Multiple 40106 oscillators with common detune control

egasimus · Joined: Feb 11, 2011 Posts: 113 Location: Bulgaria

Probably will end up as another of my ideas which I'll never get around to doing, but "as soon as I get the parts I need" (or at least no sooner than that), I'm going to make a sort of a 40106 box with multiple (7+) oscillator pairs. I would like, though, to have a single detune potentiometer for all of them. Is this even possible? Maybe with voltage sag?

JovianPyx · Posted: Sun Jun 05, 2011 4:46 am Post subject:

That is what voltage controlled pitch (as in VCO) is all about.

I assume that "voltage sag" means to starve the 40106 power supply pin by using a pot. That might work, but I don't know how well. But it really depends on what you are trying to accomplish with the 40106 oscillators. If you are looking for accurate pitch control, "sag" won't work for that. Also, if you "sag" the voltage, the amplitude will also reduce, so as the voltage to the power pin goes lower, so does the output.

If you have a solderless breadboard, you can throw one together and try it.

A 7555 CMOS timer has an FSK or frequency input, it would be easier to do it with such ICs, but they don't come 6 oscillators in a package. I think there is a 7556 which would have 2 timers per IC.

Or you can go all the way to voltage controlled oscillators.

PHOBoS · Last edited by PHOBoS on Sun Jun 05, 2011 4:54 am; edited 1 time in total

JovianPyx · Posted: Sun Jun 05, 2011 4:54 am Post subject:

Or H11Fx ICs which work similarly to vactrols.

-minus- · Joined: Oct 26, 2008 Posts: 787 Audio files: 13

I've been wondering a similar thing. I've been playing around on breadboard with Ken Stone's Chime Simulator circuit. This has three 40106 oscillators per chime modulated together with an envelope on the end result. When tuned properly it sounds quite good. I'd like to be able to change the pitch of the 'bell' somehow... but without going through each oscillator and retuning them all. A single pot to do this would be good. With CV it would be ideal... but how?

EDIT: Two new posts since I began replying! Laughing

JovianPyx · Posted: Sun Jun 05, 2011 5:16 am Post subject:

My opinion - if you want to retain the harmonic relationship between the oscillators, you need real VCOs.

egasimus · Joined: Feb 11, 2011 Posts: 113 Location: Bulgaria

JovianPyx · Posted: Sun Jun 05, 2011 6:57 am Post subject:

a 40106 is a schmitt trigger gate. The frequency of the oscillator depends on the threshhold points which are a proportion of the supply voltage. Since it's a proportion and not an absolute figure, the frequency should not change much if at all. There may be some effect from current starvation using a resistor inline with Vdd, but I doubt it will be much - but you can breadboarded it and see. The thing that will be most affected will be amplitude.

Also, starving current will likely cause oscillators that are set fairly close in pitch to lock together. That is probably undesirable in your context.

-minus- · Joined: Oct 26, 2008 Posts: 787 Audio files: 13

Interesting you mention the shruthi box. I was experimenting with these sorts of sounds a few months back. You might like to look at this schematic for an electronic tanpura:

http://books.google.com.au/books?id=AnUTQ4FT25sC&pg=PA85&lpg=PA85&dq=electronic+tanpura+circuit&source=bl&ots=gfqk0IlTc9&sig=EB7BPhUGFOxYrI0670TymJofChg&hl=en&ei=_Z7ITLq6LYWkvgPJloSWCQ&sa=X&oi=book_result&ct=result&resnum=2&ved=0CBkQ6AEwATgK#v=onepage&q=electronic%20tanpura%20circuit&f=false

I still have it in part on breadboards. I used 555's for the oscillators. It sounded kind of cheap in a way... not like those Radel digital models we see nowadays. There's a nice one here at the bottom of this page:

http://matrixsynth.blogspot.com/2007_11_01_archive.html

I'll be interested to see what you come up with... I haven't quite come back to this yet. I seemed to drift off in other drone directions...

egasimus · Joined: Feb 11, 2011 Posts: 113 Location: Bulgaria

Well, I'm not really into Indian music. It's a field I do not dare enter, for I will become lost. Besides, they have it all figured out, don't they? It's the Western culture that needs lots creative labour to fix the mess it is.

But the other day I was listening to 2/7ths of a local trip-hop band (only bass and vocals, since the remaining 5 members couldn't come) and thought - a drone box would have fit their sound very well. This made me think of a drone box that is readily usable in a more traditional musical context and still a bit unpredictable, and the concept of a light-controlled Shruthi came to mind.

I have some other questions, too. I'd figure them out myself, but I'm gonna let pride take a rest now and just ask away.

Is there a formula to determine the RC values for a 40106 osc for any given frequency, or does it depend on the particular device used? (Probably so. Gonna tune using trimmers, then.)
A 40106 overheats if it has more than 3 of its gates oscillating, right? Well, perhaps a heatsink would rectify that? (Probably so. Gonna put chips in a row, with a strip of aluminum across them, held by screws in between.)
And there was something else, but I forgot, meh.

