White or Pink noise generator circuit

[Elijah-Baley]

Hello, weird but not so new request.
I'm searching some noise effect. Something that could remember tape, or noisy and ruined old track, etc... It could be awesome if there's some occasional and ramdom crack or pop, and not just the fffff or the shhhh. But that is ask too much, I guess.

Once I built the Little Angel, a chorus based on PT2399 chip, the chorus was nice, but I never completed the pedal because the noise in the background. Silent when I was stop, noisy while I was playing. Hitting the string harder even the noise was loud.
Now I would like a sort or that kind of effect. Just the noise, not the modulation.

I see there are some white/pink noise generator. I found the Wind machine circuit, but no demo.
What circuit is available? Some common parts or at least easy to find will be better.

Thanks!

[antonis]

[ElectricDruid]

My Hiss, Crackle, and Pop chip was an answer to a similar request on an earlier thread:

https://www.diystompboxes.com/smfforum/index.php?topic=117028

There's also a white/pink noise chip:

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

Of course, both of these are microprocessor-based solutions, and at the moment getting hold of of microprocessors is like trying to tie a zebra's tail feathers to a unicorn's horn.

Still, that won't *always* be the case.

HTH,
Tom

[Elijah-Baley]

The Hiss, Crackle, and Pop sounds interesting. Close to the thing I was imagining.

I hoped in some simple circuit, without microprocessor indeed, like the one posted by antonis, even if it's just static noise. That one in particular needs a 18v charge pump and a regulator to 15v, not so easy, but ok. Then I have to add some blend or switch between the two outputs, maybe.

Does the way to make the noise dynamic? Making the noise come up when I play?

[Rob Strand]

Does the way to make the noise dynamic? Making the noise come up when I play?

Funnily enough you can use a noise gate circuit.  You need a few changes:  The audio which is gated is the noise source and the control (which goes to the rectifier, detector) comes from the guitar signal.   When you play the guitar is opens the gate.    At the output you blend the noise back in with the guitar signal.

[Elijah-Baley]

Yes, it's really funny!
The only one noise gate I built (not for me ) was a Fortin Zuul with the PedalPCB board, and it was really good.
For this project I could use some other simple noise gate. Any suggestion? Then I have to see how to invert the gating effect.

[ElectricDruid]

The Hiss, Crackle, and Pop sounds interesting. Close to the thing I was imagining.

It'd be possible to do a similar thing with purely analog circuits. You'd have basic transistor noise sources, and then convert them into pulses using comparators. Depending on the filtering you use ahead of the comparators, you could alter the rate of the pulse changes.
If those changes-of-state are high-pass filtered they become random static spikes.

There's a million ways to do this. The advantage of the microprocessor is only that you can wrap a lot of circuitry up into just one chip, not that it does anything special here that can't be done any other way.

Tom

[antonis]

You'd have basic transistor noise sources

Zener, also..

[PRR]

white

[Rob Strand]

Roland used a different type of (white) noise generator to the reverse breakdown type circuit antonis posted.

See, Q8 and Q9,
https://manuals.fdiskc.com/flat/Boss%20DR-55%20Service%20Manual.pdf

(check if C34 needs to be larger to be white --> to stay white down to low freq need to increase cap to 1uF or so)

If you dig through the archives anotherjim had a thread on a similar circuit.  At the time I remember (after doing the calculations) the noise was due the transistor shot noise and not the 2.2M resistor. (Different principle to PRR's circuit despite the similarity).

Jim used opamp amplifiers instead of Q9.  I remember he had some problems with oscillations.

Anyway the reason this circuit is a good alternative is it can run off 9V (PRR's circuit will also run off 9V).  You need to make sure the layout is noise free, perhaps even some mild shielding.   The reverse breakdown circuit really needs at least 12V.  At 9V you will struggle to get good noise or any noise without checking a lot of transistors for a magic one.

