So I think that I want to try a Ring Modulator

[Derringer]

although I will probably experiment on the breadboard with the passive method just for kicks ...

right now, Craig Anderton's seems to be a good one to build
http://desmond.imageshack.us/Himg413/scaled.php?server=413&filename=andertonringmodulatorcd7.gif&res=landing

assuming of course that the dual opamp he used is just a dual opamp and I could use a tl072 etc
and assuming that the LM565 chips I see on Ebay are legit? (Are they legit do you think?)
I do see that I will need to build a bipolar supply as well ... shouldn't be a problem.


The ring stringer looks awesome, but I don't want to bite all of that off yet.
Would anyone recommend a different ring mod to build first?

Mostly, I think I'll be using it for vocals.



/"exterminate, EXTERMINATE"

[Mark Hammer]

I've made and happily used both the EPFM1 and EPFM2 ring modulators.  They produce musically useful and interesting sounds.

The chief caveats are that:

a) There is not a lot of control over the modulating internal oscillator, other than frequency.  Often the musical quality of the result is a function of the purity and simplicity of the waveforms used.  RMs came to prominence in use by synthesists who could modulate one sine wave with another.  Guitar signals are FAR more complex than that so the sideband products will be even more complex, especially if you are modulating the guitar signal with a square wave.  I recommend a bit of filtering on the guitar signal to compensate, since there is a limit to the filtering you can apply to the internal oscillator.

Start with a 680pf cap in parallel with the 100k feedback resistor in the first op-amp stage.  That will provide a rolloff around 2.3khz.  Since that, and what we are about to do next, will take away some of the overall amplitude, reduce the input resistor from 100k to around 47-68k to compensate.  That will attenuate the bass but nothing you'd ever notice.

Now, you will see a 33k resistor in the schematic leading to a pair of clipping diodes.  We want to split that resistance up.  I understand that the actual value of the resistance will depend on the circumstances (as the EPFM notes indicate), so you may have to do some extrapolating from this.  But, as an example, let us say you split the 33k up into a 10k and 22k resistor in series.  With the 10k first, and the 22k after it, run a 3900pf cap from their junction to ground.  That will provide a second lowpass filter section, rolling off around 2khz. 

The two filter sections will keep out a lot of the pesky harmonic content from the guitar signal, letting the sideband products be more audible and gong-like, rather than chaotic hash.  Obviously, if you elect to go with a different replacement resistance for the 33k component, you'll need to work out the math yourself.  But aiming for a "split resistance" that allows you to get the optimum signal level and provide some filtering that keeps the fundamental in there and  bleeds off some of the unnecessary treble content, is the overall objective.  If you want to hear the difference, simply use a toggle to lift the ground cap and eliminate the filtering.  You could also use a 3-position toggle to select between two degrees of filtering, and none.

b) The nulling control on the RM is reasonably effective in reducing the background whine, but does not fully eliminate it.  Using a 10-turn trimpot will be helpful in nailing the optimum nulling pot.  When optimal, you won't notice it when playing, but once you stop playing you will.  For that reason, I recommend either using a noise gate after the RM, or using a momentary stompswitch to engage it. 

I have only seen such switches in DPDT form,but quite frankly, you only need the 3rd set of contacts for a status LED.  And if the RM  is only in effect when you are stepping on it, you don't really need that 3rd pole to tell you if it's on, do you?  YMMV, but the effect is somethng that very few players would ever leave on for an extended period, the way they leave a compressor or chorus or reverb on.  As such, a momentary engage switch is not really a big imposition on your playing and use of it.  If anything, it reduces the amount of work you have to do to cancel the effect (just lift your foot).


c)  This is more a recommendation than a caveat or weakness.  The lower the modulating frequency, the more "pitched" the result will sound, since the sideband products won't be as far away from the original note.  Also note that a lower modulating frequency permits the effect to be more consistent.  If you modulate at 80hz, then there will be sum and difference for almost any note you play.  If you modulate at 600hz or 1000hz, there will be little difference product for many notes you play.  That may well be your objective.  I'm just saying that if you simply want something that sounds a little boing-ey, and equivalently so across the spectrum, lowering the modulating frequency is the way to go.

