Pic Based White Noise Injector - For use with Digital Mono Circuit

[Peterae]

I plan on using this sub carrier board to which a digital mono board plugs into to create White Noise drive to one channel.

12F629 PIC microprocessor uses a dual LFSR algorithm which produces a pseudo-random stream of bits at over 90KHz. The bitstream shouldn't repeat for 142 million years.

Not my design but based on the code from here.

http://www.electricdruid.net/index.php?page=projects.noisegen

[poynt99]

Peter,

It would be great if you could make a short high quality recording of the output. I'd be interested in making a few of these if the quality is good compared to the old reversed-biased diode method.

Thanks,
.99

[Peterae]

Hi Darren
No problem, it's only been designed this morning, not had a chance to etch the board yet, but over the next couple of weeks i will work on it, i am only building the digital output, the circuit on that web site has some op amps ect to provide a white noise and pink noise output, i hope to be able to drive the digital mono directly from this and then i can control the pulse width and also add a phase delay relative to the other fet channel.

[Peterae]

I just had a thought, on another page he has code for running 8 of these at the same time in software, this would allow me to incorporate an 8 bit port and drive 1 delay chip directly as a white-noise phase delay stage, this should simulate totally my re triggering problem that is causing the random explosion effect, but without the need for re triggering thus allowing me to produce low voltage explosions.

[BEP]

Peterae,

I'm starting to think you may be dangerous

Can't wait to get on a real PC and look for a Brownian noise generator example!

Giant Stochastic resonance here I come!

[Peterae]

Well ive been called a lot of things but never Dangerous 

[Peterae]

It's also occurred to me that by using a larger port size and faster pic i can even series delay chips together to increase the resolution of the white noise for instance i could drive 24 bits and 3 delay chips each chip being a mixture of 250pS, 1nS and 5nS chips giving random phase shifts from 0-1593.75nS in 250pS steps.

My brain hasn't yet worked out if the bandwidth of the white noise is dependent on the speed of the pic, he uses 4Mhz and says upto 90kHz i can run 5 times this speed with the 429 so i wonder if i would get a bandwidth of 5 times @ 450kHz

[poynt99]

If you can get the bit stream out at a higher baud, then I would expect the BW to increase also.

.99

[Peterae]

Thanks Darren
Thats what i was hoping, which also puts me in a position where i could put a couple of dip switches in and soft set the cpu speed on power up from software and then have a variable BW.

[BEP]

You are aware the MIPS rating should be considered along with the clock?

An external clock (crystal) may limit running speed to multiples or divisors of the external clock.

I take it you will use the coil as a mixer of pulse and noise?

Non-linearity?

[Peterae]

Indeed WW
Good point i have not spent much time on this yet but agree that the t cycles Will be i think a quarter without looking it up, but he has already calculated the performance relative to t cycles and therefore increasing the clock should have a linear increase in execution time.

I hope i will have one constant pulse with the 2nd channel ruining using the white noise generator, not worked the finer details out yet but i may have to try operating this in a burst mode to simulate the re triggering, so to do this i will probably trigger the 1st channel using an Io line and at the same time run the white noise generator then after a certain time stop the generator and then back round, this should work as i can get the explosions with any master frequency.

I am at the moment not putting much time into this as i am concentrating on building controller 5 right now, once i have this running and checked out i will be able to turn my attention to this, it's been quiet handy having this at the back of my mind as the design is constantly changing

[Peterae]

Spectrum shot of 12F629 running using internal oscillator

[poynt99]

Hi Peter.

Could you show one with a 40k BW setting, and a 80k setting?

Thanks.

At the moment, it doesn't look too flat.

.99

[Peterae]

And Here is the Pic running an external 20Mhz crystal.

[Peterae]

i can show 40 and 80khz np, but first do you want a closer zoom of the 20mhz version as i have to change code over inside the chip and am currently setup for 20mhz

[Peterae]

OK zoomed 20mhz version  @ 40khz and 80khz

[Peterae]

OK another 2 still using 20mhz crystal

[Peterae]

I dug out an old 3 channel controller which is just a straight fet driver and fet stage, unfortunately my early tests only left 1 channel working, anyway here's 1 channel running 1 coil of a bifilar pancake coil

Interestingly the magnets i use over the top of the coil feel like they have a washboard effect as you move them over the surface of the coil.

[Peterae]

Is this normal, i have one face of the magnet being pulled to the middle of the pancake coil but when i flip sides of the coil it's the opposite end of the magnet that get's pulled to the middle.

[BEP]

Is this normal, i have one face of the magnet being pulled to the middle of the pancake coil but when i flip sides of the coil it's the opposite end of the magnet that get's pulled to the middle.

Yes.

A single layer planar coil has a radial magnetic polarity. If you have two conductors, one above the other like lamp cord, and connect them in series so you wind up with two windings, in certain configurations the polarity is the same as a solenoid coil.

Just think about applying Lenz's law to this 

[Peterae]

Things should get interesting when i build the 2 sub boards that limit the pulse width and drive out of phase using 2 channels of whitenoise.

Hopefully the bits will arrive Tuesday, and tomorrow i can get the little boards made up.

Also WW i have the code in hex form for the counter board if you need it.

[BEP]

Thanks but not quite ready.

I'm having some fun with a 5U4GB as a dual NRO fed by silicon rectifiers. Sometimes I can't tell if the SS or the tube is doing it. Maybe both?

I've been thinking about feeding dual planar coils. I want the phase control so I may need to break and etch your circuits.
I will be proceeding with 18F4550's controlled by 'puter via USB.

Ever thought that planar coil of yours duplicates the action of some very large speaker voice coils?

Think about one above the other with a gap between them. Feed both but one phase shifted. A single one causes Z rotation of a magnet anywhere it is placed on the coil (feed it with true AC very very low frequency). Place another one above it.

Does that bring to mind the OTPU?
I have a suspicion the rings of the OTPU were actually voice coils (epoxy encapsulated) from very large woofers 

[Peterae]

Np WW i will get round to uploading the file Tuesday so you can download at will.

Some interesting thinking there WW.

[BEP]

Most of my thinking on these things already worked on my bench.

The application of Lenz still applies, of course  :

But only to the I.D. and O.D  .

I'm quite sure the OTPU had four magnets. One above another at left and right each. The left and right were sported opposite polarities from one another. The result of vertical opposing movement of the two voice coils would have been a full sine wave on the top coil. Pulsing the lower ring with 90 deg. fields would have made the energy in the output increase a great deal.

The rule is that it takes amps and turns to make a strong magnetic field. This is not totally correct. It takes high energy. If the potential (at H.V.) is cut off near the top of the wave a strong magnetic field is also created while using short pulses with near zero 'off' time.

Lots of possibilities.

Honey do (NOW!) list is waiting....... Later.