0

I'm looking for a device which monitors a 3.3V line. When the line drops below, say, 2.7V, I would like it to switch power over to a battery backup (3V lithium coin cell.) It would also be good if it provided a logic signal to alert of the supply failure so the MCU in my project can switch over to a low power mode. It needs to be able to switch up to 120mA when the supply is high. What is the name of this device? I have a feeling it's quite common in applications like laptop computers.

Ideally it would have 5 pins: Vin main, Vin backup, GND, Vout, Logic out.

starblue
  • 6,462
  • 2
  • 21
  • 35
Thomas O
  • 31,546
  • 57
  • 182
  • 320

5 Answers5

3

Power Management Controller / Power Management IC

The functionality your after would normally be part of a full system power management IC or a battery charging/management IC.

If all you need to implement is exactly what you asked for, a single FET with a low Drain <-> Source voltage drop and maybe a couple resistor to set the trip point are all thats required.

If you need to smoothly handle bounce, such as the bounce that occurs when you plug in a DC barrel jack, you can use a supervisor/reset controller designed to deal with bounce and use the reset output to trigger the FET.

Mark
  • 11,627
  • 1
  • 31
  • 38
  • That's a bit vague... I mean a voltage regulator is a power management IC. – Thomas O Nov 03 '10 at 19:50
  • +1 - That's what I would have called it. Digikey apparently agrees with Thomas with [27 'PMIC' categories](http://search.digikey.com/scripts/DkSearch/dksus.dll?lang=en&site=US&WT.z_homepage_link=hp_go_button&KeyWords=PMIC&x=0&y=0), but this one - [Power Supply Controllers](http://search.digikey.com/scripts/DkSearch/dksus.dll?Cat=2556707&k=PMIC) looks like it contains what he wants. – Kevin Vermeer Nov 03 '10 at 20:13
  • @ree: PMIC is an incredibly general term that encompasses virtually all ICs that are not strictly logic/digital or analog. Taking a brief look at the "Power Supply Controllers" they generally appear to be specialized logic chips that don't actually handle/pass any power. – Nick T Nov 03 '10 at 20:30
  • Yeah, that's my problem. They are sequencer IC's, not battery backup IC's. I'd hate to have to go to a multi chip solution as space is critical. – Thomas O Nov 03 '10 at 20:33
  • 1
    @Tho: What was the issue with having a diode from the battery to the supply line (no diode along the main supply)? Li/MnO2 cells have an open voltage of 3.2-3.3 V, so the battery diode shouldn't conduct normally. – Nick T Nov 03 '10 at 20:52
  • Well, I need to isolate my supplies. The switched voltage only needs to go to the MCU. There is also a DSP and LED on the board which would drain the battery pretty quick. – Thomas O Nov 03 '10 at 21:02
  • 1
    you need to use a multi-channel power management controller and just get it over with... – Mark Nov 03 '10 at 21:24
  • or pepper your board with diodes, like noted. – Kortuk Nov 04 '10 at 05:08
3

Power Supervisor IC

Toby Jaffey
  • 28,796
  • 19
  • 96
  • 150
2

It sounds like a Diode-OR'd power with the line voltage also going to a digital input pin to read the line voltage.

But there's probably a chip for that.

Robert Deml
  • 3,232
  • 6
  • 29
  • 27
  • As mentioned in my previous question that wouldn't be suitable because diodes drop too much voltage. I'm looking for an IC to do this. – Thomas O Nov 03 '10 at 20:00
  • @Thomas O: How much current are you drawing at peak and when in standby, and what are the voltages? – Nick T Nov 03 '10 at 20:02
  • Peak: 120mA, standby: 1µA, 3.3V. – Thomas O Nov 03 '10 at 20:05
  • Schottky diodes would only drop 0.3-0.4 V at 120 mA (36 mW), and less at lower currents. Will your IC do better? – endolith Nov 03 '10 at 21:41
  • http://www.falstad.com/circuit/#%24+1+5.0E-6+10.20027730826997+45+5.0+43%0Av+288+320+288+208+0+0+40.0+3.0+0.0+0.0+0.5%0A172+288+128+256+128+0+6+2.7+5.0+0.0+0.0+0.5+Voltage%0Ad+288+128+416+128+1+0.42%0Ar+416+208+416+320+0+25.0%0Aw+288+320+416+320+0%0Aw+416+208+416+128+0%0Ag+288+320+288+336+0%0Ad+288+208+416+208+1+0.42%0Ao+3+64+0+35+5.0+0.2+0+-1%0A – endolith Nov 03 '10 at 23:04
  • It needs to interface to 3.3V logic, so loosing .3V on the supply is not good. – Thomas O Nov 04 '10 at 13:52
  • A 220 Ω resistor will eliminate any ill effect from putting 3.3 V into a 3.0 V clamped chip. – Nick T Nov 04 '10 at 14:20
  • On *every* IO? It's a bit of a hack and doesn't allow for high speed signalling due to input capacitance. – Thomas O Nov 04 '10 at 15:52
1

Trying to replace OR-ing diodes? How about an OR-ing controller? Such devices are generally designed for much more current.

Maxim has some interesting battery backup supervisors.

Community
  • 1
Nick T
  • 12,360
  • 2
  • 44
  • 71
0

Instead of measuring the output of your 3.3V regulator to find out when it's crashing due to low battery input, you should probably be measuring the battery voltage so that your battery doesn't get overdischarged. Done properly you can switch from one battery to another as the input to your 3.3V regulator without its output fluctuating below spec. If switching to a smaller battery you'd probably want to reduce the load at the same time.

This question is basically the same: How do I design a device to automatically switch to the backup battery?

Community
  • 1
Matt B.
  • 3,805
  • 2
  • 24
  • 28