Reliable white noise source?

[Paul Perry (Frostwave)]

The traditional one is the 'open collector' type, but... you need to select transistors, or trim levels. Otherwise, there are the digital feedback methods. Anyone got any alternatives, or new ideas?

[Nasse]

You named all I remeber, and quess the digital pseudo random number generator is more stable. In good old days National Semiconductor made a special chip with which you could easily construct a white noise generator. But unreliable noise generator is painful. I made a "drum machine" with National rhytm generator made for electric organs about 30 years ago. It used that noise chip with some transistor or fet gating to do that snare drum sound (noise mixed with bongo drum) and the circuit went malfunctioning state. I noticed I could get the snare working by keeping a finger on some component, but it was very difficult to play guitar same time with just one hand. I put a wet piece of paper in place of finger and it worked for few days before the snare spring mat died for ever. The chip was not so expensive, because I had very small income but could afford to zap all the chips few times with wrong polarity... :oops: I have seen similar pseudo random generator circuits around, done with few cmos logig chips...

What ideas else? Tune between radio stations perhaps. There was interesting "infra red microphone" circuit snippet in Elektor electronics mag years ago, produced interesting noise from light or infra red radiation. Perhaps you could just wire a microphone if you live near waterfall or at seaside. Maybe there is spiffy PIC circuit around? I think they used mechanical devices in the old days at radio and film studios...

Electret mic inside a shell does not work, I quess. I must look my archives.

[Athin]

try here
http://www.discovercircuits.com/N/noisegen.htm
maybe there is something interesing. baudline (www.baudline.com) for linux has a nice noise generator.

[Tim Escobedo]

Maybe several randomly tuned square wave oscillators unequally summed can give a pseudorandom noise source.

[Maneco]

http://chip.aeug.org/noise508.asm

[R.G.]

Microchips 12C508 PIC, about $1.50 each. Eight pin dip. Program it for the shift register pseudorandom thing, using four of the internal eight bit registers for a 32 bit overall shift register length. The thing will give a pseudorandom sequence that will repeat every 1.5... eons...

Internal clock, so it only needs +5 and ground to give you noise. Do the programming right and you get up to six unrelated noise outputs on six different pins by combining the internal bits in odd ways.

The problem with the old National noise generators was that they repeated every 1 or 2 seconds, and the human ear nails that every time as a repetitive beat.

[R.G.]

I almost forgot - you get 5V peak to peak noise at extremely low impedance out, so you can just pad it down with resistors to whatever you want. It can probably drive a pinking filter directly.

Guess which noise generator goes into the GEOFEX (tm) pedals?

[puretube]

Paul: you`re sure you tried the "open collector" with > 10V ?

[mikeb]

C'mon Paul, sounds like this is your chance to dirty your hands with some uC stuff!!!

Mike

[niftydog]

Don't those DSPIC chips have simple signal generators on board? I know, a bit OTT...

Antennas do it well too!

[Boofhead]

You basically have analog and digital methods.  Analog methods use some form of noisy analog device - eg a junction in breakdown.  Digital methods use pseudo random sequences - eg the feedback types.   There's plenty digital methods, some methods generate sequences with particular properties which are useful for different jobs.  It's not well known that if you filter the output of a simple digital feedback generator the spectra changes shape in a strange way - people have come-up with better wheels which avoids this problem.  Circuit complexity varies with both the methods and the sequence length.  How you implement the digital methods is really a separate issue eg. logic chips, CPLD, microprocessor,  which is independent of the digital method.  I suppose the question is what do you really want to know?

Reading in between the lines I guess you want a stable output.  Digital obviously produces a stable output but the sequence length needs consideration - you might need a low voltage rail.  The output of an analogue noise source can be stablized with an agc ckt - that of course adds to the the circuit complexity.  For most jobs a microprocessor solution can go a long way.

At the end of the day only you can make a judgement of the trade-offs.

[Nasse]

:? There was very spiffy looking circuit snippet in some Christmas or July/August number of elektor magazine few years ago, if I remeber it had more than one transistor generating the noise, was it two for making the noise more random...Was it 2002 or something

Maybe several randomly tuned square wave oscillators unequally summed can give a pseudorandom noise source.

There was just such circuit in some famous and biggie European elektronicks mag, and they claim it sounds like steam horn whistle! If you could make your axe sound like that trough PA, that I would call a big tone for big melody lines...

[Paul Perry (Frostwave)]

Thanks for the leads everyone! I'll let you know what I go with eventually..