"Jimmy T" wrote in news:dc4f9i$t8g$ email@example.com:
That is a neat trick, that circuit... Did you try increasing tha value of the cap significantly? What are your circuit values? The circuit will need to be somewhat optimised for the relay characteristics wrt it's dropout voltage and coil resistance.
Hello Jimmy have a look here at Terry Pinnell's library of schematics.
Here is one that might suit you..
Here is another.
After you build up your circuit and it doesn't work, if you send Terry an email "Sorry Terry it doesn't work" he will sort out your problem. He is just that sort of fellow. A real gem. Loves his electronics hobby.
relay latching current is obviously OK, and R1 can clearly pass the holding current, so C1 kinda has to be the culprit. Like Geoff says, crank up C some more. At a guess I'd pick 10-100uF. Ideally the cap energy needs to be much greater than that stored in the relay coil - not that you necessarily know the coil inductance (which varies with contact position).
So I built one. The 12V relay sucks 40ma at 12Vdc. From 24Vdc, I drop
12V at 40mA across R1 = 300R. I chose R1 = R2 = 317 Ohms (I have thousands of leaded 475R resistors, and 4 leaded 330R resistors).
With C = 100uF, the relay turns on but not off. ditto for 570uF. 1040uF and it turns off again. The relay holding current is about 10mA. With the pushbutton shorted, it oscillates with a period of a few Hz.
Of course during normal operation R1 dissipates 12V^2/317R = 454mW. quite a bit....I used 475//(475 + 475) so the single 475R dumps 300mW, which its OK with :). With the relay OFF, both 317R resistors dissipate
454mW. not so good for battery gear then eh?
OK, what about a circuit running from 12V with re-scaled R1, R2? At a guess, drop 2V across R1 giving R1 = 2V/40mA = 50R. R2 has 10V across it, so must be 5 times bigger than R1 IOW 250R. Using these exact values (thousands of 100R's): with 10mF of capacitance (5 1800uF 450Vdc caps, 2
470uF 25V caps and a 100uF 50V cap) it didnt turn off terribly well, depending on how I drove the "switch" (aka clip lead). It looks like its to do with the RC time constant...... I dont have any large electrolytics, so cant increase C above 10mF, but I suspect 20mF or so will sort it out. Not that thats a useful value in practice....
so R1 = R2 is a pretty good idea, and C wants to be a few mF.
but the circuit does work, and is about 100 years old - this is how vibrator power supplies were build, using an oscillating relay as both the oscillator and switching element.