Cd4046 intelligent ring modulator

[Oweng4000]

I'm currently building a ring modulator pedal using a ad633 and I had a crazy idea and I was wondering if it was at all feasible. I would feed a cd4046 my guitar signal and use the PLL to track the signal and then use a switch to get it to hold the note I just played so I could have the ring modulator match the key I was playing in. Anyone have any experience doing something like this?

[EBK]

You just reminded me that I need to learn more about PLLs.   

Welcome to the forum!

I'm still messing with an AD633-based modulator on my breadboard.  Interesting chip.

I like your idea of locking in a frequency using a guitar note. Probable could have lots of uses.

[iainpunk]

Seems like a great idea!!!

Maybe you could also use the PLL to track your signal octave down or up (switch) for even more options in sound.

All the best of luck
Iain

[Kipper4]

Since the CD4046 is voltage controlled I reckon you'd have a hard time matching the vco tone to a guitar note.
I feel sure it's been tried before.
Maybe do a forum search for 4046 ?

I hope you can do it though.

[PRR]

This was discussed within the last month.

"Tracking" is not as easy as we could wish. But there is/was a commercial product that let a '4046 acquire a tone, and then broke the VCO loop so it held (as long as the cap leakage allowed).

[EBK]

Sounds like tracking time isn't an issue, if I understand the goal correctly. 

[reddesert]

The CD4046 PLL is used in the Parasit Studios Guitar Synth octave circuit. Together with a shift register, it is used to generate  a 1 or 2 octave up square wave. You might look at that for ideas. There is sort of a similar concept in the frequency synthesizer section 4.2 of the CD4046 application note, see http://www.ti.com/lit/an/scha002a/scha002a.pdf

[matmosphere]

Neat idea, I'm interested to see how it plays out.

I just built the Parasit U235. It also uses a 4046 to make one or two octaves down but it has a completely different feel than the into the unknown so it might be worth looking at too.

[johnstilton]

Any updates on how the build went?

[StephenGiles]

Here is a very old scan of a sort of guitar synth using a 4046 to track input frequency!

http://www.4shared.com/dir/1321731/a7e7124c/sharing.html

It's the 2 files named CClark_synth

[Mark Hammer]

In some respects, the "ideal" ring modulator tracks the pitch of the carrier and adjusts the frequency of the modulating oscillator to keep the "distance" of the sideband products consistent.

I'll illustrate.  If the modulating source is 100hz, the sideband products of a 440hz input will be 340 and 540hz.  At 1000hz, they will be 900 and 1100hz.  While, mathematically, that sounds sensible, the audible/perceptual result becomes more "pitched" as you go higher up, and less "rubber-band-ey".  If that is the intent, fine.  But if the intent is to maintain the same sort of perceptual event everywhere on the fingerboard, then one needs to keep the same proportional interval of sideband products at all times, not the same arithmetic difference.  And that requires changing the modulating frequency according to the pitch of the carrier.  Admittedly, a trickier task, and one likely better done digitally.

I'm not very familiar with it, but others here might be.  Does the recent EHX Ring Thing do that?

[amptramp]

One problem you have with this is the second and other harmonics may be stronger than the fundamental in which case you will lock onto the harmonic.  As a note dies down, you may have a sudden octave shift in the output.  A simple zero-axis crossing comparator does not work.  You need some analog processing before the 4046 and if you don't want your guitar to sound like a clarinet, some processing to change the square wave output of the 4046 to something else.

[Mark Hammer]

As with octave dividers and octave-up fuzzes, the stiffer the string, the better the tracking, simply because it will be less harmonically rich.  Using the neck pickup and rolling off highs also helps, as does fretting higher up.

[amptramp]

You can make a local maximum / local minimum detector by using a comparator with a phase lag between the inputs.  If a signal is rising, the lagging waveform will be less than the leading waveform.  If a signal is falling, the lagging waveform will be above the leading waveform.  You detect the switchover between rise and fall to detect a local maximum or minimum.  This can be done with an exclusive OR and a clocked register to store the comparator output so successive signals can be reduced to one clock period.  Detect the change from maximum to minimum and vice versa and you get a set of rising and falling signal edges which you can use to trigger the 4046.  Notice that a waveform may go high then go down then go up then go down again in one cycle of the fundamental so you would sync into the second harmonic in that case.  The best way to reduce that is to have a rolloff that reduces the amplitude of harmonics.  This is the analog processing I was mentioning before except that it can also be digital because the digital signal itself has to be reconstituted from a square wave which has only odd harmonics to some sort of sine wave.

[MaxPower]

Going off on a bit of a tangent, if you're interested in digital audio circuits using the CD chips, check out 99IC Projects (magazine). You can download them at americanradiohistory. Link is in the Mark Hammer topic, or Google for it. Just be aware that some circuits appear in multiple issues which wouldn't be a problem except that oftentimes they are incomplete or the parts aren't labelled or the values are missing so best download them all. There aren't that many.

[anotherjim]

Just to warn - the old 1970's 4000 series exploited in those old publications were very different beasts. The 4xxxAE parts were the standard DIP I think. Later they were redesigned into the 4xxxBE and 4xxxUBE types with the UBE being closest to the old AE. As time went by, fewer types were made as UBE and now we only seem to have 4049 and 4069 inverters.
In purely digital use, the current BE versions probably won't act differently, but where some linear behaviour was used, only a UBE part can equal an AE part. The Inverter linear amplifier is still well known today, but back in the past you will also find the 4001AE NOR using each gate input in a linear control which is impossible with a 4001BE. You would have to hunt NOS suppliers for the AE or UB type of 4001.
That doesn't affect CMOS with some deliberately linear function like the 4046. The issues that plague this part are...
Different performance/range between manufacturers compared to purely digital CMOS. Particularly in the linearity of VCO control volts/hz.
Confusing marketing. The 74HC4046, 74HCT4046 "may" be compatible in a 5v circuit, but...
Some have the fairly useless on-board Zener replaced by a 3rd type of phase comparator output.