Happy ringing.

[earthtonesaudio]

Semi-related question:
The EHX Ring Thing can do some really pleasant phaser/trem kinds of sounds.  Is that something all ring mods can do or is the Ring Thing doing some other kind of DSP to get those sounds?

[Jazznoise]

They're tricky. I was trying to design my own based around FET's but it was an utter pain. The main issue is with the carrier and how much of the carrier coming through is acceptably. Unfortunately, any noise before the RM leads to carrier bleed of some sort. Mine was eventualy built around a Digipot and a 16f88, which was abit of a pain! I'd wanted an external carrier, but college projects have deadlines 

I saw a balanced FET version someone else made I want to try.

In reply to the Ring Thing the Tremolo sounds are just very slow amplitude modulation, I think the phaser/vibe sound must be doing a very slow,gentle FM as it's a DSP based pedal.

[Ben N]

Great post, Mark. That's one to file away. It may come in handy.

[digi2t]

The ring stringer looks awesome, but I don't want to bite all of that off yet.

Just built one recently, IMHO, worth the effort. There's an updated build paper that addresses some quirks and problems with the original. Do a search in the Gallery, you'll find it in "Dino's stuff".

[Derringer]

Great post, Mark. That's one to file away. It may come in handy.

+1

cool
I just ordered some 565's so I should be able to start breadboarding soon
Do you think I'll need to worry so much about filtering if I plan on using it mainly for vocal effects?

for the bipolar supply, a max1044 -which I have- is probably the easiest route, yeah?

@digi2t, heh, let me get this one under my belt first and then we'll see

[Mark Hammer]

Do you think I'll need to worry so much about filtering if I plan on using it mainly for vocal effects?

for the bipolar supply, a max1044 -which I have- is probably the easiest route, yeah?

I think your instincts are good here.  Note, however, that the back-to-back diode pair intended to keep the maximum signal level under control will still do what they always do...add harmonic content as it clips.  Although the filtering probably won't be necessary (or even audible) when using it with voice, leave it in just to make the unit more versatile.  Can't hurt.

I don't know enough about the 565 or the design to be able to say whether the signal to be modulated needs to be of a fairly constant amplitude, or simply needs to be protected from peaks over some given value.  If the latter, then those diodes may well never conduct, unless you use a fairly hot mic preamp output.

Which brings us to our second consideration.  That first op-amp stage in the stock circuit is unity gain.  With the input resistor change I suggested, that goes up to a gain of around 2x.  That will not be enough for a mere voice mic to provide enough signal to get a decent signal to noise ratio.  And it may not be enough to make the 565 chip do its thing either (again, I don't know enough about 565s to state this categorically).  So, you will either have to use a preamp on the voice mic, or else build in enough gain into the RMs input stage to get reasonable performance.

If you use the suggested 47 input resistor, and use a toggle to straddle/bridge it with a 4k7 resistor, that will give an input resistance of about 4k3.  With a feedback resistance of 100k, that will yield a gain of around 23x, which should be enough to goose the mic signal up to a respectable level to compete with the gutar signal the unit was designed for.  In tandem with the 220nf input cap, that will give a shallow low-end rolloff starting around 170hz, which shouldn't interfere with anyone other than the late Barry White, and he probably won't be using your RM.

Finally, if you are going to stick it in a stompbox with a stompswitch, then you will want to terminate that input resistor with a 1M (or thereabouts) resistor to ground if you want to avoid switch-popping..

[Derringer]

right on

yeah, when I get it on the breadboard, I'll definitely tinker till I find a proper input gain and experiment with the filtering you suggested to see(hear) what's what.

and I see on the schem, the 2.5K null-trimpot is connected to pin3 of the lm565 through an unlabeled resistor.
I'm figuring I could get away with maybe a direct 5k trim to ground or do I need to keep a limiting resistor of some sort in there between pin 3 and ground perhaps?

thank you!

