4 Latching Momentary Power Switch Circuits

written on Saturday, March 12, 2016

Not so much time ago I built a fan driver from my fume extraction (powered from 4 x AA-battery accumulators), and it became clear that the sort of ON/OFF push button circuit is required. My basic research revealed two principle results:

  1. This circuit from EDN, copied by Talking Electronics (Colin Mitchell) (or vise versa, can not really verify)
  2. Another original latching circuit from END.

Each of these circuits has its problems:

  • the first circuit simply doesn't work as is
  • the second one uses more expensive MOSFET transistors (and rare low-voltage P-Channel ones), which often do not work with 5v - or less - power supplies.

So, after some experimentation, trials and errors the following circuits were elaborated:





- and later soldered and tested (in the same order from top to bottom) using my Stripboard Manhattan technique:

Principle of Operation

In each of the circuits you have a capacitor and a target transistor. By charging or discharging the capacitor the base is either stripped off charge carriers or flooded by them depending on the transistor type. Usually target transistor starts in the OFF state, and capacitor charge has a polarity to open it. After opening a transistor (putting it in a conducting state) cap is charged with the opposite polarity, thus being able to close a target transistor.

Let take a look at the second circuit. C3 is a set capacitor - it ensures the Q2 tr-r starts in a closed state and thus the whole system is not conducting. Via R1 and R2 C2 charges to the potential of the Vcc rail. Time constant (R1+R2)*C2 defines the time which is required to allow C2 enough charge to be able to turn-on the whole system. Now we have a storage with "tons" of electrons in it; by pressing push button we inject the absolute majority of them (R5 can be considered almost as an open circuit, comparing to the base-emitter junction impedance). Q3 starts to conduct (due to the effect of positive feedback - and the whole system is conducting and working now.

Collector of the Q1 is now sitting a fraction above the negative rail potential, thus completely discharging C3 via R2 and making it almost equal to a potential of the negative rail. R2*C3 time constant defines the time required to discharge C3 to the potential of the negative rail. Finally, C3 is able "to suck in" the charges from the base of Q3, thus putting it back to a non-conducting state if the S1 button is pressed again. That's it!

So, basic principles are very simple:

  1. Find charge paths of the capacitor
  2. Find a target tr-r.
  3. When the whole system is put in a conducting state, find a discharging path and a target tr-r.

That's all!

Fixes and tweaks.

  1. 2nd circuit

    • 1 MOhm is too big for the R2 - it will have no influence on the turning-off action, while it should act as a pull-up resistor during the 'turn-off' period. Basically, the circuit doesn't turn-off with 1M R2 resistor value.
    • 470k for R2 is too big it won't allow to switch the circuit off instantly, especially if C2 has bigger value. I replaced it with 1k to speed up the overall charge/discharge times.
    • power part can be directly integrated into circuit: R9/10/11 and Q4/5 could be completely removed, value of C2 should be increased to 47uF (at least in my tests is was the most stable value) and Q3 should be switched to power transistor (or pair of BC337/8, depending on the load.
  2. 3rd circuit

    • schematics was completely re-drawn, as original one makes it very hard to read
    • 10k value of R3 in the original circuit (R1 in my variant) was a bit too big - not enough current entered the base of Q2 to open it fully, so I tried 5k6 and it worked perfectly
    • R5/R6 voltage divider should be introduced in order to prevent big voltage drop across R7 and R8 1R resistors

That is all for the moment! Please, try the circuits, build them and adapt for your own needs!


Categories: electronics

blog comments powered by Disqus