Cynosure · Posted: Sun Jun 05, 2011 7:40 am Post subject:

I have already done some testing with this. Starving the voltage will only work if all the oscillators are at the same frequency. Otherwise the detuning will effect the different notes differently. (ie. low notes are detuned slightly, and higher notes get detuned a lot).

Instead, you might be able to create a single oscillator and use dividers to create your other notes. Check out this thread I started about creating 5ths:

http://electro-music.com/forum/topic-48037.html

There are some great suggestions there and you should be able to divide things up to create full sounding chords.

JovianPyx · Posted: Sun Jun 05, 2011 7:50 am Post subject:

Cynosure · Posted: Sun Jun 05, 2011 9:37 am Post subject:

JovianPyx · Posted: Sun Jun 05, 2011 10:26 am Post subject:

Ah, ok, that's interesting to know. I haven't a clue why it works. I can guess that having a finite limit to the current available at Vdd probably changes the schmitt threshholds, those calculations assume a supply that has more current than the gate will ever need. And I agree that the resistor, especially after a certain resistance can act as a charge current limiter. I'm quite sure that it's not even close to linear so it may work in a funky manner, like all the action at one end of the pot.

What you say about inconsistent behavior is umm consistent with using a method like current starving...

As for the 5th thing, the way to get true perfect fifths that work at any pitch is to use the counter method. Even VCOs can't do that perfectly across the entire range of hearing. They can get very close, but not perfect.

Cynosure · Posted: Sun Jun 05, 2011 10:47 am Post subject:

JovianPyx · Posted: Sun Jun 05, 2011 11:33 am Post subject:

I think that's at least part of it. It can't charge any faster than that resistor lets it. But I'll bet if you put a 50K pot in series with Vdd, a good portion of the low resistance end does almost nothing, which is why I was saying it won't be linear.

Did you notice the amplitude changing when you change the inline resistance? I'd bet that it does, at least to some degree.

Cynosure · Posted: Sun Jun 05, 2011 12:04 pm Post subject:

diablojoy · Posted: Sun Jun 05, 2011 2:30 pm Post subject:

I did a mod like this quite sometime ago on a WSG for starving the supply
with 10K pot on the supply rail to provide roughly between +5v and +15v
freq control is very haphazard more for fun than anything else
the output amplitude was affected quite a lot as well there is a large differance in volume noticeable when changing the voltage
really screams at 15v and fairly quiet at 5v
edit : yeah not very linear at all as i remember, its been a while since i played with it though.
from memory output cuts out at about 4.5volts where the 40106 will stop
oscillating altogether.

diablojoy · Posted: Sun Jun 05, 2011 2:44 pm Post subject:

ha i just remembered i also tried a vactrol in there [well it was a couple of years ago] which also worked ok as far as this cct goes

JovianPyx · Posted: Sun Jun 05, 2011 2:52 pm Post subject:

Thanks diablojoy, that all makes perfect sense.

I think the only ways to get control of the 6 oscillators in a 40106 without the aforementioned side effects are the Vactrols that PHOBoS suggested or the H11Fx IC that I suggested. Of course, if those side effects are desired, then fire up the soldering iron.

As for other ways to make a drone machine like that with one-pot-controls-all one could use the oscillator in the 4046 which would provide a nice wide range. Of course, that's only one VCO per IC. 7555 timers can also be used (or 7556 for dual), but with less pitch range available than the 4046. And the 7555/7556 freq input also causes a decrease in amplitude for higher pitches. There's probably others that I can't think of at the moment, but those two are both CMOS for those of us that like CMOS. I think the 4046 might track better than the 7555/7556 method for what it's worth.

egasimus · Joined: Feb 11, 2011 Posts: 113 Location: Bulgaria

Great info in this thread Smile

Is there an IC out of which you can build more than one divider by 2? Adding a sub-octave would fatten the sound substantially, which might be desirable, but I'm reluctant to do it if I have to use a chip per voice.

And what it is that stretches the page in Opera? Very Happy

JovianPyx · Posted: Mon Jun 06, 2011 6:03 am Post subject:

Yes, and assuming you want CMOS (I would), there is at least 4013, a dual D type flip flop. Connect the Q-bar output to the D input and when you clock it, the Q output will give 1/2 the frequency of the clock.

Also there is the 4027, a dual J-K flip flop which can be set up to do the same thing. Read the datasheet to see how. J-K flip flops are very versatile and can do many things as you will see in the data sheet.

I don't believe there are quad packages for T flip flops, (T = toggle), they are made as I described above from D flip flops.

Edge triggered flip flops like these will probably do unpredictable things if you current starve them. Smile

richardc64 · Posted: Mon Jun 06, 2011 6:09 am Post subject:

blue hell · Posted: Mon Jun 06, 2011 6:29 am Post subject:

egasimus · Joined: Feb 11, 2011 Posts: 113 Location: Bulgaria

Oh.

On the other hand, the (not "a" Laughing

) Bulgarian musician's forum runs IPB 3, which automatically shortens long links. The result? It turns out people copying their own posts (and that's done a lot since you can't bump threads in "Buy/Sell", you have to repost your ad every 7 days) hardly ever think to use "quote". They copy the shortened links, then wonder why they don't work Laughing