ETI Soundbender Improvements

[Oweng4000]

http://www.fenris.talktalk.net/images/Sound%20Bender.pdf
I've made a q&D ring modulator using this schematic. Currently running into a couple of problems with the design that I would love some input on. One of the main problems is that I can't get the internal oscillator to shut up when I'm not playing. My first thought is to make an envelope follower that effectively mutes the output when there's no guitar, but I'm not really sure how to go about this. The other problem is that the blend control is very unresponsive. It appears that there really is no middle ground. Any input would be much appreciated!

[Kipper4]

Would a switch solution work for you?

Because a switch between rv4a and r10 would effectively cut off the lfo signal to output but it might pop on switching ,I don't know.

Rich

[Oweng4000]

Would a switch solution work for you?

Because a switch between rv4a and r10 would effectively cut off the lfo signal to output but it might pop on switching ,I don't know.

Rich

I like the idea but I'm nervous about the pop. I would like it to be virtually noiseless if at all possible. I'm assuming you're talking about using a chip like the 4066?

[Kipper4]

Well it's a quick thing to try unless it's boxed and hard to get to the wire from rv4.

If you have more chips why not try it on the breadboard and see if it pops.

If it does you could try a Spdt switch and have the common go to 741 input
One outside lug to rv4 lug 2 and tot her outside lug to Vb.
It might not pop.

The envelope idea might work if you use a vactrol ldr as part of a voltage divider in the place of rv4.
It's going to up the parts count. Your going to need an extra dual op amp (lm358) diodes and charge cap. Use the second half of the 358 as an vactrol led driver.
I hope I'm making myself clear.

[Mark Hammer]

Carrier bleedthrough is a common problem with most ring-modulators.  Does the nulling trimmer you used have enough resolution/precision to nail the minimum bleedtrough point?

Here's what I did with my Anderton RM, from the EPFM project book.  The carrier bleedthrough is annoying but very low amplitude.  Instead of opting for a complex noise-gate solution, I simply stuck a back-to-back diode pair in series with the RM signal going into the mixer stage.  The forward voltage of the diodes sets the threshold the RM signal must exceed in order to pass.  It adds some cross-over distortion when the RM signal exceeds that forward voltage and passes, but hell, it's a ring modulator for crying out loud.  Are you going to notice crossover distortion?  Will it spoil the listening experience?

Now, since that diode pair will subtract some signal amplitude, I would suggest reducing the value of R10 at the mixer stage to compensate, and give the RM signal a little boost.  The value would be commensurate with the amount of signal subtracted by the diodes.  So, if you used Ge or Schottkys, maybe you'd want to use 680k instead, and if you used Si diodes, maybe use 470k or even 390k.  I might also suggest strapping a small value cap in parallel with R11, like 22pf.  That should keep the klang but lose the hash.

[Oweng4000]

Carrier bleedthrough is a common problem with most ring-modulators.  Does the nulling trimmer you used have enough resolution/precision to nail the minimum bleedtrough point?

Here's what I did with my Anderton RM, from the EPFM project book.  The carrier bleedthrough is annoying but very low amplitude.  Instead of opting for a complex noise-gate solution, I simply stuck a back-to-back diode pair in series with the RM signal going into the mixer stage.  The forward voltage of the diodes sets the threshold the RM signal must exceed in order to pass.  It adds some cross-over distortion when the RM signal exceeds that forward voltage and passes, but hell, it's a ring modulator for crying out loud.  Are you going to notice crossover distortion?  Will it spoil the listening experience?

Now, since that diode pair will subtract some signal amplitude, I would suggest reducing the value of R10 at the mixer stage to compensate, and give the RM signal a little boost.  The value would be commensurate with the amount of signal subtracted by the diodes.  So, if you used Ge or Schottkys, maybe you'd want to use 680k instead, and if you used Si diodes, maybe use 470k or even 390k.  I might also suggest strapping a small value cap in parallel with R11, like 22pf.  That should keep the klang but lose the hash.

By back to back do you mean cathode to anode or cathode to cathode? Sounds like a very interesting idea that I'll have to experiment with.

[Mark Hammer]

Cathode to anode.  Essentially exactly what you would normally see going to ground in a distortion unit, only they are inserted in series with the signal such that the signal MUST pass through them to get anywhere else.

On the Soundbender, there are essentially several places where you could insert that diode pair. One is between C6 and RV4a, such that the pot adjusts the level of whatever gets past the diodes.

A second is between the wiper of RV4a and R10.    In this case, the pot could be rolled back such that very little gets past the diodes, whether carrier bleedthrough, or intended audio signal.  That could lead to some interesting effects, depending on the forward voltage of the diodes selected.

A third location might be between pin 6 of IC3 and R12.  This would have the net effect of providing crude gating of the entire RM+clean signal.  Since, as noted easrlier, the diodes will chip away at signal level, it would be a good idea to increase R13 to something like 3k9 or even 4k7, to conserve more of the signal.

I don't see any of these as necessarily superior to the others.  They just present different possibilities and challenges.  I have a Soundbender board stuffed and ready but have never fired it up.  It sits in the "bin-of-shame"; projects that were started and await finishing.  I'll get to it...eventually.

[Rob Strand]

By back to back do you mean cathode to anode or cathode to cathode? Sounds like a very interesting idea that I'll have to experiment with.

Haven't see that schematic for years.  Early 80's ETI!

From Boss HM-2.  See D1, D2, C11, R15b.  (Not D7, D8)
http://4.bp.blogspot.com/-mMD0Cdhzk0E/T-_GwsLouII/AAAAAAAABas/u3jeCR9emZk/s1600/DOD%2BFX59.gif

The diodes need to work against a resistor (R15b).  You don't want DC getting through the diodes.

[anotherjim]

Isn't the ETI circuit meant for line level use?  So you will struggle to balance guitar level signal with the  generator tone. Stick a boost/buffer in front. It's probably better to only feed a fairly clean amp channel with this. Anything high-gain will exaggerate the background leakage.
The louder the input signal can be, then the quieter the ring mod leakage will seem to be. What you currently have, is the opposite.
If your guitar amp has a series FX loop between pre and power amp, that might suit it better as it is.

[Mark Hammer]

Or just increase the gain of IC1, by making R1 a lower value.  Currently, IC1 has a gain of 10x.  Make R1 22k and you have a gain of 45x, which may provide a better S/N ratio.  You can always use RV1 to tame things.  Even better, stick a 22pf cap in parallel with R2 for a slightly smoother sound.

[Oweng4000]

Thank you for suggestions. I'll try it out tomorrow and see if it works better