ilovefuzz.com

So I'm trying my hand at some basic CMOS synth stuff and I've run into some basic questions in all the reading I'm doing.

Take the Beavis heterodyne space explorer schem:


1) What is the reason for the 1.5k resistor in series with the 10k/100k pot? I'm assuming to always have a resistance of 1.5 k min there between 1 and 2...but why? And why that value?

2) What is the necessity of the mixing diode or resistor? I tried two oscillators from the 40106 w/o any mixing resistors and the sound barely changed much. And why that value resistor?

3) Why is C5 there at all?

4) In general, what is the basis for deciding on particular pot values in a circuit like this? I get why between 1 and 2 on the 40106 -- i.e., frequency, but why say 100k for volume over 50k?

5) FINALLY, why are there no decoupling caps here?

Phew...that's a lot of questions.

colossus wrote:
1) What is the reason for the 1.5k resistor in series with the 10k/100k pot? I'm assuming to always have a resistance of 1.5 k min there between 1 and 2...but why? And why that value?

2) What is the necessity of the mixing diode or resistor? I tried two oscillators from the 40106 w/o any mixing resistors and the sound barely changed much. And why that value resistor?

3) Why is C5 there at all?

4) In general, what is the basis for deciding on particular pot values in a circuit like this? I get why between 1 and 2 on the 40106 -- i.e., frequency, but why say 100k for volume over 50k?

5) FINALLY, why are there no decoupling caps here?

Phew...that's a lot of questions.

1) You're correct. It keeps a minimum of resistance there, presumably to keep you in the range of human hearing.

2) When you're mixing multiple signals those resistors kind of prevent the electricity from backing up into your pots, and escaping out some other line. It keeps things separated. I've never used the diode method, but it looks like an even better promise of "electricity only going one way."

3) To block the DC voltage from your other gear. Of course, most of your other gear will have input caps that block DC voltage as well, but if you're mixing with a guitar or something you're going to need that.

4) Volume pots have a little to do with the high frequency response (higher value = more treble) and some to do with an intuitive rotation to volume change ratio. I'm pretty much an A100k guy.

5) You mean, like between power and ground? Not needed. I've made a million and a half circuits using these hex inverters with no decoupling caps with no trouble.

ENJOY THE SQAUREWAVES!

colossus wrote:4) In general, what is the basis for deciding on particular pot values in a circuit like this? I get why between 1 and 2 on the 40106 -- i.e., frequency, but why say 100k for volume over 50k?

I would assume, without having ever built this circuit, that the oscillator pot values are chosen to keep the range of sounds within human hearing (as Aen mentioned regarding the "stopper" resistors).

Volume pot values are chosen for a few reasons. One is output impedance (ie - how much the circuit will load whatever is next in line). Another is that the pot forms a high-pass filter with the output cap. And yet another is that the pot acts as a resistor to ground across lugs 1 and 3, so if you use a really low value, you can lose output volume (and of course, that also means you've affected the output impedance and cutoff frequency of the HP filter). It's quite interactive.

It just so happens that 100K tends to work very well for all three conditions in most pedal circuits. The A taper is chosen because it matches up with human perception of volume change (logarithmic).

Awesome. All of this makes sense to me except a couple of things.

So w/ the decoupling caps... a lot of posts at electro-music and muff wiggler say to def use a 100n decoupling cap from 9v to ground. Does that really only matter then when there's multiple chips present? Also, a lot of posts recommend, if I'm understanding it correctly, to toss a 10k/100k pulldown resistor from input to ground on any unused gates. Does that sound right, as sort of a best practice thing?

Power supply filter caps are never a bad idea. Anything with an LFO (or several of them) is going to likely be "tugging" on the power supply a bit, so it's nice to have reserve current in those bigger electros to help smooth out the thump that can sneak into the audio path. And they form a low-pass filter as well, which helps take out ticking from the LFO(s).

For any unused logic gate, you should ground the input.

Thanks! Now to figure out impedance....

Sometimes I wish I went into ee rather than the humanities. I feel like this might be a bit easier to understand. But then I wouldn't be able to write a marxist critique of audio technology, so I guess that's OK?

Impedance is complicated. At least it is to me.

I'm also a humanities guy (English degree), but over the years I've slowly absorbed some of the theory behind electronics.

Yeah impedance, at least in practice, is shutting down my brain. I feel like I get what it is via metaphor, but when it comes down to how it matters in front of me, I get lost.

Can someone tell me if I'm understanding how a 40106 oscillator is functioning correctly?

1) Cap has no charge so 0 is input which then outputs 1 (high).
2) Resistor sees 1 and current flows through, into cap.
3) Cap is filled, reaches limit of trigger, and trigger jumps to 0 (low)
4) Resistor sees 0 at one side and charged cap on other
5) Cap discharges out through resistor and trigger jumps to 1 (high) -- > and so on.

As there's more resistance , or the capacitor can store more, the frequency decreases -- and vice versa.

More or less correct, even if my terms suck? Or am I way off??

http://hackaday.com/tag/logic-noise/

Just putting it out there for beginners that hackaday has a neat series of articles tagged logic noise which have some very basic CMOS synth designs.

colossus wrote:Yeah impedance, at least in practice, is shutting down my brain. I feel like I get what it is via metaphor, but when it comes down to how it matters in front of me, I get lost.

Can someone tell me if I'm understanding how a 40106 oscillator is functioning correctly?

1) Cap has no charge so 0 is input which then outputs 1 (high).
2) Resistor sees 1 and current flows through, into cap.
3) Cap is filled, reaches limit of trigger, and trigger jumps to 0 (low)
4) Resistor sees 0 at one side and charged cap on other
5) Cap discharges out through resistor and trigger jumps to 1 (high) -- > and so on.

As there's more resistance , or the capacitor can store more, the frequency decreases -- and vice versa.

More or less correct, even if my terms suck? Or am I way off??

You've got the basic idea down, my choice would be using a 4093, make a bunch of oscillators then you get an extra input, which you can use for gating.