efficient way to disconnect battery when voltage drops?

[tautech]

The OP is on year 7 for these batteries, at least a majority of them; 1 is 10 year and a couple are less than 2 years...  Most of that longevity is a manual monitoring process.  My goal, before my health completely craps out, is to make the system smarter so that it is easier for my wife to maintain after I cannot.

Shouldn't be too difficult to cobble something together around this:
https://www.maximintegrated.com/en/products/power/supervisors-voltage-monitors-sequencers/ICL7665.html

[ja421]

Interesting chip -- thanks.  I ordered one and will experiment...

[tautech]

Interesting chip -- thanks.  I ordered one and will experiment...

Yeah, you've got Q and Q outputs but do try and find more than 1 datasheet for all the info on these.
You may be able to have enough drive for a small FET but look at using a gate driver too.
You'll need to give a regulated supply, something like an LM7805L should be OK.

I found getting the divider values just right a little tricky, some tweaks may be required. 
Pretty easy with a DMM or two and an adjustable PSU and something to monitor the output when it switches.

[ja421]

Great! Thank you for the tips.

[Codebird]

I think you might be over-engineering this. You don't need high precision for this - you just need something that will disconnect a load once the supply voltage goes under 11V or so, and turn it back on when it reaches 13V? Exact limits adjustable. It happens that I faced this exact problem myself some time ago, and still have the design I came up with. You can built it easily on stripboard and through-hole too, so cheap to build a one-off. https://birds-are-nice.me/ipfs/QmeCqXxspQN8Fa1dZ7W94nucyXA28qUcGbzc3c712PRHDa/

[ja421]

Thanks @codebird -- I ended up using an attiny85 (and most of the time it's in sleep mode); I did this as it was really easy to configure low & high voltages (via potentiometers) as well as build in hysteresis much easier than other solutions I came up with.
It is definitely not the most efficient, or elegant, method, but it seems to work.  Part of me keeps thinking "a processor will crash & hang" but so far it has proven to work well.

[Jeroen3]

You can buy battery protectors for marine batteries. They have similar situation where batteries are charged with solar energy and must not deep discharge.
Often the solar charger contains these features as well.

Stop DIY work, you're putting it outside. Which requires solid and reliable solution. Which isn't achieved on some protoboard in tupperware.

[Seekonk]

Go on ebay and search Low Voltage Disconnect.  You can get a nice ready made board for $10  with switch selectable pickup and dropout.

[Codebird]

Jeroen is right. Just about anyone here could put together a low voltage disconnect - microcontroller, or chip designed for such a purpose, or comparitors and latch as I suggested. They aren't hard to build, but if it's going to be outside then you will have to worry about weatherproofing. This applies even if you buy a ready-made board for the purpose - I can tell you from experience that even when sheltered in an outbuilding, unenclosed circuit boards do corrode. Stripboard goes fast, PCBs last longer, but all die in the end. You could build a suitable enclosure as well, with watertight seals and gaskets on all cables, but by that point you're reaching the level of expense and time investment where you're better off buying a ready-made solution, board and enclosure together.

The easiest way might be to just get a bog-standard PWM charge controller off Amazon. I wouldn't trust a low-cost charge controller for anything over perhaps 100W, a lot of them have very 'unconventional' ideas about proper battery management, but most of them do have a low-voltage disconnect and reset function.

But, if you really do want to build your own just for the satisfaction... you could just pot the whole thing in epoxy or silicone.

[ja421]

I had a $200 solar controller that was supposed to be "weather resistant"; it was inside a shed (insulated,dry, & heated); after 2 years, the components had corroded to the point that it shorted, literally blew up several components so that they were unrecognizable, and no support from vendor (schematics or repair) -- their solution was to "buy another".  So I don't hold that a "pre made" unit is somehow better than what I can come up with.  Having made a few things, and setting them in water proof project boxes & using liberal amounts of sealant where wires came thru, I am convinced that they last just as long.   I now coat all bare metal with some nail polish as well for anything that will be near the weather.

[Codebird]

It is inconsistent. I've got two pieces of stripboard in my garage - the garden light timer, and the charge controller. Both made in exactly the same way, and housed in the same room. The bell board is only slightly older, but seriously corroded - while the charge controller is entirely untouched by corrosion. Microclimate matters. But it's certainly a concern. Even if you buy an off-the-shelf unit, we all have horror stories of the 'quality' of low cost electronics.

Personally, I'd make it myself, because I know then that even if it does go wrong in a few months I can fix it and make it better. I'd also make sure to include a fuse though, just in case! This has actually been on my to-do list for a while: Making some test boards, applying different means of corrosion protection, and exposing them to the harshness of good British weather.

[ja421]

Electrolysis, I am sure, also plays a part in corrosion.   Different components may use different metals which could contribute.  It's not so much the 'off the shelf' part, it's the part that claims like "Weather resistant" should be able to handle more than a 2 year lifespan inside.  So making it myself gives me A) exactly what I want, B) helps me learn what is going on, C) gives me a lot more satisfaction, D) allows me to fix it if something goes wrong later.

On one project, I used the nail polish method & some shoe-goo to protect (or pot) all the components from vibration and it's installed on a motorcycle and has held up very well; it is not well protected, gets moisture, water splashes, rain, etc. and has held up rather well.  The wire at the far ends are the only thing that have given me a problem so far.... so I think for inside a shed or under the overhang where it never gets water on it, it will hold up, provided I don't run into an electrolysis issue.

[Red Squirrel]

Anything you buy or make you'll probably want the whole setup to be in some kind of outdoor/wet rated Nema box anyway.   The battery, charge controller, and any other controls.  May also want a small heater in the cabinet to keep it at a half decent temperature, just need to figure out the best balance as if you use the heater too much you're just using up the battery, but if you don't heat it at all the battery could freeze when it's low.   May even be worth insulating the box with some fire rated insulation. 

[ja421]

I use a small incandescent light bulb with a thermal switch to keep it from near freezing temps.  This has worked well.   The box it's in is a large ammo box that is sealed and insulated.  During the winter, I close the lid, during the summer I remove the lid & use a small (thermally controlled) 12v fan to keep things from over heating.  So on top of the individual bits (like the battery disconnect) being inside their own sealed project box, they are inside another, which is then inside a building under a roof protected from the raw elements...

Every little bit adds up; which is why I try to save as much power from every thing....  .8 amps vs 1 amp isn't much different, but when you look at a dozen things, that's over 2 amp savings...  same applies when I'm dealing with 100ma vs 80ma -- it all adds up.