Getting the SUPER SIMPLE OSCILLATOR to oscillate

The SUPER SIMPLE OSCILLATOR keeps causing headaches, because while it’s a SUPER SIMPLE CIRCUIT (and SUPER CHEAP; the active component costs you a couple of cents) it’s not necessarily a SUPER SIMPLE BUILD, as can be seen in the recurring “oscillator trouble” threads in this forum.

So I thought I’d post some suggestions for how to build the oscillator step by step, in a way that will at least tell you if you’ve managed to get it to oscillate. I’ve tested this with a variety of transistors, and have never had any problems getting things to work.

For this exercise, you need the following:

  • A breadboard, or some other means of connecting wires and component legs to each other.
  • A small-signal NPN transistor, e.g. 2N3904, BC547, or similar.
  • A relatively large electrolytic capacitor (ideally 1000 uF but e.g. 470 uF will also work)
  • A 1k resistor.
  • A LED (any single-color LED works, not RGB or colour changing varieties).
  • A voltage source that can provide at least 15 V. Ideally you’d use a lab supply, but if you don’t have that try two 9 V batteries in series. 12 V may not be enough.
  • (optional, but strongly recommended) a voltmeter.

(To move on to audio output, you also need a 100k resistor for the output, smaller capacitors, and a potentiometer, but that’s for later)

First, use the voltmeter to measure your voltage supply, to make sure you know what voltage you’re working with. As mentioned, 15 V is a good start if you have a lab supply, otherwise 18 V is easy to get from two 9 V batteries. You can go beyond this, but I’d keep it below 25 V or so.

In the drawings below, the supply is represented by the circled :heavy_plus_sign: and :heavy_minus_sign: symbols.

Next, turn off the power (or disconnect the batteries), and wire up the resistor and the LED in series:

Turn on the power again, and make sure the LED turns on. Once its lit up, measure the voltage across the LED. It should typically be 2-3 V, depending on what LED you have. If you measure the voltage across the resistor, it should show the supply voltage minus the LED voltage.

Now turn off the power again, and insert a capacitor between the resistor and the LED. Note the polarity; the − side is usually more clearly marked on the capacitor, and should go towards the LED and the − supply.

Turn on the power. The LED lights up at first, but will go out after a while. Then turn the power off, wait a short while, and turn it on again. The LED will light up again. (If you have a larger capacitor, you may need to wait ten or more seconds to give the capacitor a chance to discharge enough for the LED to come on again). If you keep doing this, you’re now the switching device of a relaxation oscillator. Congrats!

If you measure the voltage across the capacitor with the power turned on, it will show you the supply voltage, or close to it – the capacitor is fully charged, and there’s no voltage left for the LED. If you turn off the power, it can keep this voltage for quite some time. To avoid damaging things, we need to fix this before moving on, by discharging the capacitor. This is IMPORTANT, so I’ll repeat it in larger type:

IMPORTANT: Before continuing, you need to discharge the capacitor. Turn the power off, and put the voltmeter probe, or a screwdriver, or a piece of unisolated wire, or some other piece of metal across the capacitor pins and hold them there for a short while. If you want to double-check that you succeeded, measure the voltage between the capacitor pins; it should be close to zero (a few mV is normal).

With that done (you did it, right?), we’re finally ready to turn the circuit into an oscillator, by adding an electronic switch to the circuit. Locate the pinout for the transistor type you have (i.e. which pin is which; ideally you do this by looking at the transistor’s datasheet, they’re easy to find on the Internet), and identify the emitter (E), collector (C), and base (B) pins. I’ll post some common pinouts below.

Cut off the base pin, you don’t need that and leaving it sticking out but unconnected may mess with things (but if you feel experimental, you can skip this step).

Make sure the power is off and the capacitor is discharged (you did discharge the capacitor earlier, right? – just checking, since it’s really important), and connect the transistor across the capacitor, with the emitter towards +, like this:

Turn on the power. The LED should blink or pulse at somewhere between one pulse every two seconds and a few pulses per second, depending on voltage, transistor, and capacitor.

What happens here is that the capacitor charges up, at a rate controlled by the resistor, and once the voltage across the capacitor (and transistor) gets high enough, the transistor turns on (due to a transistor property with the slightly obscure name “reverse avalanche breakdown”). This very rapidly discharges the capacitor through the transistor, until the voltage drops enough that the transistor turns off again, and the cycle starts over.

Here’s an oscilloscope snapshot of an oscillating 2N3904, measured across the transistor/capacitor pins. Note how the voltage climbs to 10.7 V, and then immediately drops to 7.4 V when the transistor opens, at which point it turns off and the capacitor starts charging again. This is independent of the supply voltage, within limits.

(So if the capacitor discharges through the transistor during normal operations, why was it so important to discharge the capacitor earlier? The difference is that with only a LED in series, the capacitor will charge up to the full supply voltage, once the LED turns off, while it doesn’t get that far with the transistor in the circuit; that difference in charge can be disastrous).

