Digital latch with Zero consumption when 'off'

[TaDa]

Hi!

Is it possible to make a latch circuit that is triggered on and off by a simple push-to-make switch. Such that the latch uses zero current when 'off'...

The 'on' push actually provides the 'on' signal _and_ turns the latch the circuit on.
The 'off' push tells the latch to disconnect itself from the main supply?

I'm finding it difficult to properly explain myself but I'm hoping someone will get my drift

 

[KMoffett]

Do you already have the switch? If not, why not just use a push-on/push-off mechanical switch? Are there other requirements that demand that this be a an electronic "digital" latch.
What voltage and current are you switching. Details...details...details!

Ken

 

[LTX71CM]

It's hard to understand why you can't just use a latching switch as KMoffett suggests. But to answer your question making a latching pushbutton circuit is possible. Use a pushbutton switch to power a relay, the relay then feeds back into itself to stay on. Once power to the relay is interrupted by yet another switch (or circuit) the relay will switch off until the first button is depressed again.

 

[Reloadron]

You could use a latching relay as attached. Need to N/O push buttons though. The relay is actually two relays mechanically ganged with two coils. They come in 12, 24 and 120 volt coils AC or DC. Push and release a button and one relay latches on mechanically so there is no current drain on either coil after the relay is toggled. Latching one unlatches the other and visa verse. Each relay in the example is DPDT.

Ron

 

[TaDa]

Ok, Sorry people - I realise that my request seems a little pointless - and, perhaps, highly gratuitous - the simple solution would be to buy a small push on/push off switch but...

I arrived at this request because I wanted to make a small dolls house light. Most human sized switches are somewhat out of scale and/or ugly - so I wanted to hide a small electronic circuit inside the light and expose two metal contacts - the on/off would be achived by shorting those contacts.

I wanted zero consumption because I aim to run an LED from directly from a Cr2032 battery - I can't let this take a constant drain as the battery will go flat quickly.

Once I had the latch, my next improvement was to get it to auto-turn off - but that'll be another question

I want to make the bulk of this with simple cheap components so that _anyone_ could make them - so if anyone can think of a cheap solution to this (with minimal components) I'd appreciate it.

 

[Reloadron]

Magnetic reed switch?

Ron

 

[TaDa]

That might be a refinement on the physical switch - but it would require the magnet to be left in place to achieve the latch - which might look a bit odd in a doll's world. Interesting thought though - I'll ponder on it

 

[KMoffett]

It always helps when you explain the details of what you're "really" trying to do. It's often easier to help you solve your problem, than solve your solution.
OK, 3V supply with push-on/push-off and time out. I have just the circuit...but...it involves an 8-pin DIP microcontroller (PICAXE), a couple of small MOSFETs, a couple of diodes, a few resistors...and programming. Think you would be interest?

Ken

 

[crutschow]

A simple on/off could be done with a CMOS FF, such as a 4013 with the D input connected to the /Q output, driving a bipolar transistor to light the LED should work. The latch and transistor would draw essentially no power when off. You would need to debounce the switch with an RC filter before going to the FF clock input.

Auto-off could be done by connecting the FF Reset input to the Q output through a resistor and capacitor to ground. The RC time-constant will determine how long the light would be on. To allow immediate turn-on of the light after the auto-off you could connect a diode across the reset resistor to rapidly discharge the cap after the reset.

 

[Reloadron]

That might be a refinement on the physical switch - but it would require the magnet to be left in place to achieve the latch - which might look a bit odd in a doll's world. Interesting thought though - I'll ponder on it

It's about creativity.

When using something like a small reed switch you hide it. Possible in a hollow dowel or any piece of wood. Just the tiny glass envelope and two tiny leads. Likewise a small magnet can be built into anything.

Ron

 

[ericgibbs]

hi tada
These slide switches are really tiny.
Look at this datasheet.

 

[TaDa]

hi tada
These slide switches are really tiny.
Look at this datasheet.

Yes - they look like the ones on the sylvanian lights themselves - but they're quite pricey - I'd be thinking more of making a switch out of a staple

I really wanted a simple(ish) and cheerful solution using a couple of transistors and a few cheap discrete components.

But all the ideas and thoughts are helpful - I'll find a way

 

[crutschow]

I believe my solution which uses a FF, a transistor, and a few resistors and caps is about as simple and cheap as you can get.

 

[KMoffett]

TaDa,

I think with you're specs:

3V coin cell supply
zero off-current
N.O. P.B. switch
toggle on/off
time off
simple
cheap

you've eliminated almost all possibilities. I would be delighted to see it happen, but it's a very large challenge.

Carl's solution looks the best.

Ken

 

[TaDa]

Ok, I'll show my lack of experience here....

Can the circuit (diagram attached) be coerced to do what I want?

To turn on, momentarily short X and Y
To turn off, momentarily short Y and Z

If not turned off, choose R and C such that T turns off after, say 30 minutes

'momentarily short' could be implemented with reed switches or, more crudely just exposing a couple of contacts and shorting them with a paperclip

Am I getting this all wrong?

 

[KMoffett]

With a logic-level MOSFET (instead of a transistor) and two 3V
batteries in series (6V) and a large capacitor and a high value
resistor in parallel with the cap, you could get a push-on/fade-off
circuit...no push/on-push-off. And I don't think you would get
anywhere near 30 minutes though...but, maybe without the parallel
resistor.

Ken

 

[crutschow]

In theory it would, but C would have to be very large to stay on for 30 minutes. Say the LEDs took 10mA. Using a gain of 20 would give base current of 0.5mA. This would require an R of 2kΩ for 0.5mA and a 1V drop. A 30s time constant would require a capacitor size of 30/2k = 15mF or 15,000µF.

Also as the capacitor discharges, the LED would get dimmer and dimmer, until it finally extinguishes.

It also requires two switches. You originally stated you wanted one.

 

[TaDa]

Ok, I see where my naiveté is - C will be HUGE.
And, yes, my later solution introduced a second 'switch' - I'm thinking and typing at the same time - apologies!

I see now why the flip-flop solution is so elegant - I have drawn the attached schematic of what was described to help me understand.

I 'borrowed' the circuit from elsewhere so I'm guessing the 0.1 capacitor is 0,1uF?

I guess the current drain input R requires is not too great but to get my 30 minutes delay I'll still need quite a large R1 and C1
(I'm thinking 10Mohm and 150uF - but I'm not too sure if these are realistic or if C1 would be physically too big)

So any recommendations on R1 and C1 (and D1 and T1) would be appreciated.

Maybe the timer part is just too much.

You'll all be glad to know (or at least not totally surprised) that I'm getting the feeling I am seriously over-egging this and just put up with the switch
(or at least not bother with the timer element)

 

[crutschow]

The 0.1 cap is 0.1µF.

Your circuit looks ok but you should add a small resistor in series with the LED to limit its current. If you build it, be sure to tie all unused inputs of the FF package to ground.

The circuit may be a little marginal with one battery and you may find you need two in series.

D1 and T1 can be just about any small signal diode and transistor.

Yes, 30 minutes is a very long time to do with an RC circuit. The leakage current of a large electrolytic capacitor can be a problem, which would make 10MΩ problematic as a charging resistor. Typically for such long periods you use a low frequency RC oscillator driving a digital counter. For example the 4060 has circuitry at the input to build the oscillator along with a 14-bit counter for long delays.

 

[TaDa]

Oh dear! Given the space limitation it looks like I'm going to have to use a switch then

And that's where we started folks!

I've learned a lot and it has focused my mind - Thanks for all your input!