Chaotic / random oscillator circuit?

How about a conventional saw tooth oscillator, BUT with the thresholds set to wobble by another oscillator?

...Jim Thompson

Hmmm, yes something like that perhaps. Probably a triangle rather than a sawtooth.

More googling found this page, which is not entirely unlike the sort of thing that I have in mind:

This guy has used a PIC to control the lamp brightness. For some reason I find that idea less appealing than an analogue circuit....

Phil.

You sould maybe use two natural random number generating methods and use one to control the temporal length and the other the brightness.

e.g., you could amplify some noise and use that as a random number generator(could the number of peaks or use a threshold).

one

One approach I've tried is a 555 clocking a 8 bit serial-in, parallel-out shift register (74LS168?) wired with three XOR gates for pseudorandom counting. I then used some resistors to twiddle the 555's capacitor's charging slope depending on the register's state. Seems pretty quirky, eh? The problem is, it is of course deterministic, which means when there are more outputs in a high state, it runs faster, and the pattern repeats ad nauseum every 255 clock pulses, which isn't a very long cycle (at 1Hz average, that's about 4 minutes huh?).

Speaking of chaos, it might be possible to produce a chaotic result by interlinking three of these devices. Each one is connected to the other two. If the values are right, it will become a chaotic three-body system. (Note that a two-body system is always well defined and easy to solve, while certain three-body problems are stable because they can be expressed as a two-body problem with a minor perturbation. Certain three-body problems result in something extreme happening (in astronomy, one object might be cast into space; in electronics I suppose the result would be saturation), while certain others remain together, and oscillate in a sort of pattern, but the pattern is somewhat unpredictable.)

If you want truely random response, you might try a type T flip-flop clocked by a Geiger tube. Put it near a piece of granite, the uranium content will provide enough radiation for a few pulses per second, give or take. It'll obey Poisson statistics, if I remember my distributions correctly.

Tim

-- Deep Fryer: A very philosophical monk. Website @

Maybe a very long pseudorandom generator with 3 DACS that pick off

Good Luck! Rich

Well, not entirely true. Depends on what you mean by chaotic. Many n-body problems are not all that chaotic, and of course are, in theory, completely deterministic. I tooka computational physics class where we had to model such things and most 3-body models where actually quite stable(one of our goals was to investigate it). It might start off in some chaotic state but usually ended up in a relatively stable state.

I'm not sure if that would be a good method because either you would have to know of an chaotic state that wasn't just unstable. although maybe there is a variation of this method that would work which would be to introduce random errors in the calculations but then you'll need some random number generator in the first place. Actually simple (psuedo)random number generating algorithms are not difficult to encode and would seem random to most people.

Although I guess for a lamp its not that big of an issue ;) Now that I think about it, it isn't going to hurt if the lamp isn't truly random. In fact it might be kinda cool to do something like that where you could put several lamps together in a room and they would learn there position and use them as initial conditions for the N-Body problem. The intensity of the lamp could be the magnitude of the force.

I thought about that too but its not truly random! It is Poisson(which means its not truly random too) but you also have to take into account external radiation. Of course that might be considered a feature rather than a problem. The main problem here is consistency and I would imagine using Johnson noise would probably be the easiest and most consistent way. Although I'm not sure how cool it would be and it might be more fun to have such a think syncing to radiation

How about three slow, unsynchronized triangle waves? You could do that with three dual opamps.

John

means

No, it is completely random, and it means it is completely random.

No matter how random a signal is (as if there could be degrees of it), there must always be an average periodicity associated with it. This is a statistical property of random events.

Zener noise is nice, but I remember setting up an amplifier using it as a source and measuring the "frequency" in the couple hundred kHz range average. As white noise, there is nonzero LF noise, but it's not as large as it could be.

Maybe an RF FET, with big 1/f noise, would be a good source for this.

Tim

-- Deep Fryer: A very philosophical monk. Website @

A pseudorandom sequence generator on a PIC, of course.

I think I found your problem.

-- Many thanks,

Don Lancaster voice phone: (928)428-4073 Synergetics 3860 West First Street Box 809 Thatcher, AZ 85552 rss:

Please visit my GURU's LAIR web site at

Well, I suppose it depends on how you use the term. By truly random I mean something that follows absolutely no ability to describe or analyze its results. Radiation does follow patterns and there is most likely some internal mechanism that makes it absolutely deterministic.

Wiki says

"Hidden variable theories attempt to escape the view that nature contains irreducible randomness: such theories posit that in the processes that appear random, unobservable (hidden) properties with a certain statistical distribution are somehow at work, behind the scenes, determining the outcome in each case."

and thats sorta my view but it also says

"is used to express lack of order, purpose, cause, or predictability in non-scientific parlance. A random process is a repeating process whose outcomes follow no describable deterministic pattern, but follow a probability distribution." "The term randomness is often used in statistics to signify well defined statistical properties, such as lack of bias or correlation. Monte Carlo Methods, which rely on random input, are important techniques of computational science.[1] Random selection is an official method to resolve tied elections in some jurisdictions[2], and is even an ancient method of divination, as in tarot, the I Ching, and bibliomancy."

Which I agree with too but I think truly random != random. i.e., if your going to say truly random but mean the above then you should just say random.

Lets not argue over semantics. I do not believe "truly random" is a technical term and personally I do not believe anything is truly random... but obviously there are many things that are random.

I agree... but random != truly random in my book. If it did I would not add meaningless qualifiers. Of course you might have a different interreptation of truly random than I do and thats fine. Just trying to clear up how I interpret it. (its kinda pointless since it's no big deal though)

I'm not sure? What is the most simplest method to generate noise and what type of noise is that? I would imagine that any type of noise would work but they would just have different characteristics.

Maybe something I might try is

V |

• --|\ | | \ R | >

| | /

- --|/ | GND

(with feedback of course)

and then use some peak or threshold detector afterwards going into a counter(maybe add a filter or whatever else). I've not done this before I so I have no clue if it would actually work like I imagine.

Ha! Yes, Don. The truth is that I spend weekdays doing digital things, and using a PIC for this would just be boring. Finding quirky analogue circuits is a more interesting passtime for the weekends.

Phil.

How but using a tube then? lol...

How about a recording of some "random" intensity information? Maybe take a candle, a photocell and digitally sample. You could just play it back with a PIC and I suspect few people could figure out that it was looping.

I also did something like the RGB lamp thingy using a PIC, it looks pretty cool doing smooth rainbow color changes(even if I do say so myself) ;-)

e.g.http://mysite.mweb.co.za/residents/cyb00746/chaos/chaos.htm. But it

Here is my stream of consciousness idea. Crystals make low jitter oscillators. Relaxation oscillators jitter. Could you have a PLL attempt to lock to the xtal oscillator, then somehow derive the phase difference, which should be jittery.

e.g.http://mysite.mweb.co.za/residents/cyb00746/chaos/chaos.htm. But it

In about 1970, we had a problem to generate white noise for an audio recording. We selected a TV channel that was all snow, then modulated that volume to sim the engine thrust, worked perfect., though that was at audio frequency.

I appreciate Mr. Endecott's challenge = randomize at the Hz level. Got me thinking! Ken

You mean like:

Thanks for the link; it's a flashing RGB LED.

Have you actually seen one of these? The impression that I got from other descriptions was that they cycled through the colours, fading from one to the next, with a deterministic pattern. But that particular description suggests that it may actually be (pseudo-)random.

I also fear that they won't look great if you put them in parallel....

Phil.