Thursday, 7 May 2020

Synthesizer Build part-31: NOISE MODULE.

A very easy to build noise module with 5 different sorts of noise, two of them being 'Grainy' noise with adjustable graininess. 

This is a module I adapted from the MFOS Noise Cornucopia schematic by Roy Wilson. It's turning out to be quite a popular project because I'm getting lots of feedback from people who built it and are really happy with it. Especially the addition of the Grainy Noise.
So, I needed a good noise source in my synth and this one seemed perfect. The original schematic has a random gates section which I didn't need but which you can easily add on if you want it. But I left that out. (I made a separate layout for the Random Gates Generator section which you can find further down the article) I also changed the transistor used to generate the noise and I changed the way the transistor is integrated in the circuit. My way is simpler and generates 200mV worth of noise right at the emitter of the BC547.
This was a one day build for me. I spent the morning adapting the design and making a stripboard layout. Then I built it in the afternoon and by 8pm that same day I had a good functioning noise module built into my synthesizer. The layout I made worked right from the start. No troubleshooting needed.
In the layout below the transistor is shown as a schematic symbol, and not as it's normally shown in the TO-92 package, to make it clear that the collector is not connected. In fact, you need to cut off the collector leg completely to stop it working as an antenna. I've put the pin-out of the BC547 in the layout to make this extra clear. You don't need to choose a BC547 that gives you the best noise results. Any BC547 will work fine. That's an other reason why I changed the transistor arrangement from the original 2N3904 with capacitor. This works much better. Should you experience hum or something, from the power supply, then you can put a 10µF or a 22µF electrolytic cap between the base (ground) and the emitter (+ to emitter).

Here is the verified layout.

(Last revised 16-May-2020: Added grounding wires to output jacks and pinout to noise transistor.)

Print only:

Bill of Materials. Instead of the TL084 and TL082 you can also use the TL074 and TL072 opamps:

Here is the altered schematic, made from the original Noise Cornucopia design:

I changed the 500K trimmer, used to set the amplitude of the noise, for a 200K panel potmeter so you can use it as a level control. In my panel I used a 500K panel potmeter but that really is too high a value. When I turn the potmeter 1/3rd open, the amplitude reaches it's maximum at 10V peak-to-peak and the rest is just maximum volume, so I think it's better to use a 200K potmeter. (However I haven't tested it with a 200K potmeter) If you only have a 100K potmeter you can try changing R5 from 10K to 4K7 to get the gain right, and then it should work with a 100K potmeter, but I haven't tried this either so no guarantees there.
The opamps used here are not critical. The schematic says to use TL074 and TL072 but I used the TL084 and TL082. I think you could even use an LM324 instead of the TL074. 
This module is designed to work on a dual 12 Volt powersupply but it will work equally well on 15 V.

How Grainy Noise works:
The Grainy Noise consists of very short pulses with an amplitude of plus and minus 5V. The Opamps IC2 a and b are set up here as voltage comparators which are being fed on the non-inverting inputs with white noise and on the inverting input with a voltage that can be set with the Graininess Potmeter to between 0V to + or - 8.25V (roughly). Each comparitor has a diode on the output so handles only one part of the voltage phase (either positive or negative). So if the voltage on the negative inputs is very high, the noise will only occasionally go over it and we'll get only a few Grainy Noise pulses. When the opamps are not producing a pulse they are at rest in either full positive or full negative voltage on the output pins but those voltages are being blocked by the diodes. So only the Grainy Noise pulses are being fed to the output. As you lower the voltage, the threshold will become lower and the noise will tip over the boundary more often creating more and more Grainy Noise pulses.

The 'Grainy' noise is a real asset to have. It's very useful because of its harsh sound. It sounds a bit like the noise you get from old TV sets. If you look at the scope image in the video you can see that most pulses from the Grainy Noise go into negative voltage. The more you turn up the Graininess, the more pulses you get that go positive and that's what is used to create random gate pulses. In fact only the positive noise pulses are used in the Random Gates generator and the negative ones aren't used because that diode has been left out. See the original Noise Cornucopia Schematic for that. The Highpass grainy output is a bit low in amplitude. That's not just in my build but other example videos show the same thing. Maybe a different type of capacitor would make it better but to really change it you should put it through an extra opamp and give it some extra gain but I didn't bother with that. I don't think I will be using that output much anyway. If you change the resistor to ground you also change the highpass filter so I don't think that is advisable to make it louder.
Btw, the LowPass noise is not exactly the same as Pink Noise in my opinion. The LP noise has more rumble (bass) in it I think but you may have a different opinion on that. I leave that open :) I think to call it LowPass and HighPass etc is more intuïtive than to assign different colour-names to the noise. 

I've made a separate layout for the random gates section of the MFOS Noise Cornucopia design, for those interested in adding this on.

Here's a new video I made with sound samples of the different types of noise:

And finally some pictures of the finished product. As you can see the finished module is very small. In fact it is only 3 centimeters wide so it won't take up much space in your modular set-up:

Okay, that's number 31 done. A very satisfying build because everything worked right from the get go. Any questions or remarks? Please put them in the comments below or post your questions on the EB Projects Discussion and Help Facebook Group.
See you on the next one.


  1. Hi,

    can I also use the 1N4148 as an alternative to the 1N914? Thanks in advance

    1. Yes, I used 1N4148 diodes too. No problem.

  2. I don't have any TL082, but I have plenty of TL072, which they say is the "low noise" version of the 82. I'm guessing this won't work quite as well with the 72? Your thoughts?

    1. No problem mate. You can use them, it won't make any difference. =)