[Eric Wasatonic] had a box of SWB2433 transistors that he had very little information about. In order to discover their properties, he fired up his curve tracer to compare these transistors with more common ones. He noticed the SWB2433 exhibited negative resistance while the similar curves of a 2n3904 didn’t. Then he reverse-biased the two transistors: the negative resistance region on the 2n3904 was less than that of the SWB2433, but it was there, and a 2n2222 had a bigger region. Using this knowledge, he developed a relaxation oscillator circuit which uses a negatively biased transistor.
Using one transistor, one resistor and one capacitor, he describes the circuit and how the components affect the frequency of the sawtooth wave the oscillator creates. [Eric] uses the oscillator to build a simple LED blinker and shows what happens when he changes the transistor and adjusts the voltage or resistance. He also shows the circuit as a tone generator and adjusts the tone by replacing the resistor with a potentiometer. And then, for fun, he modifies the circuit to show the oscillator as an AM transmitter. Check out his video after the break.
Of course, [Eric Wasatonic] didn’t come up with this idea (he links to this Simple LED Flasher Circuit page in the description of the video) but [Eric] shows us the process he took discovering the negative resistance of the transistor and the results of the circuits. Even though there are other ways to blink a LED and other ways to make an oscillator, [Eric]’s circuit is about as simple as it gets.
[Thanks, fede.tft]
that + constant current sink = perfect sawtooth for a class D amp
gonna be a heck of a glitch amp. Temperature unstable, EMI sensitive
A few years ago I decided I wanted to start learning electronics. I didnt have any idea about parts to order or where to start and I didnt want to spend heaps of money just to try something. So I thought I’d salvage some stuff from hard waste around my neighbourhood, they are full of components. I made a small breadboard out of computer IDC connectors and ribbon cable and made one of these simple led blinkers from salvaged parts, resistors and all. I learnt that a lot of electronics is actually about getting ‘close enough’ and that sanding, gluing and soldering IDC connectors into makeshift breadboards is a pain in the arse not worth a few dollars with free shipping for an ebay special. But I did sit watching that blinking led with mangled leads with bits of solder still stuck to them for about half an hour. I wonder how many of my neighbours worried about what the kid running into the evening with their old electronics was upto.
Interesting. I wonder if this avalanche business is damaging the transistor?
I believe the correct technical term for the function above is sharkfin wave.
This looks like the flash pattern of the red LED on a lot of immobilisers and old car alarms.
Hey I think I used that exact curve tracer at Penn State back in the early 90’s and it was old then :)
At one tie, one could buy LEDs that blinked. The flasher electronics was right on the same chip as the LED. You could also get LEDs with a current limiter, so no need for an external current limiting resistor. I have no idea if you can still get them, though I bet if commercial use required such things, they can get them, just not available as parts to the masses.
For a long time there was the LM3909, intended for flashing LEDs, but a whole slew of secondary applications. It was less versatile than the 555, but for a flasher, was very low current. It lasted a long in production, but never saw much application. Though once it was dropped, Michael Covington came up with a discrete component replacement, trying for minimal component count.
Michael
MakerFaire sells a soldering badge kit IIRC also available through AdaFruit that has these LEDs. The LEDs are undoubtedly available in single part quantities.
I remember making a flasher with a LM3909 back in the late 1970s. It was very efficient. In regards to this circuit, I can think of one that is even simpler, although not so stable. I’ve seen situations with LED’s and an “almost-dead” battery where, unloaded the battery voltage is just a bit over the drop voltage of the LED and current flows until the voltage drops to the drop voltage and current stops. The battery recovers a bit and then the cycle repeats. I’ve seen that often with numerous 3×1.5 volt pocket LED flashlights. Thats a simpler circuit. :-)
Any chance the SWB2433 is a type of unijunction transistor? I think I read somewhere that the non-programable unijunction transistors can be used to detect a magnetic field.
It looks as if you could add 4 extra components and have a version of Chua’s circuit. Or 2 more and perhaps have a sine wave output?
This is due to the well known Tunnel effect. Its also called an Esaki Oscillator. Ludic science on youtube showed a circuit built with jellybean parts:
https://www.youtube.com/watch?v=JqAZt97ZQQA
Thanks for info. I built one with a BC337, 18 volt battery and started at 15K resistor. All worked well at first BUT the transistor seemed to slowly lose its properties. I had to keep dropping the resistor value to keep it flashing, else I just get a dim glow. Current from batteries is so low I doubt they were running out, almost as though I was slowly damaging the transistor. At 9 volts no oscillation at all. Happened over about 5 minutes. Are we wrecking transistors?
1a- Transistors have a negative region beyond collector breakdown voltage. When collector breakdown is reached, amplification effect on base makes collector voltage to drop.
2. SWB2433 is NPN silicon VCBO = 50V – VCEO= 20V according data at:
https://br.ebay.com/b/Small-Switching-Transistor-Transistors/4666/bn_119901163
Continuing former comment:
1b – Negative resistance in reverse connection.
A transistor can work reversed, this is, using emitter as collector. HFE is very small, though. And breakdown voltage for the emitter-base junction is about 5V. So a transistor connected reversed has a VCBO of about 5V and should also have a negative resistance region.
1c- POWER TRANSISTORS usually do not have a negative collector resistance region. When collector breakdown is reached, voltage continues rising slowly as current is increased.
Thanks for your posting about relaxation oscillator. Had already forgotten it is possible to do it with negative impedance device.
I first saw it at university lab exercise using a tetrode back in 1971. The oscillator by negative plate dynamic resistance was called dynatron. It made use of a region of negative plate resistance on tetrodes. One could connect a pentode as a tetrode to make use of this feature.
The characteristic of negative resistance can be used to store one bit of information on a transistor. The transistor is taken to the breakdown and VCBE takes a lower value. It can be reset by reducing collector voltage under a certain minimum and allowing it go go back to a high value. The transistor is triggered into breakdown by opening the base or having a resistor from base to emitter and no bias voltage.
More about breakdown voltage on transistors.
While most transistors have a breakdown voltage of about 5V for the base emitter junction, the actual collector junction breakdown voltage for a transistor is hard to predict. Manufacturer only guarantees BVCBO is ABOVE certain value. When selecting transistors to meet a requisition, they verify the transistor DOESN’T BREAK at a certain voltage. They don’t find out at what voltage will it actually break. So a transistor rated at, say, 35V might break above 300V, as I verified once with some Texas transistors TIP29A.
And here is a paper from 1964 about using transistors in avalanche mode to make use of breakdown to change from one stable condition to another, and a graph of negative resistance region.
https://lib.dr.iastate.edu/cgi/viewcontent.cgi?article=4858&context=rtd