How to design efficient electronic switch

Hi,

I want to operate a switch that will transfer a 5.5V and 500ma while the on/off voltage is about 0.5V. The main design concern is the voltage/power drop on the switch. I tried to do it with a mosfet but the gate source voltage was to high (2.1V) and also the voltage drop on the mosfet was too high (1.5V) resulting in a dimmed leds.

I would appreciate any help to design a more efficient electronic switch.

Thanks.

Run your 0.5V into a comparator, like LM358, with the other input set for 0.25V. Have the output of that drive a Low Rds MOSFET, like AOI518 from digikey.com, to sink current from the LED cathode.

Thanks CrossRoads.

I Think i didn't explained my self cry well. I need this 0.5V signal to power the switch that will power the entire circuit. there will be no power prior to the 0.5V operating the switch.

With your solution i need to power the LM358 all the time. Can you think of way/mosfet that can perform like my description? I know there's a mosfet with 0.25V on Vgs but how do I calculate the voltage/power drop on it?? My last mosfet is the 2N700 and it dropped 2.0V on it with load... (and 1.5V without load)

Thanks.

you can't really power anything with 0.5v with today's silicon semiconductor technologies...

so you need a "external amplifier" as a driver to drive your circuit, like a comparator (which requires external power) or just a simple transistor (which requres external power, too).

maybe you can find a low Vgs and low Rds mosfet, to minimize the voltage drop? (like 0.5v compared to 1.5v)

if you don't care about the color issue (LEDs change color slightly when different voltage applied), try a smaller current limiting resistor (well if you have one), or just simply parallel 2 or even more FETs.

Is the 5.5V on and available to power anything? Or is that only available to be switched on?

If I remember correctly, germanium transistors turn on with about .5 volts. There are still some available (expensive) . Selection is very limited and I don't know if you can get any that will handle that much current.

You will need quite a lot of base emitter current. The germanium I worked with did not have very high current gain.

pegwatcher: If I remember correctly, germanium transistors turn on with about .5 volts. There are still some available (expensive) . Selection is very limited and I don't know if you can get any that will handle that much current.

You will need quite a lot of base emitter current. The germanium I worked with did not have very high current gain.

Germanium's about 0.2V, you can even run oscillators at about 0.1V with a pair of them I think I read somewhere!

Hard to find these days though. And I think the leakage current is substantial and very temperature sensitive.

There are micropower comparators that draw very little current from the supply...

I was thinking that .1 to .2 V was VCE Sat. Some leakage, yes. Temperature sensitive, yes, and NTC. Silicon was a great step forward.

@CrossRoads - The 5.5V is only available after the switch is opened. May be I can use the paralleled FET's to be the main switch of a soft latch circuit?

That way i can use the 0.5V to open them enough for the 5.5V to open the main FET switch. But the problem remains, how do i minimise the voltage drop??

Right now the 2N700 is dropping 1.5V with no load and 2V with load (less on) that's a huge waist of power...

Hi, where is the 0.5V coming from that you want to use to switch on the load?

Tom...... :)

giliep: @CrossRoads - The 5.5V is only available after the switch is opened.

What is the maximum time the switch will be held closed for? The 5.5V supply could be used to charge a capacitor when the switch is open which would power the comparator/MOSFET circuit when the switch is closed. Only a few uA is needed to power a micro-power comparator and MOSFET gate you see...

@Tom - it's coming from an old diskman. I want it to make the leds on when i hit play

@MarkT - the switch can be held close for a long time. hours, days...

Hi.

@Tom - it's coming from an old diskman. I want it to make the leds on when i hit play

Is it coming from the headphone socket?

That will be audio/AC and it will be more than 0.5V, how and where did you measure the 0.5V?

Tom..... :)

TomGeorge: Hi.

@Tom - it's coming from an old diskman. I want it to make the leds on when i hit play

Is it coming from the headphone socket?

That will be audio/AC and it will be more than 0.5V, how and where did you measure the 0.5V?

Tom..... :)

Rectify it, store it in a small .1uf? Or so cap then use that pulse to latch your circuit (via feedback) or a flip flop, but I think I came across a circuit ages ago that simply powers on your device when it sees a signal...

And again why can't you use the 5v source? Amplifying the signal would be easy.

Why on earth didn't you say from the beginning the control signal was headphone audio???

"on/off voltage of about 0.5V" strongly implies a DC logic signal

Now we know its an audio power signal, how about putting it through a small audio transformer to boost to 5V or so, rectify to charge a capacitor, use that voltage to drive the gate(*) of a MOSFET, possibly using a comparator to get rapid switching.

(*) protect with resistor/zener divider to prevent overvoltage.

Hi,

@MarkT - Sorry about that. I thought since i measured 0.5V with a voltmeter that would be the important thing. I can't use a transformer since i don't wan't to use external power prior to the play button being hit. It suppose to be an outdoor activation so i need to use as less power as i can. Does this transformer can work solely on passive components?

@cjdelphi - I can't amplify it cause i want to save battery, so i can use the longer for the leds. Other wise the leds will be constantly on and the battery will dead in no time.

Using a comparitor would be ok, they require a little current.

But using a transformer from the ac source to boost the voltage should work.

Its all rather unwieldy though, a far better approach is to have a separate power source so the switch can be straightforward and simple - but if there isn't separate power its better than nothing.

I am reminded of the coherer, an ancient but clever device used to detect signals in telegraphy and early wireless - it sort of might do this task almost(!) http://en.wikipedia.org/wiki/Coherer

Hi

I am [u]reminded[/u] of the coherer,

Careful showing your age.

Tom........ :)

The shock and awe of electricity, a bbc doc.

The coherer is covered quite well, the first version worked fine in a colder climate but in india the coherer would rust and stick together.

http://youtu.be/uZY3L50VNps

All 3 parts are well worth watching if nothing more than to brush up on your electronics history... it left me with a strong hate for Marconi though and how patents ruined science forever.