Whitey

[Kipper4]

I did some forum searchs and I know some of you have messed with white noise before.

I have on the breadboard a variation of Ricks Wind Noise machine.
Google that all you like the pic definition make it hard to read.
Jellybean roulette here I come.......
I'll post up what's on the breader later tonight.

The goal is to make a wind/wave sound machine that would fit in a 1590B with modulation.


I found it a little unstable on the board and am looking for something more consistant with regard to freqauncy.



I found this

http://schematicsforfree.com/archive/file/Oscillators%20and%20Generators/Pink-White%20&%20Other%20Noise%20Generators/Pink%20&%20White%20Noise%20Generators%20-%20Misc.pdf

My assumption is that more parts means more stable...

Talk to me people.

[Mark Hammer]

Don't forget these ones.  Not all white, but pink and wind noises are OK as well:
http://paia.com/manuals/docs/PTman-doc/sm08forPT/9751is_200.pdf
https://www.paia.com/talk/viewtopic.php?f=6&t=197&sid=244aaa2f9c47b86b4bc45d46323a1cfa
https://www.paia.com/talk/viewtopic.php?f=6&t=177&sid=244aaa2f9c47b86b4bc45d46323a1cfa

[ElectricDruid]

I've had a crack at this too. Ok, I've had a few cracks:

https://electricdruid.net/which-noise-chip-do-i-need/

Plus this, of course:

https://electricdruid.net/adding-vintage-hiss-crackle-and-pop/

There's basically two techniques amongst all those designs you've found: transistor/zener noise sources, and shift-register based noise sources. My firmware noise gens all use the same LFSR algorithm they're implementing in hardware.

One thing to watch out for with LFSRs is that the noise is only flat to about one third/half-ish of the output rate, so if you want flat noise across audio, you need a sample rate of 50-60KHz upwards. The output is a cycle of 2^n-1 pseudo-random bits, so if the shift register is only short and the output rate is fast, you can get audible cycling. The MN5837 used only a 17-bit register, so the cycle was 2^17-1 = 131071 bits long, which is only a little over two seconds. I've never heard one, but it's been described as "Chuffing like a steam train". Some of those hardware shift register designs are much too short to be much good for white noise, but they might be interesting for semi-random pitched sounds.

The only novelty I can bring to this is doing pink noise on a chip, rather than using a filter. This uses a different approach called the Voss-McCartney algorithm, described here with further references on the page:

https://electricdruid.net/noise2-white-pink-noise-source/

[ElectricDruid]

Oh, one other thing to say is that you can get some interesting noise effects by feeding noise to a VCO's frequency CV input (so it becomes an unpitched VCO, essentially) or to the Cutoff CV of a highly resonant VCF. This gives opportunities to explore different noise "characters" using different VCO wave shapes or filter types.

[PRR]

> MN5837 ... ... been described as "Chuffing like a steam train".

A very slow steam engine maybe.

I used one for years, mostly for metering EQ. Yes, it bops up 1dB every 2 seconds. I suppose in a music application that could be monotonous.

[Kipper4]

Starting to think the reason a lot of my trials (about 5 so far) with white noise generators are not working is because I'm only using +9V and I think the transistor breakdown needs more volts to do its thang.
Keep buggering in.

[antonis]

it bops up 1dB every 2 seconds. I suppose in a music application that could be monotonous.

I can assure you Paul that most of my riffs are more monotonus..

[anotherjim]

Starting to think the reason a lot of my trials (about 5 so far) with white noise generators are not working is because I'm only using +9V and I think the transistor breakdown needs more volts to do its thang.
Keep buggering in.

A voltage doubler would be the way to go to use BJT breakdown. Once you use a BJT that way, you should cut the unused leg off at the root since it won't be much good as a transistor  - or so it's said, I've never tried. Still, it suggests that only cheap common types will make more sense -  don't waste any rare mojo parts on it.

[amptramp]

I have  General Radio 1390-B white noise generator that uses a 6D4 gas triode (thyratron) as its noise source.  I didn't get the optional 1390-P2 pink noise filter but the online manual has the schematic for it, so if I ever need it, I can build it.  I bought it from a music store that was selling off everything.

[Kipper4]

Great stuff guys.
I'm liking the noise chip Tom. I'll have to try that.

Thanks @Mark Hammer still got some of those to build. Top info as ever.

@PPR 12 of our English sheckles per chip. I'll leave that while I investigate further.
Not dead in the water just on a back burner. Appreciate the input sir.

