Super Simple Oscillator: Simple except when it isn’t

What Look Mum No Computer (Sam Battle) calls the Super Simple Oscillator is described in this video (which followed up on this video) and this web page. That may be a case of too much information, because some of what is said and shown seems to be inconsistent with itself. I’m writing this post to try to make things a little clearer, and to bring you this warning:

If you’re looking to get into building electronics, maybe thinking about noise boxes and synthesizers, an oscillator consisting of only four parts might seem a great place to start. “I definitely can build that!” you might be thinking, and you’re right — you can. Whether you can get it working or not is another question, and if you can’t, it might not be your fault!

People have come to the LMNC Discourse forum looking for help getting the SSO working. Sometimes we’ve been able to help. Sometimes we’ve just had to say, “Look, this circuit is weird. We can’t see that you’ve done anything wrong and it may be you haven’t. Why don’t you build an APC instead? It’s more likely to work.”

So let’s take a look at this strange little oscillator.

I don’t know who came up with it first. Dick Cappels claims to have had the idea in 2006, but it wouldn’t surprise me if someone else had done it earlier. Kerry Wong wrote about it in 2014. Both of them wired it as an LED flasher running at sub-audio frequency. Perhaps LMNC was the first to make an audio oscillator out of it.

The basic idea here is pretty simple (but you can skip this paragraph if it confuses you). You take an NPN transistor and wire it “backwards” with the middle leg unconnected, with a capacitor in parallel and a resistor in series. When you apply voltage the transistor doesn’t conduct, so charge builds up on the capacitor until the voltage across it reaches a certain point. Then the transistor goes into avalanche mode, suddenly becoming a conductor; the charge on the capacitor discharges through it. That means the voltage on the transistor drops, it leaves avalanche mode and stops conducting, and the whole cycle repeats. The time it takes to charge the capacitor, hence the oscillator frequency, depends on its capacitance and the resistance in series. If you add a second resistor at the top of the transistor and connect its output to a speaker, you’ll hear a tone.

So far, so simple. So where’s the problem? Part of it is that, even as simple as this circuit is, there are several versions of it shown in the links above, at least one of which is wrong. The other part of it is, as I said, you’re connecting the transistor backwards with one leg disconnected. That poor transistor wasn’t designed to be used that way. It should be no surprise the results can be unpredictable. Some transistors don’t reach avalanche until a higher voltage than others, and if you don’t supply enough voltage, it won’t oscillate. Some transistors don’t seem to work at all. Wong wrote:

Another interesting thing I noticed is that the same transistor from different manufacturers or even different batches can behave quite differently. For instance, while I could get one batch of 9014’s to oscillate at around 12 to 13 volts consistently, the 9014’s from another batch would not oscillate at all. So your results might be quite different than what I have here.

In other words, you can build this ridiculously simple four- or five- or six-component (we’ll get to that) circuit in a couple minutes, and then spend hours trying to figure out why it doesn’t work, when the answer is, because it doesn’t want to. So should you build one? Sure! Why not? But should you expect it to work? Be pleased if it does. But it might not.

If you have a soldering iron and a piece of stripboard or perfboard you can throw this thing together quickly. Even more quickly if you have a solderless breadboard — and if you don’t, it’s a good thing to get. On a breadboard the circuit looks like this.

Let’s talk through this. Starting on the left, a wire connected to the + terminal of a battery comes in. Parts connected to the same vertical column on the breadboard are connected together, so that wire connects to the left end of a resistor (the part with colored bands, further to the lower left). The right end of that resistor connects to the middle leg of a potentiometer (in back). The pot’s left and middle legs are connected via a short wire. The right leg connects to the left leg of a transistor whose middle leg connects to nothing. Its right leg connects to left leg of an LED. Its left leg also connects via a wire to the right leg of a capacitor, and to another resistor whose other end connects to a wire that goes to a speaker. The left leg of the capacitor and the right leg of the LED both connect to two wires, one of which goes to the other terminal of the speaker, the other of which goes to the – terminal of the battery.

But we need some details. What kind of battery? What kind of LED? What kind of capacitor? What kind of speaker? And so on.

Start with the resistors — the color bands indicate what kind they are. The brown, black, red, gold one is 1k and the brown, black, yellow, gold one is 100k. Resistors are unpolarized, which means they work equally well either way around.

The pot is 10k. It gives you a means by which to vary the resistance and therefore change the frequency. With some resistance values the oscillator stops working, so if you hear nothing, try turning the pot one way or the other.

The capacitor is a 10µF electrolytic. Or not; you can use different capacitor values to get different frequencies and if you use something like 330µF the frequency is so low it’s sub-audible and you can see the LED flash. This part is polarized, so it has to be installed the right way. In the above diagram, the longer of the two legs goes on the right. The shorter leg is on the same side with the white stripe and goes on the left.

The LED is, well, an LED. Pretty much any run-of-the-mill LED will work, if you install it right. The long leg goes on the left, the short leg on the right. The body of the LED has a flattened side which goes on the right, but it’s hard to see.

How about the transistor? Well, that’s a big part of the story.

It needs to be NPN, not PNP. One of the most readily-available NPN resistors is the 2N3904. That’ll work, but it will need about 12V to make it oscillate, so your battery has to supply more than that. 18V batteries aren’t so easy to come by, but 9V ones are, and if you watch the LMNC video you can see how to use two of them to supply 18V to this circuit. The exact voltage doesn’t matter, I’ve used 24V for instance, but 18V is easiest and safest. Don’t try to use a wall wart transformer unless you know what you’re doing.

Most other transistors also need about 12V to work, but a few come in lower. The SS9018 only needs about 8V, so can — maybe — work with a single 9V battery.

You also need to install the transistor the right way around, and this is where it gets fun. Your transistor probably will look like this:

(Only smaller.) The three legs are called e, b, and c (emitter, base, and collector). Most of the time they are e, b, c in that order from left to right when the flat side is up. The 2N3904 and SS9018 are e, b, c, for instance. But if you have a BC337? It’s c, b, e. Just to piss you off. In the above breadboard layout you want e on the left and c on the right, so usually you have the flat side toward you, but with an oddball such as the BC337, the flat side is away from you.

One of the diagrams on the LMNC page is a schematic in which the transistor is represented as a D-shaped symbol, and it’s labeled 2N3904, but the way the D is facing, if it’s a 2N3904, it’s backwards. The stripboard diagrams on the same page have the transistor facing the right way. See, you’re not the only one who gets confused.

As for the speaker, use a cheap paper cone thing from your electronics supplier, not your Bose stereo. Or an expendable cheap phone speaker, or something. Headphones or earbuds aren’t recommended.

Now, if you compare the connections in this layout to the diagrams in the pages I’ve linked, you’ll find some lack of agreement. Mostly that’s because it doesn’t matter.

The capacitor, for instance, can go from the transistor emitter to ground, or it can go from the emitter to the top of the LED. You’ll get a change of frequency, but it’ll work either way.

The 100k resistor can connect to either the emitter or the collector. Again, you’ll get different frequencies, but either way can work.

You can even get rid of some of the parts. You can skip the pot and just connect the 1k resistor directly to the emitter, for instance. You won’t be able to vary the frequency, but it’ll oscillate. Don’t omit the 1k resistor, though, because if the battery connects directly to the transistor it’ll probably fry it and the LED too. Ask me how I know.

You can omit the LED. I think it’s there because the circuit started as an LED flasher and LMNC just left it in, because LEDs are cool and because if it doesn’t light up, you know something’s wrong, but you can leave it out.

Or you can add parts, to make the circuit behavior more interesting, but do that only once you get the basic version working.

So in the simplest form the circuit is two resistors, a transistor, and a capacitor. Four components; what can go wrong? Well, the transistor or the capacitor can be backwards. Or you can have a bad connection. Or a bad wire. Or a bad speaker. But also, as indicated, the voltage can be too low, or the transistor can simply not want to work. If it doesn’t oscillate, check your connections, swap out components, raise the voltage (safely and within reason). And if it still doesn’t work, smile, say “stupid simple oscillator”, and walk away. Go build an APC. It’ll work.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s