hi all i m new here. please help me out. i have 5 led's and want to light them such that only one is on at a time rest off. but the sequence shud be random please help me how to do this and please do not suggest pseudo binary random sequences ie the sequences that are cyclic but appear to be randomdue to large repetition time. pleez help me as early as possible and add a word on generalising this to more led's

thanks

 

Originally posted by joseph123@Jan 6 2005, 11:05 AM

hi all i m new here. please help me out. i have 5 led's and want to light them such that only one is on at a time rest off. but the sequence shud be random please help me how to do this and please do not suggest pseudo binary random sequences ie the sequences that are cyclic but appear to be randomdue to large repetition time. pleez help me as early as possible and add a word on generalising this to more led's

thanks

[post=4446]Quoted post[/post]​

Hate to tell you, but there is no truly random number generator.

If you want to get into the math of it, only way to have a true random number generator would for the period of be infinite. No way to get there. The generator would have to run from now until the day the universe ends.

All you can do is make one with a huge period that is cyclic.

p.s. If someone knows of one, I would love to know about it myself.

 

This simpliest way of creating damn nearest truely random generator, I've found is to build a random bit generator using the avalanche noise in a reverse-biased PN junction. Then use the bit signal to clock your LED sequencing circuit.

There's plenty of websites explaining how to create a random bit generator.

 

use very hi fr. oscillator ( e.g. 3 inverters in a ring ) and sample it with a much lower fr. signal. It gives "random" 1s and 0s.

 

Originally posted by nbucska@Jan 21 2005, 12:38 PM
use very hi fr. oscillator ( e.g. 3 inverters in a ring ) and sample it with a much lower fr. signal. It gives "random" 1s and 0s.

[post=4703]Quoted post[/post]​

All your doing is aliasing a signal using a low sampling rate. You will hit a cylic pattern in it eventually.

Cell phones frequency alias a lot. Their GHz signal aren't sampled in the GHz range. That fast of switching would suck way too much juice, so they sample at a much slower rate using the same bandwidth and end up with a frequency aliased signal matching the GHz signal but down in the KHz or MHz range. Its neat stuff.