COTS Battery Backup

one of

mple

hermal*

I'd be more than happy to, only that was a long time ago, and that along wi th many other documents of the time have long since disappeared.

From memory there wasn't much to it. An op-amp driving a power transistor a s an emitter-follower, then a diode to protects against the battery feeding back when the power was off. Control was from a high resistance feedback path fr om the battery terminals (so negligable drain on the battery) and somewhere in that lot was the thermistor and a preset pot.

Current limiting was crude - just a resistor in series with the transistor collector.

Not a lot of help I'm afraid :(

Hmmm, thinking a bit more. The thermistor and preset weren't directly in the feedback but in the potential divider producing the reference voltage.

Actually it's remarkable what you remember once you actually start thinking about things. One resister from the battery + to the inverting pin of the op-amp. Something like 470k I think, and the same from there to the common ground. A resistor from a zenner reference to the non-inverting pin of the op-amp, then resistor - preset - thermistor from there to ground. No idea what the values were, nor what specific thermistor it was

Folderol wrote on 8/29/2017 1:33 PM:

I believe I saw a modern copy of that circuit. Instead of the opamp, pot and many other components it has an MCU. Otherwise identical...

Fascinating.

Had they also swapped out the current limiting resistor for a light bulb, like I did? A simple 'device' that not only provides protection but gives you an immediate idea of what sort of current is passing, lighting up really brightly if there's a short cicuit.

If not, presumably you could just cut a hole in the right place and attach one.

I've had a small desktop APC one go like that, there was no way the batteries were coming out without completely trashing the case. Even if there had been I wasn't about to subject another set of batteries to circuitry that did that.