R.G.'s Pseudo-Random LFO

[NPrescott]

So after breadboarding Rick's Wind Machine I thought I'd have a go at R.G.'s suggested pseudo-random LFO to modulate the frequency instead of the pot.
http://geofex.com/Article_Folders/LFOs/psuedorandom.htm

Now I know I'm showing how green I am but I have a few basic questions:
1) Should the mixing resistors Rm be the same value for each oscillator? I assumed so to bring them to equivalent output levels.

2) Where should I begin with Rt and Ct values? I've made several tries with values all over the board but I'm not seeing any modulation at the output, instead I get my voltage halved (approximately). I'm beginning to suspect I've done something wrong for this reason.

I'm sure there is a (relatively) easy explanation regarding frequency and the resistor/capacitor values I'm simply having trouble finding it. I guess it's about time I take a look at the Art of Electronics again. If anyone can suggest any readings specific to the topic I'd be grateful.

Thanks

[R.G.]

So after breadboarding Rick's Wind Machine I thought I'd have a go at R.G.'s suggested pseudo-random LFO to modulate the frequency instead of the pot.
http://geofex.com/Article_Folders/LFOs/psuedorandom.htm

Neat! That's actually what works in the old PAIA surf and wind synth, so it will eventually operate when the kinks get worked out.

1) Should the mixing resistors Rm be the same value for each oscillator? I assumed so to bring them to equivalent output levels.

This is the answer you didn't want to hear. It depends.

Yes, if you want them all at equal levels. No if you want one frequency to predominate. Consider a surf synthesizer. You want a major beat of the wave frequency. You want a much slower beat of the waves coming in sets, and (at least) a minor, faster beat of the variation in waves so it doesn't sound so canned. You'd use different resistors to scale the levels.

2) Where should I begin with Rt and Ct values?

I split this because it's two different questions. The frequency of a Schmitt trigger gate oscillator is F = 1/ (R*C) *(a whole bunch of junk about the power supply, upper and lower thresholds, and other stuff). The "bunch of junk" reflects that the oscillator charges from about 1/3 of the power supply towards 2/3 of the power supply, and then flips direction. The "about" changes with power supply, temperature, phase of the moon, etc. To get it approximately where you want it, breadboard one oscillator with its output driving a 4.7K and LED. That will let you see the oscillator frequency down to about 15-20 cps. Then dink with resistors and caps till you get one, and from there you can get the right values from the rest of your oscillators by scaling that "RC" term. It's the product of R and C that matters. A good place to start would be with RC about 1 second, and picking a cap to start with. For instance, with RC = 1S and C= 0.1uF, the value of R is ... 10 megs. Hmmm... slightly impractical starting point. 1uF/1M is also 1 second. 0.1uF and 1M is 1/10 second.

An RC of 1S would produce a frequency of approximately 3 Hz, guessing about the effect of that "other stuff" term. I really ought to do the math, I guess, but it's actually faster to do it on a breadboard once and scale it from there.

I've made several tries with values all over the board but I'm not seeing any modulation at the output, instead I get my voltage halved (approximately). I'm beginning to suspect I've done something wrong for this reason.

I'm guessing that (1) you have had frequencies way, way above a LOW frequency oscillator and (2) you may not be adding the outputs to get a voltage in the range you need.

This last is a subtlety. For any circuit driven by an LFO, you need to figure out what voltages the driven thing (filter in this case) needs to run through its full range. Then you have to dink with your LFO to provide that range. The CMOS oscillators are particularly easy for this because their outputs scale to whatever voltage range you need within the size of the power supply, and they can have a midpoint set by setting the bottom resistor of the scaling mixer to any point inside the power supply range. This is because the output of the CMOS gate is best modelled by a 2PST switch. The output is switched to either power supply or ground. A series resistor on each of the oscillators going to a single mixing point pulldown resistor both scales the outputs independently and makes their midpoint be whatever you want.

