Creating Sleep Mode in Battery Powered Consumer Device

[groundhog1]

Hello Everyone,

I have a product that will have an LCD display, and is controlled by microprocessor. It is powered by battery (possibly 9 volt). It will draw about 20mA at about 5 volts.

I am planning to use a switching DC-DC step down regulator. The problem is, that I don't want to use a hard on/off power switch. I was planning to have a software monitored push button, where the uP will be in sleep mode (power saving) and wake up upon activity at the button.

But if the uP is monitoring the button, doesn't the buck regulator have to be on all the time, wasting power? I have done this sleep mode before, but the uP is connected directly to the battery power.

Maybe there is a different way to implement this sleep mode function? Maybe I use the shutdown on the regulator and have everything off, then use something else to monitor the push button? Is there a button monitor IC that I can connect directly to the battery and it can activate the buck regulator upon button activity, thus turning everything on?

Thanks,
groundhog1

 

[IRstuff]

Power in CMOS devices is a function of frequency.

So if you slow down the processor, it consumes less power. During sleep mode, you could crank the frequency way down, which would allow you to duty-cycle control the regulator when the voltage drops too low.

TTFN

 

[Comcokid]

A true low quiescent current battery-powered sleep mode with voltage regulation ahead of the processor gets tricky.

As IRstuff suggested - You could implement a 'burp-mode' where the switch regulator is switched off and the processor runs from a capacitor. At short intervals, the processor wakes, turns on switch power to re-charge the cap, and then shuts down. However, the average sleep-mode current may remain relatively high unless the design is carefully thought out.

Another approach is to use a boost switching scheme. Use ~3 volts for the battery. Use a Schottky/inductor boost scheme that when OFF, the battery voltage passes through the regulator and you get sufficient Vcc to operate the processor. When ON, the regulator boosts the Vcc voltage to 5 volts.

Another approach is to use a very low quiescent current linear regulator. If you typical ON time is very short with extremely long sleep times, this can prove to be quick, simple and give long battery life. Use regulators such as the Analogic Tech AAT3220, AAT3221, or Seiko S-812C series which have operating currents of about 1uA.

 

[richs]

Hi!

How about this. Have a PNP transistor from the positive
9V lead to the supply of the regulator. The emitter
going to the battery, the collector going to the
regulator.

Have a momentary normally open power switch in parallel
with the collector of an NPN transistor. These are
both put in series with the base of the PNP transistor.

In operation, the user pushs the button, allowing current
to flow into the switcher by pulling the base of the
PNP transistor lower. The microprocessor, after
power up, holds down the PNP transistor with the
NPN which is "wire or'ed" with the pushbutton.

When the processor is finished with itself and wants to
power off, it lowers the base to the NPN, which will
cut off current to the base of the PNP which then
drops power. Of course, you will have a series resistor
between the base of the NPN transistor and the output
of the microprocessor (buffered if necessary) and a
series reisistor between the base of the PNP and the
collector of the NPN.

This should work regardless of the supply voltage assuming
the transistors break down voltages are not exceeded.
An old 2N2222 NPN and a 2N3906 PNP ought to do the trick.

Cheers,

Rich S.

 

[groundhog1]

Those are all very good suggestions!

I love the low quiescent current regulators. Looks like the Seiko can handle larger input voltages and would be better for this application.

I also am very intrigued with the BJT switching circuit. I think that is very inventive.

But, I can imagine the possibility of the uP doing funny things if the uP half way turns on, or if the person doesn't hold the "on" button long enough. It will be difficult to turn off in this state.

Also, on turn off, as the uP is powering down, you need to insure that the base of the BJT is nailed down low, long enough for the circuit to fully turn off. What if on turn off, the uP turn-off control line fluctuates?

Maybe a debounce type circuit..
I guess I should just try some of these out in the lab.

Thanks,
groundhog1