Thursday 7 May 2020

Synthesizer Build part-31: NOISE MODULE with 5 TYPES OF NOISE + Random Gates.

A very easy to build noise module with 5 different sorts of noise, two of them being 'Grainy' noise with adjustable graininess.  Works on dual 12V so Eurorack friendly.

(The Random Gates section is located half way down the article)
This is a module I adapted from the MFOS Noise Cornucopia schematic by Ray 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. The transistor's Emitter-Base breakdown voltage is exceeded thus the transistor is operating in avalanche mode, creating nothing but pure noise.
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 might need to choose a BC547 that gives you the best noise results. I heared through feedback comments that there can be differences between transistors but you should get noise with any transistor. It's just that some transistors produce more noise then others. I myself put in the first transistor that I had, and didn't choose between them. It worked fine as you can se in the video. Should you experience hum or something, from the power supply, then you should resort back to the transistor arragement in the original design as shown in the original Noise Cornucopia Schematic

Here is the verified layout.

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

Stripboard only. 
For extra clarity: connect the 'Base [B]' of the transistor to copper strip 'I' and the 'Emitter [E]' to copper strip 'G'.
The cut on position D17 in the layout above, has been moved to position D21 to make it more visible.



Below is an overview of the cuts and the wirebridges alone. This is seen from the component side! As ever, mark the cuts on the component side with a black waterproof marker and then stick a pin through the marked holes and mark them again on the copper side. Now you can cut the copper at the marks with a sharp hand held 6 or 7mm drill bit.


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:


The noise output from the transistor goes through a highpass filter consisting of the 100nF capacitor and the 2 MegaOhm resistor. This creates a filter cutoff frequency of 0.8Hz letting through all the frequencies and rejecting any offset voltage. Should you experience an offset voltage after the filter then lower the resistor value from 2M to 1M. That will make the cutoff frequency 1.5Hz.
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 on the front panel. 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 and starting to clip, 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. I've had confirmation that this solution works just fine.
To be honest, you don't need a gain option in a noise module like this, so you can just as easily forget about the Gain potmeter and put in a trimmer, set it so the output of pin 7 gives +/-5Vpp noise level and leave it at that. That's also how it was intended in the first place. The gain option was just my own idea.
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. The pinouts are all the same. 
This module is designed to work on a dual 12 Volt powersupply (so ideal for Eurorack systems) 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 LoPass and HighPass etc is more intuïtive than to assign different colour-names to the noise. It's also a useful type of noise to have because a lot of people prefer mixing LoPass noise into the signal path instead of white noise because the LP noise sounds less muddy.

RANDOM GATES GENERATOR:
I've made a separate layout for the random gates section of the MFOS Noise Cornucopia design, for those interested in adding this on. The two 7 pin headers are there to provide a choise in randomness of the Gate signal. See the original Noise Cornucopia article for the schematic drawing. The signal has the most randomness if you place the jumper on the lower settings. The higher up you go the less random the Gates get. 
NB.: Place only one jumper on the pinheaders!! 
If you have a rotary switch with 7 positions you could use that, instead of the pinheaders, and make a feature of it by placing it on the front panel. That's up to you. Let me just say also that I have not built this random gates module myself but I've gone over the noise cornucopia circuit schematic with great precision and it's a very simple layout so it should work fine. If you have built this layout please send me some feedback about how it's working. I've had some feedback saying if it doesn't work like it should to put a 2,7MOhm resistor in parallel over the 10pF capacitor.

Personally I don't find this type of random gate circuit very useful. It's not synchronized in any way and you only get one output with a random pulse train on it. I would prefer the Yusynth 8 Random Gates project where each pulse has it's own output. What I would prefer even more is to pair a Sample and Hold circuit with a noise generator like the one above. But it's all up to you of course. 
(Check the comments below to read about Tim's findings when he breadboarded this circuit to test if he could build it with 7 outputs instead of one.)