I'm sure there is a (relatively) easy explanation regarding frequency and the resistor/capacitor values I'm simply having trouble finding it. I guess it's about time I take a look at the Art of Electronics again. If anyone can suggest any readings specific to the topic I'd be grateful.

Reading A of E is never a loss, but in this case I used the CMOS Cookbook and Fairchild semiconductor's app notes AN118 and AN140, which are both reprints of the old RCA CMOS app notes, I think.

[chi_boy]

Don't forget to keep us updated!  Pics and samples if you can.

Good luck!

-GP

[chi_boy]

I stumbled onto this page and thought it may give you a place to start for the LFO.

http://members.shaw.ca/roma/train_horn.html

It's still higher frequency than you would want, but its a place to start.  Had any luck yet?

[NPrescott]

So, small insubstantial update;
I've got the oscillators interacting as expected as shown by the suggested LED. However, and this is a big however; I can't seem to manage the VCR using an NPN transistor.

I've looked at R.G.'s suggested 2n5088 VCR and I'm getting nothing. Rm was made to be the suggested 100k and consequently the LED no longer lights up. My DMM still shows a voltage variation so I continued on with the transistor. I've tried small values for R1 and R2, tried it with values being equal and with small and large differences and I simply can't get a varying resistance.

I'm unclear on which points along the transistor are to be used as a resistor - my thinking was after the large electrolytic capacitor marked Requiv and ground but it hasn't yet worked so I've stalled to look up anything I can on transistors as VCRs but it's slow going. Sorry for the delays.

[chi_boy]

CONGRATS!!!

Don't downplay the significance of getting the LFO working. That's a big part of the project.

Personnally, I would move away from the LED at this point. There are a lot of white noise circuts online. Some are transistor based and some are opamp based. A different circuit may be easier to add the LFO.

From what I've tried, Transitors are trickier than opamps. But then again it's all about compromise.

[NPrescott]

So I've got it working to some degree! I never quite managed R.G.'s suggested VCR and I had plans to try a VCF instead when I got to reading up on Tim Escobedo's PWM, which doesn't have much to do with anything except for the EZ LFO he recommends.

It struck me that it looked rather like a simplified version of what I was trying so I gave it a shot. Success!

Well, sort of - I wired the oscillators through a 2n5088, (which reminds me that I should give the 2n3904 Tim E. suggests a try). I took the collector to the first portion of Rick's Wind Machine's frequency section (0.0047 caps) and ran a ground wire from the second portion (opposite ends but I'll get to that). I removed the frequency pot because it was getting in the way, and I later found it was interfering with the operation of the LFO's.

Somewhere along the way however I got mixed up and the LFO's don't control the frequency like I was expecting, and instead of sweeping vowel sounds like the frequency pot gave before I now have a more wind and surf machine the likes of which R.G. mentioned I believe.

What I know is limited but I've found that by wiring the LFO's through different sides of the frequency caps provides for different results, on one side you get the surf-y sounds and little else, and the other way round you get background white noise as well as the surf. Either way I like it! If I had to guess I think I might be in some way or another bypassing the noise generator when I get nothing but the LFO, if anyone knows better please correct me.

I've recored a brief (~1:30) clip for those interested (specifically George  ). I have it wired as mentioned for background white noise as well as the sounds of surf because I prefer it - digital spring reverb added around :34. I apologize for the sound quality as the best I could do was a laptop mic - but being white noise you get a pretty good idea of what it sounds like. I'm not altering anything in the video save to turn on the reverb, there aren't any controls any more!

http://www.aronnelson.com/DIYFiles/up/wind_and_surf.mp3

For the sounds in the clip I've only got two oscillators running with Rt for both oscillators at 1M ohm, Ct is 10 uF and ~12.5 uF with Rm for the first ~600K ohms and the second at 430K ohms.

Thanks to R.G., Rick, George, Steven and everyone for the help, and I'll try my best with any questions but I don't claim to know a lot.

geez that is a long post, feast or famine I suppose.