I can testify for @Antonis

Now Jim I came across an old thread of yours investigating white noise.
The higher voltages do seem common to the bjt noise source designs.
Oh poo I've been playing transistor roulette mostly with a bag of low gain bc109,s
And many others too and Once trailed they were put back in the bag with the others.
I wished I'd isolated them now then I wouldn't have upwards of 200 jellybeans to check.
At least I have time and patience to get it sorted.
Those metal can bc109s where bought for a npn fuzz design and treble booster iirc.
Just seems odd that each Can had the hfe written on top.
That must have been some 7yrs ago when I started this journey.

Amptramp is that a valve?

I came across also a vellman kit PDF for a pink noise generator but c-mos.

I also note some have chosen to design using the

Tlc2272

Hmm the search continues. I'll probably whip up a perfboard 7660 charge pump for the breadboard.

[Kipper4]

I hate it when that happens. I'll just leave this here for future reference.

https://www.velleman.eu/downloads/0/illustrated/illustrated_assembly_manual_k4301.pdf

Again I note the 12-15V supply.

[amptramp]

@Kipper4

Yes, the 6D4 is a tube and it is mounted in a magnetic field.  Gas tubes are generally good noise generators, so I expect someone to come up with a VR150 / 0D3 noise generator design.  Further information is here:

http://www.ase-museoedelpro.org/Museo_Edelpro/Catalogo/tubes/records_nw/6D4-CV1949/6D4.pdf

[willienillie]

Okay I gotta ask, why would you want to build a white or pink noise maker?  What is the use of such a thing?

[anotherjim]

Hey Rich, that Velleman kit uses CMOS shift registers to make the noise. It ought to work fine at 9v.
It has an 8v regulator. The reason for the onboard regulator isn't clear unless the designer was thinking of keeping the Schmitt trigger master oscillator stable. I don't know about you or anyone else, but for music noise, I couldn't care less about precision. I would leave the 78L08 out and fit a wire link across. Keep the input protection diode.

The CD4006 shift registers are now NOS supply only - get these while you can!

[Kipper4]

Thanks guys

@Willie I would like to develop a low parts count circuit suitable for newbies to make their own pcbs from perfboard and be able to make in a teen class maker environment as a sleep aid for people who suffer from conditions like aspergers and autism.
I wondered if this might encourage suffers to take up a means of earning doing small signal audio or repairs to promote enterprising minds in the market place.
of coursse the schematic and instructions would have to be freely available and creative commons so schools could use it commercially.

Thanks for the info Tramp. Nice one.


Hey Jim
Does it matter if the cd4006 would be the CD4006Be type which im guessing is buffered.?


I'll porbably just use Toms chip and promote the circuit and chip at the same time in the end.
But  meantime i want to understand how it all works and why so i'll likely keep some trials on the breadboard until I settle on a design thats more reliable than what i have now.


Thanks again

[willienillie]

@Willie I would like to develop a low parts count circuit suitable for newbies to make their own pcbs from perfboard and be able to make in a teen class makes environment as a sleep aid for people who suffer from conditions like aspergers and autism.

Well that makes sense.  I was wondering if this was a techno music trend or something.  Those people seem to dig monotony.

[anotherjim]

I'm pretty sure all the CD4006 there ever was will be buffered - and it wouldn't make any difference in this application if they weren't.

[Kipper4]

Phew that's a relief since I bought 10 from China. For a 5

[ElectricDruid]

This looks pretty good for an introduction to the theory behind these LFSR hardware generators:

https://www.eetimes.com/document.asp?doc_id=1274550

There's lots of other links on google, but the trouble is finding one that makes the light come on *for you*.

T.

[PRR]

> ....I'm only using +9V and I think the transistor breakdown needs more volts to do its thang.

B-E breakdown is usually very near 7V. A 9V supply is very marginal for getting current in and getting random voltage out. When the battery sags it is sure to cut-out.

> I can assure you Paul that most of my riffs are more monotonus..

I did say "musical applications"??

> why would you want to build a white or pink noise maker?  What is the use of such a thing?

Random noise into a sample-hold gives a sequence of random values. Run these to a voltage controlled oscillator. You may get a "more melodious melody" than antonis's playing. Or: it avoids the monotony of the Twelve Tone Scale, since it plays all in-between pitches equally often.

Also good for background boops in Sci-Fi movies.

The idea of working with differently-wired people's brains is novel. It may also have use to distract from tinnitus.