Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Dead battery indicator with minimal parts

  1. Jun 13, 2005 #1
    I need to make a battery monitor with minimal parts.

    A green LED should be lit during normal operation.

    The device has two batteries (designed for 9V, but it could probably use others) in a split supply (-9, 0, +9). I need to monitor BOTH batteries, and have the green light go out and a red LED come on when ONE of the batteries drops below a certain level.

    Because digikey makes you buy small parts in bunches of 10, it would be really great if I could use the 2 or 3 transistors left over to build this thing with. It seems like a really clever person could figure out a circuit with only two transistors and 5 or 6 resistors, but I am not that person. I'm more of an op-amp kind of guy. I understand the basics, but I am not too good at predicting complete transistor behavior, as evidenced by burns on my thumb and forefinger from a hot transistor earlier today.

    I was thinking something based off of OK/Low battery indicator, but when I tried it, it didn't work. Apparently a small enough battery means there is not enough voltage to light both the red and green at the same time, so only the red lights? That's not how it worked when I did it, though. Both lit. I don't understand.

    Any ideas?

    And when I say "minimal parts", I mean BJTs, resistors, and the LEDs.
  2. jcsd
  3. Jun 14, 2005 #2


    User Avatar

    Staff: Mentor

    I think that you should be able to do the + side monitor with 3 NPN transistors, the 2 LEDs and about 7 resistors, as long as the ambient temperature doesn't vary much. The - side monitor circuit would be the same, but would use 3 PNP transistors in a complementary fashion.

    Make the voltage "reference" with an NPN transistor with a B-E resistor and a C-B resistor, where the ratio of these resistors gives you an overall multiplication on Vbe. Like, if both resistors are equal, then the Vce will equal 2xVbe (where Vbe is around 0.7V, depending on temperature and Ic). Feed this voltage reference with a resistor from your +9V supply, and set the overall Vce to be a reference voltage into the next part of the circuit.

    The next part of the circuit will be like a cross-coupled differential NPN pair, where each of the two transistors in the pair has an LED coming down to its collector from the +9V supply, with a series resistor right before each collector. The emitters of the two NPNs then go through resistors to ground. The base of the right NPN is connected to the collector of the left NPN, and the base of the left NPN is coming from the voltage reference circuit. You then tune up the resistor values so that the left Green LED NPN stays on until the +9V supply drops too much to supply the voltage reference properly, and then the left NPN shuts off, which turns on the right (Red LED) NPN.

    I only sketched the circuit on a Post-In note, so it could be missing some resistors here and there, but it looks like it basically will work. And I agree with you about that other circuit that you linked to -- it will turn on both the red and green LEDs after the trip point is reached. -Mike-

    Edit -- Better is to hook the voltage reference collector directly to +9V, and connect the emitter to GND through a resistor. Then you get a constant drop from 9V, and the emitter voltage is what you use to drive the base of the left diff pair NPN. Yeah, that makes more sense.

    Edit2 -- Oh, and a feedback resistor connection from the collector of the right NPN back into the base circuit of the left NPN. So that when the right/Red side does come on, it helps to hold off the left/green side. You can adjust all the resistors to give you some amount of hysteresis on the trip point.
    Last edited: Jun 14, 2005
  4. Jun 14, 2005 #3
    Hmm.. It really needs all that?

    Here is a very halfbaked version of what I am thinking. Yes, I know this particular circuit would never work. I want it to be crude but I think I'm overcrudifying. Just help me think it through...

    The orangeish sections are the two voltage sensors. When the voltage drops below a certain threshold, the base voltage, because of the divider, drops below the 0.6 V turnon of the transistor, turning it off.

    The blue section is supposed to be an OR gate, but wouldn't really work as shown. It would have to be the opposite, so that one of the voltage sensing transistors turning on would indicate a low battery. Otherwise it's a NOR instead of an OR. Also the voltages wouldn't work because one is driven by the other supply.

    Could Q2's collector go to V+ instead of ground? Then when the transistors turned on (however that would magically work), one of them would pull the OR gate down to either ground or V-. Either would work, so it would be a crude but functional OR gate.

    Then the red section turns on the red LED, of course, and that pulls the input to Q4 low, turning off the green.

    Did I at least get that stage correct?

    It seems like you could collapse the green section into the first section, though, making it three transistors. Q1 and Q2 become two alternate paths for the green LED current, which... oh wait that wouldn't work either.

    Is there any way to provide the opposite functionality of Q1 and Q2? (With PNPs perhaps)? In other words, it would have a pull-down resistor so that it could OR and the transistors would turn ON when the base voltage dropped too low.

    Does this inspire any ideas?

    Another is to have the green and red connected back to back and antiparallel and one side connected to ground. When the other side is held high, the one LED lights, when you pull it below ground, the other lights.

    Attached Files:

  5. Jun 15, 2005 #4


    User Avatar

    Staff: Mentor

    >Hmm.. It really needs all that?

    Actually, the circuit you show in your attachment is very similar to the one on my Post-It. Except I had one for the + battery and one for the - battery. Do you want just one set of LEDs for both batteries?
  6. Jun 15, 2005 #5
    Ahh.. yes I only want one green LED and one red LED. The green will light when both batteries are full and power to the circuit is connected, the red LED will light when either battery drops below a certain level. Can you scan your post-it or sketch a picture? You can draw a schematic like I did with Klunky schematic editor, screenshot, save, and upload.
  7. Jun 17, 2005 #6


    User Avatar

    Staff: Mentor

    Hey, that's a cool link. I wondered how you made your drawing so easily.

    I'll just throw it into Spice so I can figure out ballpark resistor values, then post a screenshot. I probably won't get to it until Monday, though. PM me if I forget. My version right now is for the positive side only. I'll see if I can think of a simple way to OR in the negative side too....
  8. Jun 17, 2005 #7


    User Avatar

    Staff: Mentor

    Okay, here's the quick version -- no optimization or anything. My first swag at resistor values worked well enough that I didn't change anything. The trip points are about 10V on the way up (for Vbatt = V1 supply) and 7V on the way down. You can tune the resistors to set the trip points and hysteresis wherever you want for your particular battery. Again, keep in mind that the trip points will move with temperature, so depending on how sloppy you can let your application be, you might want to add some temperature compensation parts (but that means you'll need to start adding at least a zener diode, for example).

    In the screenshot, the supply is at 10V, which is just enough to keep the left green led on as the supply decreases. When it gets down to about 7V, the left green LED snaps off, and the right green LED snaps on. Then as you raise the supply back up, the LEDs snap back the other way when you cross up through about 10V, and stay that way up to 12V.

    Hope this helps. I didn't spend any time thinking about the negative battery stuff. You can either just complement this circuit for PNPs and add 2 more LEDs to get the separate indication for the negative battery, or just complement the voltage reference (the leftmost NPN) to a PNP version down in the negative battery stuff, and figure out how to use that to OR into the control stuff for the NPN differential pair. -Mike-

    EDIT -- Hmmm, the Upload feature didn't seem to work for me. Is there a trick to it?

    EDIT2 -- My file was >100KB. Now it's small enough to upload. Hope this helps! -Mike-

    Attached Files:

    Last edited: Jun 17, 2005
  9. Nov 12, 2007 #8
    I think you can use a transistor as the key to protect your circuit
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook