EPFM ring mod

[martin beer]

My first post in a very long time...
I'm thinking about a ring mod, and the maestro looks like a big undertaking, so I reckon the Anderton design (1980 edition)might be a good idea.  I was wondering if anyone had tried tweaking this design.  I've found good stuff so far in older posts by Mark Hammer, and this websitehttp://www.lynx.bc.ca/~jc/pedalsBalancedMod.html
In the EPFM book, Anderton suggests that an external carrier cannot be substituted, and I'm trying to figure out why this is.  Has anyone here tried?  JC's diagram shows the output from the VCO section (pin 4 of the 565) joining to the input of the phase detector section (pin 5 of the 565).  Is there an obvious reason why I couldn't just earth pin 4 and send an external carrier to pin 5? 
Or should I just build it and see?  I have a certain amount of faith in the EPFM designs, since I use an envelope controlled super tone control filter just about every time I play electric bass...

[col]

I like the one from Penfold's book "Practical Electronic Musical effects Units" which is based around an LM13700N and classed as a "Metal effects Unit". You can get some very strange sounds from it by using the pots he recommends as trimmers as full sized and altering the settings.

[Mark Hammer]

It may be the case that the requirements for matching the form/impedance/amplitude of the carrier to the input on the 565 are just too complicated to justify adapting it.  Personally, I find enough opportunity for mods in the stock design that I'm not especially motivated to pursue external carriers. 

Incidentally, since you mention using the EPFM envelope follower, consider using it to control the mix of modulated vs unmodulated signal.  For instance, if harder picking reduced the combined parallel input resistance (fixed resistor plus LDR) of the RM signal to the mixer stage, you could reserve ring modulation for when you pick hard.  Think of it as dynamic intermodulation distortion, instead of dynamic harmonic distortion.

[martin beer]

Cheers for the quick reply, Mark.  I guess my main worry about the square wave osc is that it might sound too obviously squarish and stuttery when used at very low frequencies.  In an older post you described extending the range to cover lower frequencies.  Have you tried this with the EPFM ring mod, and if so, how did it sound?  Since I'm mostly interested in the more metallic and odd harmony sounds, this might not actually be too much of a worry.  And yeah, using the envelope follower did cross my mind, although I hadn't thought of that particular application.  It sounds potentially interesting...

[nordine]

martin,

i've been toying with an MC1496 based ring mod, and from what i'e gatheed till now, i concluded that: lefting aside the 'carrier rejection issue', the second MOST important thing on a ring mod is a proper and good carrier wave ....that means, if it can go from square to triangle to pseudo sine, and can alter its symmetry as well... your ring mod can really expand into cool territories

[Mark Hammer]

The thing that one must always remember about ring modulators is that they were developed for use with "pure" electronic signal sources; i.e., oscillators.  All of that sum and difference math presupposes that one knows precisely what the frequency content of the carrier and to-be-modulated signal are, and that unusual and fleeting harmonics and multiple notes are not part of the equation.  Of course that is about as far away as one can imagine from what happens when you plug in a guitar and modulate it with a carrier that contains lots of harmonic content.

The fork in the road that this presents is whether one wishes to approach the RM as a source of dischordant and mildly unpredictable noise, or whether one wishes to approach it as a tonal shaper for what is intended to remain a relatively pitched signal source that goes up by X amount when you want it to, and down by X amount as well.

My own rather limited experience with the two EPFM circuits tells me that as the modulating frequency gets lower, and the filtering of the audio signal source gets stronger, the second type of approach (pitched) gets more feasible.  The lowering of the carrier frequency range is accomplished by increasing C2, the cap that straddles pins 1 and 9 of the 565.  You will note that JC Maillet uses a .1uf cap on his suggested circuit, and Craig Anderton has a .05uf cap in that same position in the stock EPFM version.  It might be worth sticking a .15uf and .1uf cap in series and wiring them up to a 3-position SPDT on-off-on toggle as a range selector.  In the centre position (both caps in series) the combined capacitance is roughly equal to .06uf, while in each of the side positions you would have two noticeably lower carrier ranges.

The other thing to consider is how to achieve an audio signal more closely approximating a steady-state tone source.  There are two aspects to this: the bandwidth of the input, and the dynamics of the input.

You will note that the original has a unity-gain input stage IC1a followed by a back-to-back pair of 1N4001 diodes to set a maximum signal level going into the 565.  There is essentially NO filtering on the way.  Part of the reason for this is that Craig conceived the design as one where a pristine copy of the input would be mixable with a ring-modulated version to retain the "pitched" quality.  One of the downsides of that is that the absence of filtering makes the ring-modulated signal messy, as overtones are modulated by other overtones.  Are there simple ways in which the filtering of the audio signal might be improved to yield a more listenable RM sound?  Happily, yes.  here are some suggestions.

You will note that everything between the audio input and pin 2 of the 565 has this uncanny resemblance to the MXR Distortion+ and DOD 250.  Both of those units have a cap to ground in parallel with the diodes to roll off highs.  So what you can try is to piggyback a cap with the 6k8 resistor to ground (straddle the resistor between the legs of the cap before sodlering them into place on the board).  The precise and desired value of the cap would depend on what value is selected for R12.  You will note that this resistor has an asterisk beside it in the book, denoting that it should be selected on the basis of input signal level.  My guess is that you will probably want a rolloff somewhere around 1.5khz, so if you stick with R12=33k as shown, then you'll want a cap of .0033uf (3n3, 3300pf).

At the same time, IC1a's feedback loop also presents possibilities for rolling off highs in a productive manner.  As noted, Craig provides a pathway for specklessly clean signal to show up at the mixing stage, although there is a send-receive loop along the way that can be used to create sonic mayhem of your choosing.  Personally, I've never used the send/receive loop.  If one is similarly disinclined, then there is nothing wrong with sticking a small cap in the feedback loop of IC1a to roll off the highs of the original signal.  What you mix back in to IC1b won't be clean and crisp, but then clean and crisp is not what you're looking for with a ring-modulator anyways.  If the mixable "clean" signal is essentially the fundamental with a couple lower order harmonics, that's not so bad.  After all part of the purpose of that clean path is to blend in some fundamental to retain the pitched quality.  Consequently, consider sticking a .001uf (1nf, 1000pf) cap in parallel with the 100k feedback resistor to tame the treble even more.

Finally, there is the matter of carrier bleed-through.  Having a multi-turn trimpot for the Null control is particularly useful, but sometimes not enough.  Depending on the carrier frequency used, you may find you want to either a) stick a feedback cap in IC1b to roll off some of the "whine", b) reduce the value of C6 to roll off the lows and reduce the audibility of the drone (if that's where the carrier is), or c) both.  Any of these options will undoubtedly change the tone of what comes out of the output jack, but then what's coming out of the output jack is not exactly a close relative of what went in, either, so I don't see any great sacrifice there.