[Mark Hammer]

Not sure why the resistor is unlabelled , but in the book it is listed as 5k6.  A 4k7 fixed resistor and 5k trimpot would let you nail the sweet spot.  Or, if you can score one, go with 5k6 and 2k5.

You could use a 10k trimpot, but then it becomes a question of having one little twitch of the thing overshoot or undershoot the optimal resistance to ground.  Better to go with the smallest possible value trimpot so that fine-tuning is easier.  Or, use a 10-turn trimpot if you have one.

[Taylor]

Semi-related question:
The EHX Ring Thing can do some really pleasant phaser/trem kinds of sounds.  Is that something all ring mods can do or is the Ring Thing doing some other kind of DSP to get those sounds?

In my experience, this is really only possible with a frequency shifter (single sideband modulator). The only analog one I know of is the Bode monster - not a very friendly build. Maybe Puretube knows of some obscure version that's simpler to build?

But there is ready-to-go frequency shifter code for the Spin FV1 on Spin's website. I have tried it and not only does it sound great for atonal stuff, but the subtle, watery phaser-ish sounds are very nice.

[Jordan A.]

Semi-related question:
The EHX Ring Thing can do some really pleasant phaser/trem kinds of sounds.  Is that something all ring mods can do or is the Ring Thing doing some other kind of DSP to get those sounds?

In my experience, this is really only possible with a frequency shifter (single sideband modulator). The only analog one I know of is the Bode monster - not a very friendly build. Maybe Puretube knows of some obscure version that's simpler to build?

But there is ready-to-go frequency shifter code for the Spin FV1 on Spin's website. I have tried it and not only does it sound great for atonal stuff, but the subtle, watery phaser-ish sounds are very nice.

A proper ring modulator can get into some nice subtle "phaserish" sounds at times, with an audio rate signal in one side and an LFO in the other.  Picture the LFO starting at zero, and then rising - the audio signal will start at zero amplitude and then increase in amplitude with the LFO, then decrease in amplitude as the LFO goes back down hill.  The interesting part happens when the LFO goes through the zero point and starts to swing negative, the signal will return to zero amplitude and then increase in size but with its phase reversed as the LFO swings down, which can sound quite different from ordinary tremolo AM.  Dialed in just so it is quite beautiful and shimmery.  I think that might be what earthtonesaudio is talking about.  If that is what he is describing then any four quadrant multiplier with at least one DC coupled input can do it.

On another note, I've just been messing with some nice Analog Devices chips and I'm pretty impressed.  I have an AD534 on a breadboard set up as a ring modulator, and the bleedthrough is pretty fantastic with no trimming whatsoever.  With one input grounded and a 10 volt PP signal in the other I see about 14mV PP of bleedthrough at the output, with the signal at the ungrounded input reduced to 1 volt PP the bleedthrough is buried in the noise floor, which is about 4mV PP, that seems pretty darned good for a messy breadboard and no trimming.

My transformer/diode ring modulator is loaned out to a friend right now so I can't measure it, but I'm sure the bleedthrough and noise were quite a bit worse than that even after careful trimming.

For guitar or vocal use I think bleedthough is one of the most important considerations, nobody wants an annoying and ever present whine along with their audio.

The AD534 is hard to get and expensive, but the AD633 isn't too bad, and would make a super simple yet very high quality ring mod.  The transformer/diode approach has a unique sound that I like a lot as well.  I have not tried an OTA ring mod yet, anyone mess with those?

cheers,

jordan

:edited for clarity, hopefully:

[Earthscum]

A proper ring modulator can get into some nice subtle "phaserish" sounds at times, with an audio rate signal in one side and an LFO in the other.  Picture the LFO starting at zero, and then rising - the audio signal will start at zero amplitude and then increase in amplitude with the LFO, then decrease in amplitude as the LFO goes back down hill.  The interesting part happens when the LFO goes through the zero point and starts to swing negative, the signal will return to zero amplitude and then increase in size but with its phase reversed as the LFO swings down, which can sound quite different from ordinary tremolo AM.  Dialed in just so it is quite beautiful and shimmery.  I think that might be what earthtonesaudio is talking about.  If that is what he is describing then any four quadrant multiplier with at least one DC coupled input can do it.

