loading
Make a cute + expressive 1-bit noise synthesizer with a logic gate and a few other spare parts.

Watch a demo video to get a better idea of what I mean, or listen to some improvisations: 321.

Step 1: Materials

Here are things you'll to do it exactly the same way I did. Some things, like capacitor size, will vary. Other things could be totally different (you might try a different logic chip, for example).

Tools

  • Soldering iron + solder
  • Hot glue gun + glue
  • Knife/razor
  • Wire Cutter
  • Breadboard

Materials

Extra

  • Oscilloscope
  • Power supply (for watching current draw)
  • Multimeter
  • Sound system or headphones

Step 2: Make FSRs

The interface to this device consists of four force sensing resistors. They can be anywhere from about 2x2 cm to 4x4 cm. If they're bigger, just make sure the wire is longer.

Plusea makes some really cool DIY FSRs, but I used my own style, because I don't have the same materials available. For Plusea's FSRs, you're working with around 200 ohms - 200 kiloohms max. For mine, you're working with 5-15 megaohms max. Commercial FSRs are in a similar range (e.g., this one is around 20 megaohms without a load). This maximum resistance is important to keep in mind when you're selecting capacitors.

Step 3: Prototype: Make a Square Wave

NAND gates have two inputs and one output, and the 4093 has four NAND gates on it. Here's how they're connected. When both inputs are high, the output is low. In every other situation the output is high. The 4093 also has Schmitt triggers on its inputs, which helps make it resilient to noise and more likely to oscillate regularly.

There are a few ways to get a NAND gate to give you a square wave. We'll use a way that requires one gate, resistor, and capacitor per oscillator:

  • Wire one input to high (9 V)
  • Connect one input to ground via a capacitor
  • Connect the output to the grounded input via a resistor

If you watch the output of this gate on an oscilloscope/listen, and switch out capacitor and resistor values, you'll notice you can approximate the frequency of the square wave by f = (1 / (r x c)). For example, when c = 470 pf and r = 10 megaohms, f = 212 Hz (measured at 250 Hz). This is where those numbers maximum resistance numbers come in handy for getting an idea of what kind of capacitance value to choose.

In short, if you use this circuit and change the resistance you can change the frequency. Furthermore, you can turn it on and off by wiring input 1 to ground instead of high.

Step 4: Prototype: Make the Circuit

There are two main approaches to making this synth.

  • Figure out what you want in advance, and draw up a circuit that works like that.
  • Make some sound, and start intuitively adding and removing components however you feel.

Surprisingly, the second approach works really well sometimes. The first Nandhopper I made happened that way, and I think it's the most interesting. Unfortunately, if you do things the second way it might be really difficult to get the circuit off the breadboard and onto the chip. When you put everything on the chip capacitance changes slightly between everything, so anything that's fine-tuned might not work anymore.

The first approach is a lot more predictable and repeatable, so I'll describe two layouts that have come from that perspective.

One layout just uses each of the four NAND gates as independent square wave synthesizers. In my experience, they're not genuinely independent and there is always some sort of leakage between the channels -- which can make things really interesting. Each pair of gates is mixed into one of the stereo outputs.

The other layout uses two stages: one driving oscillator, that is always running, and three oscillators stuck in a circular/triangular feedback loop. That's the one I breadboarded here.

Pick one of these, or design your own (or improvise your own, according to your ear).

Step 5: Solder Everything Together

Now that everything works on the breadboard, we're going to move it over to the chip.

Start with the power connector. Leave maybe ten centimeters of wire, just in case something goes wrong and you need to cut it off and reuse it. When you solder the ground and 9 V in place, it will help remind you where things are.

From here, I like "transplanting" the circuit: you have the breadboard next to the chip you're soldering onto, and slowly move components from the breadboard to the chip. This is one way to make sure you haven't forgotten anything. Make sure you don't forget that the orientation is different if your chip is upside-down and your breadboard is right-side-up!

Next, add the capacitors. I wrap the legs around the chip so it stays in place, then solder them on once they're where I want. One of the capacitors might have to stretch if it's going to ground, just make sure the leg doesn't touch anything it shouldn't.

Next, add any extra wires. If you're doing the independent oscillator design, you'll need wires for giving each gate its logic high.

Finally, add your resistors and FSRs. Resistors can be fit snugly under the chip, FSRs should be coming off the sides of the chip's pins.

Don't add the 1/8" connector yet, since you want to make sure the circuit works first. Solder some temporary connections (long wires) to the 1/8" connector.

Step 6: Plug It in and Finalize

Plug it in! Grab a 9 V battery (if you haven't already been using one), some headphones, and test it out. It should be outputting a 0-9 V signal, while sound normally runs at +/-1V. But sound systems also tend to have capacitors in line with the input that center things around 0 V and remove any DC bias. You should still be careful with volume: start with it off, and slowly turn it up.

As you improvise you might discover a lot of strange things: touching sensors together might cause weird sounds; depending on how you've wired it, touching the sensors without applying force might even cause some strange sounds.

