I picked up one of these chips with my last ED order just for shits and giggles. It looks like an easy build, so I was going to throw this up into a stripboard layout and share it here, but I wanted to ask a few questions first:
-
Is there any reason not to use a simple, resistor based voltage divider rather than the 78L05 which the circuit suggests? What’s the benefit of the voltage regulator here? Is it necessary?
-
I’d like to replace R1 with a potentiometer. What values of C4 would I want to use to create a full-spectrum sweep with a 100K pot? Is that possible?
-
Is there any other low-hanging fruit of cool stuff I can get out of this chip/stuff into this circuit?
Cheers!
Ahhhh. Okay! That explains it!
100 ohm resistor and 100 nF capacitor would be around 16000 Hz, same capacitance with 10000 ohm resistor would be around 160 Hz. I recently made a noise module like that, but it actually doesn’t sound so good. The sweep range is bit too much to have a nice control, and as the noise is filtered, the level of the output lowers quite much. When filtering white noise, there isn’t much energy left when most of the high frequencies are removed.
You might get away with using a resistor based voltage divider and then buffer that with an op-amp follower - but it’s probably bush league.
For your filter you can calculate the cutoff frequency with Fc = 1 / (2 * pi * R * C).
What you’ll want to do is put the 100k pot in series with the 5k6 R1 - that way you will have a range of R = 5k6 → 105k6.
Plug that into your Fc calculations, you’ll get:
1 / (2 * pi * 5.6e3 * 1e-9) = 28KHz
down to:
1 / (2 * pi * 105.6e3 * 1e-9) = 1507Hz
Maybe not all the way down to where you want it, but play with the formulae to figure out what you need.
One other cool thing you could add (and something I’m working on in the background now) is a S&H circuit. so you get random sample and hold outputs.
Cheers
1 Like
I think that the Electric Druid noise chips actually output digital noise, so that the output signal directly from the chip is either 0 or 5 volts. Low-pass filtering this gives white noise in range of hearing. But, if you would apply a sample&hold circuit directly to the chip and then filter (with the “white noise” filter at upper range of human hearing) it, would that give a approximation of low-passed noise without affecting its level?
Invest a couple of $ in a bunch of 78L05s (and 78L09s, if you’re interested in building guitar pedal circuits) from Tayda, and you have your +5 V (and +9 V) supplies sorted for a year or two.
1 Like
According to the datasheet, the pink noise is already analog so should be sampleable without any more filtering needed.
Thanks for all of the replies, all!
Looks like it’s a 4-bit output so unless you add a bit of filtering, you’ll get one of 16 levels. Might be good enough, but if things feel a bit limited you know why.
(That PIC has a 5-bit DAC, but the datasheet says 4 bits so I assume they do the calculations with 4 bit resolution. No, not going to analyze the code today 
)
I have a designed a PCB for this
and will have spares if anyone wants. They should be arriving soonish, JLC dependent.
No panel, I’ll just make a wooden one.
2 Likes
I found this guy’s project, he’s able to make some very cool percussive kicks/stomps and swooshies: Noise & VC Tone Filter Module - Things Made Simple He’s using an EQ chip I’ve never heard of, the TDA1524, but I’d probably want that functionality in a separate module.
2 Likes
This is indeed very dodgy! The typical TL074 can provide at ±15V VCC maximal voltage swings of ±10V into a 2k load (Output voltage swing in the datasheet). This means that the op-amp can provide 5mA (the short-circuit current is higher). Whether that is enough depends on the IC, but I do have my doubts about that. The datasheet does not provide any information, except that one should use a voltage regulator. I would follow that recommendation.
1 Like
Rather than starting a new thread I figured I revive this one 
I’m looking to build an ultra-simple table-top pink and white noise generator with the Noise2 chip, powered by 9v. I want to keep the build as simple as possible, so rather than having a biased TL82 or 72 or similar, I opted for an LM358P (which I had on hand) which would allow me to avoid having to bias anything.
I’ve breadboarded this and it works fine based as far as I can tell (I don’t currently have an oscilloscope so I’m saying it’s “fine” based on my by-ear judgement and some multimeter tests). I’ve also looked through the LM358 datasheet and it seems like it can handle voltages near zero, so there should be no issues there.
Here’s what I came up with for a circuit:
I added two trimmers before the outputs because my intended end user is someone who’s making “noise” so they’d likely be plugging it into pedals and thus they might want something to adjust the output depending on what they’re plugging it into.
I’m not sure if I’ve missed anything in the design, and since I’d like to eventually share this with the wider world, any suggestions for improvements or important things I may have missed would be greatly appreciated.
-Gabe
I don’t think that’s the best way to use the trimmers, connect the output of the opamp to the top pin, the center to your output socket and the bottom pin to ground. Maybe add a 1k resistor in series with the center pin to the output socket.
1 Like
Good call, @craigyb. Will do. 
Yes — what you have just varies the output impedance, the effect of which will depend on what it’s plugged into downstream (perhaps including no effect at all).
@craigyb’s recommendation to replace the pots wired as variable resistors with pots wired as attenuators is better, and may be okay for going into guitar pedals. But you’d still be varying the output impedance, and if the downstream thing is for instance a synth module, that 100k trimmer is going to be too large.
The way I might do it would be to put the attenuators before the op amps. Then you have nice low impedance outputs, in fact maybe too low unless you add a 1k resistor between each op amp and its jack. This wouldn’t be any more complicated really, and would make the circuit more versatile.
(This assumes the Noise2 can drive 100k loads. Probably can I’d think, but if not, you could go with inverting amplifiers, with variable resistors in the feedback.)
1 Like