That's pretty much what this does: http://www.diystompboxes.com/smfforum/index.php?topic=95766.0
Was a "proof of concept" thing, but actually worked well enough I figured I'd post a thread about it. Just a Phase shift oscillator swinging the input bias back and forth on an Octup! with similar results to what you just described. The bias has to keep moving, though, or it just settles back into an octave up.

[Jordan A.]

A proper ring modulator can get into some nice subtle "phaserish" sounds at times, with an audio rate signal in one side and an LFO in the other.  Picture the LFO starting at zero, and then rising - the audio signal will start at zero amplitude and then increase in amplitude with the LFO, then decrease in amplitude as the LFO goes back down hill.  The interesting part happens when the LFO goes through the zero point and starts to swing negative, the signal will return to zero amplitude and then increase in size but with its phase reversed as the LFO swings down, which can sound quite different from ordinary tremolo AM.  Dialed in just so it is quite beautiful and shimmery.  I think that might be what earthtonesaudio is talking about.  If that is what he is describing then any four quadrant multiplier with at least one DC coupled input can do it.

That's pretty much what this does: http://www.diystompboxes.com/smfforum/index.php?topic=95766.0
Was a "proof of concept" thing, but actually worked well enough I figured I'd post a thread about it. Just a Phase shift oscillator swinging the input bias back and forth on an Octup! with similar results to what you just described. The bias has to keep moving, though, or it just settles back into an octave up.

Cool!  Might have to try it sometime,

jordan

[earthtonesaudio]

Cool, nice to hear it may be possible with a simple multiplier.

I was thinking of building something like this: sim

[Derringer]

got the 565'sin the mail the other day and finished breadboarding it tonight

damn is that carrier frequency noisy -- you warned me Mark but I had no idea

I don't have a 10-turn trim on hand so hopefully one of those will help a bit

Otherwise, the momentary switch seems like the way to go.

The tone switch didn't seem to make much of a difference, but then again I only had it run through a Little Gem. Should the tone switch be really noticeable?

sounded very cool when I plugged in a microphone though 

[DiscoVlad]

Would a 4046 be a suitable replacement for the LM565? Just wondering out loud.

[earthtonesaudio]

The pinout is different but I think that the 4046 is the preferred replacement.

[Derringer]

mouser directed me to this chip: NTE989, which appears to have the same pinout as the LM565
http://www.mouser.com/ProductDetail/NTE/NTE989/?qs=0rBfBll4ZW1nO1IdMAKWyA%3d%3d

I'll order one on my next shipment and see if there's any difference


ALSO... I see that GGG has a PCB layout for this effect and uses a 5532 IC instead of the RC4739
http://www.generalguitargadgets.com/diagrams/epfm09.gif
but, there's not parts overlay to see what goes where.
With some time I know I can decipher what goes where

but if anyone has a diagram that they could share, that would definitely be appreciated

[iandy4]

I built one recently that sounded incredible using the classic two transformer (xicon TM018-R's worked best out of the trannys I had on hand) and four diodes in a ring.  If you google ring modulator schematic it should pop up a bunch of times.  All I did was add an opamp gain stage at the guitar input and used a signal generator as the carrier signal (with no gain stage, just worked the volume on the sig. gen.).  It sounds incredibly cool when I used the sweep function and when compared to my roommates way huge ringworm it was definitely sounding as good tone-wise.  I recall having switched the carrier and output signal around which resulted less of the carrier signal bleeding through.