If nothing works, check for broken connections or things in contact that aren't supposed to be in contact. Debugging these is probably more work than it's worth, and if it's not obvious you might be better off building a new one.

If everything works correctly, leave things plugged in and start hot gluing everything in place. Start with the audio connector, and then soldering "real" (shorter) connections between the chip and audio jack. Then hot glue everything that looks like it could ever come undone. This synth isn't meant to last for a long time, but the glue will help when you inevitably drop it. I leave things plugged in just to make sure I'm not breaking anything as I apply the glue.

Step 7: Variations and Notes

Variations

  • Try using something that isn't a NAND, but still has a truth table that allows you to make oscillators (for example, XOR, NOR or NOT).
  • Multiple Nandhoppers wired together: you've got ten fingers, right?
  • A different embodiment: imagine a sphere covered in FSRs with the same NAND-based synth in the center.
  • Try TTL instead of CMOS logic (running off a regulated 9 V battery, or some other power supply in the standard TTL range).

Notes


Thanks again to Dane Kouttron, who explained to me why this worked a long time ago.

If you have any comments, explanations of important things I didn't cover, or ideas for how to make a really responsive and varied noise synth out of a few components -- post below!
<p>Hi guys! This is awesome but I am serious amateur is this the simplest way to make a synth. I currently working on a project with e-textiles and want to make a tshirt that is the controller for a synth and this is the perfect. Any help? Cheers! </p>
Is there a way I can avoid the 4093? I don't seem to have that ic...
<p>Yes,if you have an other nand gate chip such as 74HC00 (pinouts are diffrerent)</p>
I love the 'dead bug' style of soldering everything together.
what exactly does this do?
kyle, this is awesome ... this thread is now old, and I see you're on to some truly amazing 3d scanning stuff.. but I'm still curious about this 'megaheterodyne capacitive' sensor you use to drive the LEDs... can you offer a schematic or even just a little more detailed discription of what that circuit is about?<br />
Hey uberklok -- yes, this is old, but I'd like to come back to it at some point :)<br /> <br /> The &quot;megaheterodyne&quot; capacitive sensor is basically this:<br /> <br /> http://www.thereminworld.com/pics/schematics/simple.jpg<br /> <br /> Running in the MHz range rather than KHz (KHz is the standard range for heterodyne sensing, a la the classic theremin design).<br />
I just love to look of this thing, and the tiny size! great intructable. But I need to find a better source for buying parts, Norway is so horribly expensive!<br />
The foam where my ICs was, is not conductive. I bought couple of 555s and they were on that black foam, and I tried does it have any conductance and it didn't..... Why yours was but not mine=0 : D<br />
thank you. <br />
Is there a way to wire a pot, so that I can change the pitch with that?
The four FSRs control different aspects of the pitch. Instead of using FSRs, you can use the center tap and one end of a pot, yes.
Cool. I was planning on doing a project with about 8 of these.
You used a 4093? I have a handfull of 4011's. They are the more widespread type, methinks. I suppose it will still work with a different NAND?
I've only worked with Schmitt-triggered gates, so I can't say whether it will work. Give it a shot!
I looked up the datasheet on them. My particular flavor of 4011 has schmitt-trigger inputs, so I guess it would work. The next problem is why there are two near identical chips in the 4000 series!
Triple modulated feedback loop?
I wonder if you could put a NAND device in between two microphones and an amplifier to get a weird mix of two sounds?
Totally, you'd just need to amplify them a bit before doing the logic (otherwise only the loudest sounds would mix). Audio is generally +/- 1V but logic runs at 5V or 9V, with cutoffs for HIGH and LOW at 1/3 and 2/3 the voltage (in the case of schmitt trigger inputs). I'm not sure what it would sound like...
It might sound like reverse audio clipping, but, not being an audio engineer, I don't know.<br/>A more do-able way may be to have an opamp act as an analog NAND.<br/>Lemme try an ascii-art schematic:<br/><br/> R1<br/>in1&gt;--<sup>v</sup>v<sup>v---\</sup><br/> | R3<br/> R2 L<span class="underline"></span>_o----<sup>v</sup>v<sup>v-------------------- |</sup><br/>in2&gt;---<sup>v</sup>v<sup>v--/ | |</sup><br/> | |<br/> \ |<br/> L---(- in) |<br/> (opamp) (out )&gt;------*------------&gt; circuit outputs<br/> | ------(+ in)<br/> |<br/> |<br/> V <br/> (GND)<br/>
Crap, it didn't work.
But that did!
conductive foam...very weird whats next sinking wood ,floating metal ,conductive glass,non flammable gasoline?
