Post by thread:[PIC] Twinkling LED's and other randomized Effects

'[PIC] Twinkling LED's and other randomized Effects'

        I have heard of shift registers being used along with feedback

to generate a pseudo-random bit pattern.  A PIC would be excellent for

setting up such a scheme because any two or three bytes can be shifted

left or right and ganged in to a shift register so that a bit at one

end marches through the register and ends up at the other end.  How

does the feedback loop work to scramble the bits so that the output at

least appears random?

        If the bits in the shift register start out as all high or all

low, will that prevent the entropy from starting?  In the project I am

thinking of, even the apparence of randomness is good enough.

Martin McCormick WB5AGZ  Stillwater, OK 

OSU Information Technology Division Network Operations Group

_______________________________________________

http://www.piclist.com

View/change your membership options at

http://mailman.mit.edu/mailman/listinfo/piclist

At 09:54 PM 10/11/2004, Martin McCormick wrote:

>         I have heard of shift registers being used along with feedback

>to generate a pseudo-random bit pattern.  A PIC would be excellent for

>setting up such a scheme because any two or three bytes can be shifted

>left or right and ganged in to a shift register so that a bit at one

>end marches through the register and ends up at the other end.  How

>does the feedback loop work to scramble the bits so that the output at

>least appears random?

Someone mentioned my AVR code that does this, easily adaptable to the PIC.

>         If the bits in the shift register start out as all high or all

>low, will that prevent the entropy from starting?  In the project I am

>thinking of, even the apparence of randomness is good enough.

The only thing you can't do with a maximal length generator, is to start it 

with all zeroes. That's a "stuck" state.  Anything else you load in will be 

fine.

That means that an N bit register can only generate (2^N)-1 conditions, but 

you can always make the PN gen longer, and then take some bits out of the 

middle, and those will generate all 256 possible states with equal 

distribution. 

_______________________________________________

http://www.piclist.com

View/change your membership options at

http://mailman.mit.edu/mailman/listinfo/piclist

>How does the feedback loop work to scramble the bits so that the output at

> least appears random?

The feedback loop is basically an XOR of 2 or more bits in the shift

register.  The output of the XOR is connected to the start of the shift

register and will generate a pattern (depending on how many bits are

used and where the taps are placed).

Take a look at the following app note.  "Random" is relative; it depends

on the number of bits in the shift register before the pattern repeats.

http://www.maxim-ic.com/appnotes.cfm/appnote_number/1743/ln/en

> If the bits in the shift register start out as all high or all

> low, will that prevent the entropy from starting?

Starting all ones is ok, never preload with a zero, as the XOR will

never generate a 1 output to feed back into the shift register.

Also google on (random "shift register" "511 pattern") for other articles.

Ken

spam_OUTklumiaTakeThisOuTadelphia.net

{Original Message removed}

>

>Take a look at the following app note.  "Random" is relative; it depends

>on the number of bits in the shift register before the pattern repeats.

It's not random at all.

It's completely and absolutely predictable.

Radioactive decay is random.

Although it's gaussian in large groups, the time between decays is totally 

unpredictable.

OTOH, if I know the register configuration (tap points) and what state it 

is now, or what state it was at any time previous, then I can predict the 

next state exactly.

The interesting thing is that if you only look at a small subset of the 

bits, then it appears random. If the register is long enough, and it 

doesn't take that much, it's repeat time can exceed the age of the universe.

This makes it very useful for secure radio links.

_______________________________________________

http://www.piclist.com

View/change your membership options at

http://mailman.mit.edu/mailman/listinfo/piclist

On Oct 11, 2004, at 7:54 PM, Martin McCormick wrote:

> I have heard of shift registers being used along with feedback

> to generate a pseudo-random bit pattern.

there is some C code to do 5 PWM "random" outputs on a 12c675 here:

    http://www.geocities.com/westfw/electronics/flames.c

It's supposed to look like fire.  It worked pretty well, IMO.

On the other hand, I would think stars would be different.  In

particular, fire spends a lot of time down near zero brightness,

whereas stars, if they twinkle at all, tend to have very brief

lessenings of brightness, along the lines of

        brightness = min(128, 255-random(128))  (where 128 is max brightness)

BillW

_______________________________________________

http://www.piclist.com

View/change your membership options at

http://mailman.mit.edu/mailman/listinfo/piclist

> whereas stars, if they twinkle at all, tend to have very brief

> lessenings of brightness, along the lines of

> brightness = min(128, 255-random(128))  (where 128 is max

> brightness)

I've been using these as RNGs (by seeding timers with intervals)

http://home.clear.net.nz/pages/joecolquitt/white_noise.html

Data generated from the one on the left is used in a game of chance

and passes government requirements for randomness (NB Schottky

is now BAT46, +/- supply is from a MAX232 in the circuit)

In a starfield, attention is not likely to be on just one star, unless you

really really don't have a life, and something less than truly random

would do. A cinema in town has a starfield in the roof with what must

be a simple system and pattern of ons/offs - it's not as convincing as

in/out fades would be

_______________________________________________

http://www.piclist.com

View/change your membership options at

http://mailman.mit.edu/mailman/listinfo/piclist

> Radioactive decay is random.

Even that's under fire :-)

New Scientist September 25 2004 has several articles on randomness and how 

much of it isn't or mightn't be. People are attacking the randomness of 

quantum mechanics *apparently* implicit in the outcomes of Bell's various 

experiments.

________________

NB the following do NOT disprove randomness but are in fact a feature of it.

1.    Measure the distance between a randomly chosen pair of stars.

Measure another similarly chosen pair.

Express the distances in whole numbers.

See if the numbers so obtained have any common factors.

Do this for a large number of star pairs.

Take the proportion that don't have a common factor.

Multiply by 6.

Take the square root.

This gives you Pi :-)

Works with distances between "random" bomb craters too ;-)

________________

2.    Take a floor with regularly spaced  lines on it.

Drop a needle etc randomly on the floor.

Note fraction of drops where needle crosses a line.

Proportion is a factor of needle length, line spacing and, of course, Pi.

(That one's more obvious to me).

        Russell McMahon

_______________________________________________

http://www.piclist.com

View/change your membership options at

http://mailman.mit.edu/mailman/listinfo/piclist

> >Take a look at the following app note.  "Random" is relative; it depends

> >on the number of bits in the shift register before the pattern repeats.

>

> It's not random at all.

> It's completely and absolutely predictable.

True.  Hence the word "random" in quotes.  Actually, it is typically known

as a

psuedo-random sequence.

> OTOH, if I know the register configuration (tap points) and what state it

> is now, or what state it was at any time previous, then I can predict the

> next state exactly.

>

Absolutely.  A common usage is to send test patterns (such as the well

known 511 pattern) down a modem link to determine the bit error rate

of the link.  The receiver must know what the result should be in order

to determine how many errors occured.

Ken

.....klumiaKILLspam@spam@adelphia.net

{Original Message removed}
William "Chops" Westfield writes:

>there is some C code to do 5 PWM "random" outputs on a 12c675 here:

>    http://www.geocities.com/westfw/electronics/flames.c

 Thank you very much.  I got it and will have to look it over to see

how it works, but that's the kind of problem I like.  

Martin McCormick WB5AGZ  Stillwater, OK 

OSU Information Technology Division Network Operations Group

_______________________________________________

http://www.piclist.com

View/change your membership options at

http://mailman.mit.edu/mailman/listinfo/piclist