Here's the layout I made for this section: 


Here's a demo 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:






To finish I want to direct your attention to a great video by Moritz Klein about building noise modules where he explains the theory behind it very well.  Click here to see the video on YouTube.

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.

If you find these projects helpful and would like to support the website and its upkeep then you can buy me a Coffee. There's a button for that underneath the menu if you're on a PC or Mac. Or you can use this PayPal.Me link to donate directly. All donations go towards the website and projects. Thank you!

23 comments:

  1. Hi,

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

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

      Delete
  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?

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

      Delete
  3. I’m a little confused as to what the 7 pin headers do in the random gates module. Is that to connect the noise and gates modules?

    ReplyDelete
    Replies
    1. No they provide a choise in randomness. The further up you place the jumper the more random the Gate signal gets (or the further down, I'm not sure).

      Delete
    2. I've added the explanation to the text.

      Delete
  4. Thanks so much for answering my question. I just wanted to say that I love your site. As a newbie in this craft, this is my go-to site. It is a huge help and an invaluable resource. There is nothing else like it on the web! Thanks again

    ReplyDelete
    Replies
    1. Wow thank you very much. I'm glad you like it so much!

      Delete
  5. Eddy, heeft you granular noise signaal ook een offset. Ik heb +1.2 volt offset op de positieve en negatieve spikes.

    ReplyDelete
  6. I have built a composite of your design, Ray's (I like his diligent use of power filtering capacitors) and the pink noise filter found on https://syntherjack.net/make-some-pink-noise-generator/ It is supposed to be pinker than Ray's design.

    I notice that you don't use power input filtering in your stripboard circuits. Do you not notice noise where noise is not wanted or is it not a concern for you?

    ReplyDelete
    Replies
    1. Hi James. Yes you are right, I usually don't include power filtering because I don't often need it. I use a linear dual powersupply that is very well filtered but in those modules that really need it I do include them. And I have started to include them now as standard in the latest projects I publish. It was something I just didn't concern myself with when I first started to build my synthesizer and I haven't gotten into trouble yet ;)

      Delete
  7. Hi, can I use the 2N2222 instead of the BC547 for the 5-noise module?

    ReplyDelete
    Replies
    1. I'm not sure you'll get the same noise from that transistor but you can try it. Do remember the pinout of a 2N transistor is reversed from that of a BC transistor so it'll be E-B-C instead of C-B-E.

      Delete
    2. Hi, I built it with a socket and try different transistors. I'll let you know which ones I liked the most. But I will for sure get a BC547 when I order the next components :D

      Delete
  8. Hi Eddy, do you think I'd be correct to assume that if I give all 7 outputs of the CD2024 on the gates generator their own output jacks and indicator leds, i'd have 7 different random gates on one panel? It seems logical to me that it'd work but I might be overlooking something.

    ReplyDelete
    Replies
    1. Yes, I think that's correct unless I myself am overlooking something too LOL. You can also build the 8 Random Gates project by Yusynth that I have on the website. Just to be sure. That one works very well.

      Delete
    2. For those who want to do this, I just breadboarded it and it works

      Delete
    3. I do recommend against this though. when I breadboarded it I used my finger as a noise source, on the module itself the random circuit is really unstable for me and really only worked on high frequencies that aren't really useable anyways. Lots of people online seem to have trouble with the gates as well, and having seven ungated/unsynced random gates just isn't that useful musically. Like you said, just build the Yusynth module. I wanted to do this because I use standard widths for panels and the panel would be rather empty with just the noise module, so I'll fill it up with the simple double sample & hold circuit by Skull and Circuits and normall the inputs to the white noise.

      Delete
    4. Thanks for your reply Tim. This is very useful information to share with those who might think of trying this too. I added my personal thoughts about this to the article too where I share your idea to use a sample and hold to pair with the noise generator instead of a random gates circuit. That's much more useful to have.

      Delete

Note: comments are moderated and do not appear straightaway. Your first comment is not allowed to contain any links.