hahaha try lighting gasoline after it has been sitting on the pavement for 5 minutes. i tried and it didnt work :(<br/><br/>floating metal = boat/ship (already invented)<br/><br/>sinking wood = ebony (already, well, not invented but...)<br/><br/>and as for conductive glass; ------<br/><br/><a rel="nofollow" href="http://www.teralab.co.uk/Experiments/Conductive_Glass/Conductive_Glass_Page1.htm">http://www.teralab.co.uk/Experiments/Conductive_Glass/Conductive_Glass_Page1.htm</a><br/><br/>yeah, i know, im a smartass, but im bored and it is not even .5hours into 2009 lol<br/>
conductive gas?
Old radio tubes and television tubes are filled with gas. Electricity would jump through the gas from one metal part (Anode) to the other (Cathode).
I thought they had a vacuum in them, thus, the name: vacuum tubes.
Though I hate using wikipedia as a source "...an electrical signal by controlling the movement of electrons in a low-pressure space. Some special function vacuum tubes are filled with low-pressure gas." Vacuum tubes are generally a partial vacuum, whereas the existing gas inside is conductive.
Neon and fluorescent lights, not to mention mercury vapor.
Well i am fresh out of ideas...
neon lights?
now what...umm....underwater computer?
Submarines!!!!! um, those underwater-treasure-hunting-robot-things. <br/><br/>oh!! and this<br/><br/><a rel="nofollow" href="http://www.freepatentsonline.com/4658358.html">http://www.freepatentsonline.com/4658358.html</a><br/>
LOL!
very cool, i recently made one of these type of nandsynths, combined with an 8 step sequencer based off the baby10, added a few features like lm386 high gain amp, int/ext clock, CV out, clock out, clock in, individual pulse outs, all on a bolt style patch bay, and called it the NB8.. <a rel="nofollow" href="http://www.jnabeats.com/circuitbending/NB8.htm">http://www.jnabeats.com/circuitbending/NB8.htm</a><br/>
Awesome! Post some pictures of the internals, too! I really like the "wetness" of the raw sound you're getting, and the overall aesthetic of the device. You should also add some capacitive bends -- pairs of wires you can touch that "break" the circuit in weird ways. Long live nandsynths :)
i have to say, the LM386 is one of the nicest sounding, most organic gainy opamps, really contributes to overall liquidity of the tones you get when you push a lot of gain through the circuit! the TL072 is much more 'dry' &amp; clinical when compared. The only other one I've messed with is LM741 and its signature sound is somewhat in between the two, IMO. Yes.. should have included guts-shots.. but I have a nasty unprotected quad RC filter circuit just hanging out up top I need to tape up/shrinktube (when I 1st did, it grounded out).. should have put it on the board but ran out of room =[ <br/>
I haven't used the TL072, but I'd agree with your LM386 and LM741 assessments :) Best with the heatshrinking; messy or not, I'd still like to see some guts :)
It'd be cool if you hooked this up in a guitar signal chain and ran the guitar output through it and out the amp. There'd be some interesting sounds I'm sure
I've been messing around with this; I used on/off switches and pots instead of force sensing resistors. I like the way it sounds with random effects added:<br/><a rel="nofollow" href="http://www.youtube.com/watch?v=BMhaM8cBsF4">http://www.youtube.com/watch?v=BMhaM8cBsF4</a><br/>
Nice! Which layout did you use? It doesn't sound like they're all independent oscillators, but sometimes it's hard to tell the difference between other schematics.
My chip is the same chip as yours (same name, works the same) but loads smaller, and therfore is too small to fit on the breadboard. How can i get it on my breadboard? Help!
Could you post the name of the company that makes the chip, or a link to the datasheet? You could try a solderable protoboard instead.
[http://www.rapidonline.com/Electronic-Components/Integrated-Circuits/Logic/4000-Series-CMOS-logic-(surface-mount)/34486/kw/Quad%20NAND%20gate http://www.rapidonline.com/Electronic-Components/Integrated-Circuits/Logic/4000-Series-CMOS-logic-(surface-mount)/34486/kw/Quad%20NAND%20gate]Its on this page : <a rel="nofollow" href="http://www.rapidonline.com/Electronic-Components/Integrated-Circuits/Logic/4000-Series-CMOS-logic-(surface-mount)/34486/kw/Quad%20NAND%20gate">http://www.rapidonline.com/Electronic-Components/Integrated-Circuits/Logic/4000-Series-CMOS-logic-(surface-mount)/34486/kw/Quad%20NAND%20gate</a><br/><br/>its the 4093<br/><br/>cheers<br/>
The issue is that you're using what's called a "surface mount" component, which can not be placed on a breadboard. You could solder wires to the individual pins, but I recommend you just but a through-hole 4093.
Cheers, i'm going to the store soon so i can see the parts - hopefully i wont buy the wrong one again.
could you use wire wool intstead of the conductive foam?It may work...
Only if the wire wool has a significant resistance. If you take a multimeter to two points separated by a half inch, and the resistance is a couple kiloohms, you're set.
Hihi funny project. Makes me want to plug the iron this afternoon... :-)
I'm not quite sure I understand the need of a NAND gate. Is it just to produce a square wave?

About This Instructable

47,222views

133favorites

License:

More by kylemcdonald:Structured Light 3D Scanning Nandhopper 1-Bit Noise Synth Lo-fi Arduino Guitar Pedal 
Add instructable to: