What did you screw up today?

[nickel-city-fab]

I have long harbored this fantasy of getting an old "commercial" UPS from a data center, the kind that requires its own rack. You can easily find 30kW units. And making a large, huge, Li-ion battery bank out of 18650 cells, for the whole house (since most of the data center UPS's output 230 single phase sine). They are not cheap. On an auto xfer switch.

UPS batteries are a consumable.
......
The makers seem to count on the grid being reliable. I suspect the design is only expected to have to operate a maximum of 5 or 6 times in it's "lifetime", so that IS the "lifetime". Naturally, when the UPS quits working well enough, most just pitch it and get another, rather than replacing batteries, if they even realize there are batteries inside. That fact is not lost on the makers either, who do, after all, sell UPS units, and would like to stay in business.

[J Tiers]

I have long harbored this fantasy of getting an old "commercial" UPS from a data center, the kind that requires its own rack. You can easily find 30kW units. And making a large, huge, Li-ion battery bank out of 18650 cells, for the whole house (since most of the data center UPS's output 230 single phase sine). They are not cheap. On an auto xfer switch.

You'd do fine with old fashioned lead-acid.

The fascination with Li cells is one I do not share. I'd rather have NiMH any day in most things. Not least because mass market Li cells are known to fail and burn, even these days. Not a lot of them, but why ANY? Because of their construction. There are types that do not, but of course they are more expensive....'nuff said.

The only thing that saves Li cells is the BMS. If NiCD and NiMH had BMS, they'd last forever and work very well. The lack of BMS makes it too easy to reverse charge a cell in the pack by overdischarge, which quickly ruins the pack.

[PStechPaul]

I have a Harbor Freight automotive booster/air compressor that is nearly 20 years old, usually kept on charge, and it still works adequately. I use it only a few times a year, sometimes to crank my truck engine, and more often to run the compressor to top off the tires on my car. It's a cheap unit, probably about $40 new, so I doubt that it has a very sophisticated charger circuit, and I would expect a UPS to have one at least as good or better. IMO, a UPS should be used only for as long as it takes to finish your immediate work on a computer before shutting down. I use a laptop so I can continue working for a few hours on battery, and that's usually enough to last until power is restored. Maybe once a year there might be a longer power outage, and for that I could use my generator.

Basically I'm saying that SLA batteries in a UPS should last at least 5-10 years, and perhaps it would be wise to analyze the parameters of the charging circuit in your UPS, and if it does overcharge the batteries, replace it with a good float charger.

[nickel-city-fab]

In which case, the old standby, Trojan T-105 batteries. Most often found in floor scrubbing machines, and fork trucks. Available in 6v and 12v styles, with some huge amperage ratings (I forgot the details). Lately I've been looking at the new batteries at work: 8D-1400.

You'd do fine with old fashioned lead-acid.

[Joel]

I have long harbored this fantasy of getting an old "commercial" UPS from a data center, the kind that requires its own rack.

I just used four $17 lawn mower batteries on mine. Got the UPS off Craigslist for cheap.
I built it so that I didn't have to get out the generator except for long outages (to keep the fridge and a couple of other things running). Works great.

[danlb]

Basically I'm saying that SLA batteries in a UPS should last at least 5-10 years, and perhaps it would be wise to analyze the parameters of the charging circuit in your UPS, and if it does overcharge the batteries, replace it with a good float charger.

I've worked with both home and commercial grade (Liebert, APC and others) and can tell you that the SLA in a typical home UPS just does not last. Units from APC, Cyberpower, HP, Triplite and others typically run the battery all the way down when there is an outage. They run it down to the point where the battery is damaged. To make matters worse, most of them will "test" the battery by switching the load to the battery and measuring the voltage drop. If the battery is bad everything crashes.

I mark my UPSes with the date of the battery change and replace them on a two year schedule. It makes no sense to hope that the batteries are still good at 5 years when it means system crashes if the batteries are weak.

Dan

[J Tiers]

YES.... 100% correct. Made cheap , intended to work only a few times before crapping, if that.

The charging circuits are no different than the power circuits (and may use some of the same parts).... cheap, and barely better than the ones in old battery line trimmers etc.

The gel cells need a different charger than a wet cell battery.

[Joel]

They run it down to the point where the battery is damaged.

As I am sure you are aware, one can go to the advanced power settings in Windows and have the computer sleep, hibernate or shut off, at whatever percentage desired. I think I have mine set to hibernate at 40 or 50% (of battery remaining). You can set when it notifies you, etc. - lots of handy settings are to be found there.

[J Tiers]

As I am sure you are aware, one can go to the advanced power settings in Windows and have the computer sleep, hibernate or shut off, at whatever percentage desired. I think I have mine set to hibernate at 40 or 50% (of battery remaining). You can set when it notifies you, etc. - lots of handy settings are to be found there.

Unless it is monitoring the UPS, the computer will not be aware of the battery state of charge. The assumption here is that we are using a non-battery (non-laptop) PC, which depends on having continuous AC input.

Otherwise, why even have a UPS?

[Joel]

Unless it is monitoring the UPS, the computer will not be aware of the battery state of charge.

Of course.
It is one cable, and it comes with the UPS. Windows handles the rest, and does a nice job of it,

The assumption here is that we are using a non-battery (non-laptop) PC, which depends on having continuous AC input.
Otherwise, why even have a UPS?

Assumption?
I was specifically referring to my tower computer and external Trip Lite UPS's.

[danlb]

Yes, Windows can do graceful shutdowns. So can Linux. I run a small personal server farm, as well as multiple Tivos. The Tivos don't have a great way to shut them down, though I made up a script that will telnet to the tivo and send the appropriate remote control button press events for sleep mode when the power fails.

The problem with all of this is that the consumer grade UPS is likely to report 100% charge even when heavily sulfated. That results in a very sort run time, and the computers crashing as they are trying to shut down.

My solution to the problem is to use two UPSes in series. The first is a pure sine wave which feeds the modifed sine that the computers plug into. When the power fails on either one, the computer gets the "running on battery" signal and starts the shutdown procedure. The network equipment is left in an on state since they pull fairly little power.

Dan

[J Tiers]

Of course.
It is one cable, and it comes with the UPS. Windows handles the rest, and does a nice job of it,


Assumption?
I was specifically referring to my tower computer and external Trip Lite UPS's.

None of the ones I have had were interconnected with the computer, so that never came up. I prefer to decide what to do, and that may not be just closing all the programs and shutting down. Some programs may need specific things done before shutdown, easily accomplished in say 5 min, but windoze may not do what is needed.

For instance, completing something that is being done with a remote computer may not go well if just shut down mid process, where a couple minutes of human intervention may do fine. Most UPS' with a good battery will last 5 minutes if properly sized.

And, as Dan says, if the battery comes up to voltage, even of it is after 1 minute of charge, the UPS decides all is well. Of course it may shut off in 10 seconds of actual operation......... Even a good connection with the computer and windoze probably will not work well with that situation.

[Joel]

It has been 2 decades since I had one that didn't. I have 5 in current use, plus the big guy posted above.
I only mention it in case someone wasn't aware that there were alternatives to letting a UPS run its batteries flat (and decimating their lifespan).

[Cenedd]

I have a compound screw-up for you. I've got one of those 4-sided (cheap) diamond plate sharpeners that sit in a box/holder. I've previously used it with oil and enough had accumulated in the bottom that the lowest face gets covered - which sort of works out quite well. Anyway, that was on a shelf and somehow got knocked over without me noticing. So there's thick brown oil over the shelf, down the side, between the shelf and the supports, dripping on things below That's one.
In starting to clean up, also now on my trousers. That's two.
Went in, emptied my wallet, keys etc out of my pocket, changed, and threw them in the washing machine in the hope it would come out if I was quick. Went back to cleaning up. Hmmm, my phone isn't in my pocket....oh now, I couldn't have....yup! *sigh* Not only that but it came out the pocket and cracked the glass as it repeatedly impacted round the drum. It STARTED as IP-rated!
Turns out that the heated bed of a 3D printer is good for loosening the glue that holds the rear cover on so you can get it apart to get the water dried out. I've nearly got everything functional again

[Cenedd]

Turns out, I'm not done yet! Got the last of the condensation dried out and stuck the back on again. Not quite got it perfectly aligned so I heated it again to take the back off and try again. The adhesive let go suddenly at one end and not the other and the back plate shattered. Back to the 2014 phone until parts arrive I guess. (Yes, I'm painfully aware this is a 1st world problem)