Undervoltage protection circuit

I'm looking to build a circuit to protect a battery from being discharged too deeply. I'd like to have the load disconnected whenever the battery's charge drops below 12V or so.

Is the circuit attached below the right way to go about this? (With the voltage drop across the zener and the LED at 12V and the resistor limiting current to 25mV or so.)

Any other advice/criticism/suggestions would be much appreciated as well.

I think you have the right idea, but you want something that will turn all the way off quickly and stay off. The transistor circuit will have a slushy area, where it might cycle on and off (which is probably worse). This is because when the battery load is removed the battery will increase in voltage.

How automated do you want this? You could put an automatic cutoff with a timer, for example, so the battery can build up a charge, or put a manual pushbutton reset.

Good points. Either of those solutions (push-button reset or time-out) would work for me. Could you point me to an appropriate circuit?

Ready made? Nope, though I'm sure they exist. My thought is a relay. I'll sketch something up, but it is likely one of the other guys will have better thoughts. What kind of current are we talking about here?

I appreciate your help. The load will never draw more than 5A. I was trying to avoid relays so as not to take the hit on current draw from the battery but I would certainly reconsider if someone has a good idea that involves one.

Does anyone have any experience with Maxim's battery monitor chips? Are they worth looking into for this project?

Not familiar with the chips, but they might be a good idea.

Here's what I came up with. The circuit will not reset if the battery isn't above the set voltage, and the LED will light if tripped. I kept everything very low current.



You'll need something better than that zener, a friend pointed out the TL431, which is programmable.

I've attached a list of zener diodes, but they run ±5%, which is pretty bad for this application.

That looks like it's just about what I need. Thanks! A couple questions:

Is there an easy explanation for why you need both Q1 and Q2? And when you say "better than a zener", is that just in terms of turning on quicker?

D'oh! I missed your explanation about the zener. Programmable does sound like a big improvement.

The tolerance sucks on a conventional zener. ±5% in a 11V zener translates to ±0.55 Volts, quite a bit for this application.

I made a set/reset flip flop using transistors. I thought about using a CMOS chip, CD4001, and decided against. Q1 and Q2 flip, they are always opposite states of the other, which means it is digital, either on/off, never analog or in between (which was what I was commenting on about your original design).

Let me know what you come up with. Won't hurt my feeling if you use something else.

I think I get it. Together, Q1 and Q2 form a signal buffer in to Q3 which switches the load.

If that's correct, Q1 and Q2 can be whatever little transistors I have lying around and I can use something really beefy but with a low Hfe for Q3. Does that sound right?

Yep. Right on target.

I'm much obliged. I'll follow up when my regulator shunts arrive and I can start playing with the circuit.

I was going to sleep last night when a problem occurred to me, we don't want Q2 collector resistor R4 interfering with the voltage sense circuitry. It's an easy fix, a low power diode, as shown.



Odd how the subconscious keeps chewing on a problem after its laid it to rest.

I'm not too comfortable with the voltage drop of the diode allowing Q1 to turn off, but I see several ways around that, such as a Schottky diode for example.

Yep; makes sense. I can easily envision the headaches the debugging would've caused. I appreciate the update.

PM me, I'll snail mail you a Schottky. They are cheap enough, but not available from the local Radio Shack.

That's a very kind offer; I appreciate it.

I'm having a little trouble seeing why a Schottky is necessary. Will adjusting the shunt regulator to account for the extra voltage drop not be sufficient to allow the use of a regular diode?

Transistors care about voltage. If the voltage across the base emitter is over .7 then they are current controled. With a conventional diode my thought is the base emitter is close to being on. I want to get that voltage firmly below the .6V threshhold to make sure it is off.

A regular diode might work the first time, or not. I want it to work the first time.