Once you’ve seen it blink, turn off the power, and add the rest of the components in LMNC’s circuit. You can start with the potentiometer, and test the blink circuit before moving on. Then replace the capacitor with a smaller value (the relation is linear, so a 1000× smaller cap means 1000× faster oscillation), and finally add the output resistor. Good luck!

13 Likes

And yeah, common pinouts: 2N3904 has the emitter to the LEFT if you have the flat side/markings towards you, BC337 and BC548 have it to the RIGHT. The base is in the middle on all three (cut off in the photo):

(actual transistors tested in circuit)

(cannot find my camera charger so you’ll have to do with a mobile snapshot where you cannot really tell what’s what without the annotations, may fix that later :slight_smile: )

9 Likes

Nice tutorial, a couple comments:

That sentence is likely to be so much gibberish to many of the people needing this writeup. You do address that some more below, but a few more words here about e.g. what a pinout is, what a datasheet is, and how to locate the latter and how to find and understand the pinout might be a good idea. Or just omit that sentence, skip all the talk about pinouts and datasheets, and instead say below “you need to know which way around to put the transistor, here it is for some common transistor types, if you have a different transistor look here (link to a post about how to find and read a datasheet)”.

Discourse turned your paren-c-paren into a copyright symbol…

6 Likes

Good guide. You bet your banana I’m referencing this if I ever decide to make this.

5 Likes

very good idea to make this thread, nice work !

3 Likes

maybe add the polarity of the LED, like this :
led polarity

6 Likes

Another comment:

A problem here is that there are several circuit diagrams on that page, with and without CV, with and without tone control, stripboard and schematic — and they aren’t very consistent with each other or with the sources, e.g. Kerry Wong’s blog. Some show the capacitor from emitter to ground, some show it from emitter to collector. Some show the output resistor connected to emitter, some to collector. The schematic appears to show the transistor connected incorrectly (or else depicted from underneath). In fact the capacitor can be connected either way and it’ll work, and the output resistor will work either way, but it causes confusion. People have also gotten confused by for instance the LED in the stripboard diagram whose size leads them to think it’s connected to strips further apart than it is. Maybe a clearer, definitive schematic and stripboard diagram should be added here. And/or a Fritzing breadboard diagram:

5 Likes

maybe we should also add a link to this thread in all the threads that have already been opened on this oscillator.
if you agree I will do it

7 Likes

I do spell out what I think you should add next, but my thinking here is that if you take your time to go through these steps one by one, and you get them to work on the breadboard, you have a better understanding of how the core components fit together and what they do, and are better equipped to build and debug existing breadboard/stripboard layouts. But having additional tips and tricks in the comments cannot hurt, so keep them coming.

4 Likes

Super difficult oscillator.

4 Likes

Nah, not really. A similar tiny-step-by-tiny-step guide for e.g the CEM3340 would be a LOT longer.

4 Likes

Hey @fredrik thanks for this! I followd your instructions step by step and managed to get everything to work!! Now, where should I insert the resistor to go to audio out?

4 Likes

There’s a junction where the potentiometer, the transistor (emitter), and the capacitor (+ side) all meet. Connect the output resistor there, other end to speaker.

1 Like

IT WORKSSS!!! Though at a veeery low level. OMG finally!!!

4 Likes

Simple to build, simple to get it working — if it wants to. Impossible if it doesn’t want to.

I think it’s a great circuit for beginners to try, if they understand that some transistors just don’t want to oscillate at the voltage they’ve got, and some don’t want to oscillate at all, and if this step by step approach works right up until the transistor goes in, and then doesn’t, they should just try another transistor and/or say “okay” and go build a Schmitt oscillator or an APC or something.

In that sense it’s not an ideal beginner’s circuit. But it is easy to build, and might work.

5 Likes

Congratulations! Here’s a couple more projects to try out:


4 Likes

Thanks man! I’ ll check them out asap! Now the real deal is gonna be to transfer the circuit on the stripboard eheh

2 Likes

think about the current path through the components, like in the tutorial and it’s okay go

4 Likes

I agree that the circuit design is very simple. Getting it to work clearly isn’t. Literally my first attempt at soldering in 1971 was a multivibrator circuit using one or two OC71 germanium transistors (germanium was cheaper and so popular with British hobbyists.) It worked first time. That was a simple build.

This design, though ostensibly simpler, exploits one of the breakdown modes of a transistor, and as such it isn’t reliably reproducible. If you’re using a transistor in avalanche mode to make a multivibrator, you’re doing it the hard way.

2 Likes

The classic transistor “astable multivibrator” has eight components and eighteen connections that need to be done correctly, plus a bit more if you’re adding a LED. This has three components and six connections, plus the LED. It was probably not just the circuit itself that made your build work the first time :smiley: (Have you tried building this one?)

(multivibrator in quotes since that’s a silly name for a square wave oscillator, as noted elsewhere.)

3 Likes