Low Voltage Cutoff Circuit

[mc172]

Yeah, another one.

I need a low voltage cutoff circuit that you can enable after the battery has been connected and latches off once the threshold has been crossed. Then, draws absolutely minimal current as it may be a few months before I can get to the thing to take the battery out. The battery will be charged on its own.

I've attached two slightly different schematics of something I'm messing about with.

I have built both and the first one (sch0) works as I'd expect if C3 is fitted, not if it isn't. The second, sch1, well I'll get on to that.

It's a pretty straightforward comparator circuit but I've moved the "measure" or "sense" element to the other side of the output FET, which works as a latch. The idea is that once the FET turns off, the sensed voltage is hopefully very low, which keeps the FET off. This does work quite nicely on the real thing.

The reed switch is the "enable" device. Without C3 fitted, the FET conducts as soon as you connect J1 (the input) to the power supply, whether the reed switch contacts are closed or not. I haven't had a chance to play around with values of C3 yet.

Well what I'd really like to do is move everything to the right of the MOSFET apart from R6. I tried this in "sch1" but now the thing really is always on as soon as you connect the battery. The comparator has a push/pull output stage - not sure an open collector stage will help me as I have mulled it over and have come to the conclusion that it's the low side doing it.

I'm going to get this on an oscilloscope tomorrow to try and capture what's going on at startup but I thought I'd ask here if anyone has tried this before.

A thought I had whilst driving home is that if the FET even starts conducting slightly, it will start doing a comparison and then the output of the comparator , which is trying to drive high relative to the low supply voltage, will end up pulling Q1 down.

Any ideas would be great - someone must have done this before... Thanks in advance.

[Ian.M]

The TLC3702 comparator has a push-pull output,  so when powered down its output will be dragged down to near 0V via the output stage upper MOSFET's body diode to Vdd, turning on the MOSFET Q1.  To avoid that you'd need either a comparator that is true open collector or open drain with no diode to its positive rail, or add an external transistor or MOSFET to drive Q1's gate.  The disadvantage would be slow turnoff as only the 1Meg gate pullup can discharge the gate.

I suspect you are using a three cell LiPO pack, so the approx 60uA worst case discharge current if you leave the comparator and reference powered is highly undesirable.   That's about 430mAH/month which will kill the battery unless you set the cutoff point well above 3.0V/cell.

[mc172]

Thanks for your reply, Ian.

Aha, I had never considered that. Clever.

Do you by chance know of any low power, true open drain/collector comparators?

The separate transistor solution is quite an easy thing to bodge on and a great suggestion so I'll try that tomorrow and report back.

Cheers.

[stmdude]

I'm designing something similar right now actually.

The (cheating) way I approached it was to use a voltage detector from the NCP300 series that controls a low RDSon MOSFET (as I'll be pulling a bit of power through it).
I haven't built and/or tested it yet, but it's a pretty simple circuit.

Depending on your load, you may want to put an RC filter in front of it, as to not trigger from any sudden power-spikes.

[nfmax]

I designed a similar circuit myself, to protect a 12V lead acid battery. Details at https://www.eevblog.com/forum/projects/battery-undervoltage-lockout/msg1060148/#msg1060148. It might give you ideas, maybe.

[Marco]

Is the complexity with the reset switch really necessary? With a ltc1540 you can have the low voltage cut out continually running with submicro-ampere current.

[mc172]

Well, I put an N-channel MOSFET in, swapped pins 2 and 3 and problem solved. Cheers.