier of moderate gain ...

that AGC circuit also looks cool ...

the compressor idea is interesting, too much comp and you'll easily loose the sharpness of the octave ... but interesting nonetheless ...

cheers ...

[g3rmanium]

it sounds similar to the old-school circuit for sure ... did you try the second squarer circuit in the app note by any chance ??

I thought about it, yes, but I don't really know what the difference would be. If you have 180 ° or 0 ° of phase shift, you'll get the same output. What do you think?

[Eb7+9]

yeah, not sure either ... maybe it tames the baseline drift at onset of signal, maybe not ...

[g3rmanium]

yeah, not sure either ... maybe it tames the baseline drift at onset of signal, maybe not ...

Here's what I was thinking:

In -> Soft limiting -> Downscaling -> Voltage offset -> X² -> Out

Signal limiting could be done with a JFET, maybe something like the Fetzer Valve.

Next -- scaling the signal down to something like +-1 V.

Apply a bias of 2 V so that we're still within the rail limits even if squared.

[markusw]

What sounds pretty cool at least with the classical 4-diode 2-xfmr ring mod is splitting the signal and
* feeding it once directly into the ring mod
* running it through an opamp wired as comparator (actually just one more part with your bipolar ps) and feeding the "square"-waved guitar signal as a second signal into the ring mod

I could imagine that it should give pretty cool sounds with the AD633 too. Also you would have one signal with a stable amplitude thereby avoiding/reducing the lack of sustain you experience with feeding the non-processed signal into both inputs. Probably you will have to lower the "square"-waved signal you get from the comparator to avoid overloading the AD633 but this just means two more resistors wired as a volume pot (or a trim pot).


Have fun,

Markus

[g3rmanium]

Ok, so I've tried the bias voltage. Which didn't work. It probably just needs more current going in. There's this sentence in the application notes: "The differential X and Y
inputs are converted to differential currents by voltage-to-current converters."

[g3rmanium]

have you tried chords yet?

Yes, sound samples can be found in this new blog entry.

[R.G.]

Welcome to using analog computation circuits.

When you see a circuit like this that has three power pins on the chip - a V+, a V- and a ground - it means that the circuit uses the ground internally for other than a simple bias in most cases. If this is like many Analog Devices computational circuits, you're going to have to give a what it believes is a real bipolar power supply consisting of a V+, a V- and a ground, said ground being able to sink and source current from either power supply without moving. That's the fly in the ointment for chips like this.

Consider trying it with two batteries and if that works, give it a power supply that fakes two batteries. Getting a synthetic ground like we use for opamps to work might be hard to do.

[g3rmanium]

If this is like many Analog Devices computational circuits, you're going to have to give a what it believes is a real bipolar power supply consisting of a V+, a V- and a ground, said ground being able to sink and source current from either power supply without moving.

Isn't that what the XP Power chip does?