[anotherjim]

I remember some of the experiments Rob refers to and that Roland/Boss scheme. It was tricky for breadboarding to need high gain to raise component noise because of stray crap picked up in the front end, AC hum in particular.
Now, someone suggested splitting the BJT bias resistor and adding a bypass cap to 0v. It seems counter-productive but I think the idea is to limit the amount of noise at the collector being cancelled by the negative feedback to the base, It still had big hum problems. Bear in mind the DR-55 drum machine it came from was battery power only and in a nice steel case so hum pickup wasn't really a problem for it.

[Rob Strand]

At first it made me wonder if this bias scheme would help fend off a some interference.
Pull C1 of course.


The idea is to reduce the impedance in the base circuit and decouple it more from the supply rail.

If we make R1 smaller, say in the order of 22k to 33k then it tweak R2 for biasing.  Using the Roland
part values, if R2 is already 2M2 then adding R1 isn't going to change the biasing much. 
Maybe choose around the C2 = 4.7uF zone.

The effectiveness in removing supply noise is fairly clear and can be controlled with C2.

I'm not sure just how much improvement it will provide against external noise.  While we are reducing
the base resistor somewhat, the transistor input impedance r_pi is already around 30k.
If we make R1 too small, ie. much lower than r_pi,  it will reduce the wanted noise level.

[antonis]

the transistor input impedance r_pi is already around 30k.

Isn't that impedance exaggereted a bit, Rob..??

Supposing about 800μA Collector quiescent current, r_π of 30k calls for a transistor beta of about 1000..
(923 to be more precise..)

P.S.
You do refer on r_π = β/g_m = β*V_T/I_CQ or am I wrong..??

[Rob Strand]

sn't that impedance exaggereted a bit, Rob..??

Supposing about 800μA Collector quiescent current, rπ of 30k calls for a transistor beta of about 1000..
(923 to be more precise..)

You equations are OK but there's a couple of details that push it down.   (For the Roland circuit) The spec'd transistors have a gain of about 300 and the circuit is running of 6V.  Collectors bias at about 2.9V to 3V  so Ic about 300uA, and then re ~ 87 ohm and rpi ~26k.

[antonis]

(For the Roland circuit)

[Elijah-Baley]

Thanks guys.
I'll read your messages soon trying to get all these information, and then see how I could design a first schematic. This is pretty unfamiliar for me, it's not the usual guitar signal schematic.

[amptramp]

I have a General Radio 1390b noise generator which uses a 6D4 gas triode (thyratron) as its noise source.  It is available with a plug-in pink noise filter which I don't have but the schematic for it is in the service manual, so I could make one if I was truly inclined to use it.  You can find the manual here:

https://manualmachine.com/generalradiocompany/1390b/4650800-service-and-user-manual/

Mine came from an audio service shop that was being closed down.

There are some simple digital feedback pseudorandom noise generators that can be outputted through a DAC to give another source of white noise that would probably be acceptable for audio work.

[ElectricDruid]

There are some simple digital feedback pseudorandom noise generators that can be outputted through a DAC to give another source of white noise that would probably be acceptable for audio work.

You don't even need a DAC! Here's one example, from the illustrious TR909:

https://electricdruid.net/tr-909-noise-generator/

If it's good enough for Roland, it's good enough for me!!

[iainpunk]

i never finished it but i was working on a popcorn noise pedal, that had the possibility to have the noise level follow the amplitude of the wave, sort of, by flipping the phase of the signal every time a ''random'' pop would come along. this was complemented by an offset which basically added a DC square wave with the same intervals of the phase flipping.

it had some trouble, and quite a low headroom.
i might pick it back up after my holliday

cheers

edit: https://www.diystompboxes.com/smfforum/index.php?topic=128523.msg1238188#msg1238188

[anotherjim]

Some of us had some fun using just a PT2399 as a noise loop generator. With enough gain, the self-noise builds up quite satisfyingly.