COTS Battery Backup

rickman wrote on 8/14/2017 5:26 PM:

Here's a link to the specs. APC even has a load vs. run time graph.

The 110V input voltage may be an issue, though!

I see it claims 70 watts out and a little under 36 minutes run-time at that load, i.e. just 45 watt-hours capacity. From the quoted size I see that it might accommodate a motorcycle size accumulator, which are typically 7.5 watt-hours capacity for a 12v SLA. You can get 25 watt- hours from a similar-sized LiFePO4 cell, but they certainly aren't fitting those: they retail for $US 260, but they only want 40 bucks - unless they sell them without a battery fitted.

Actually, the problem is both input and output. I forgot that it would require connector adapters on the outputs for line powered devices. I wonder if they have a 220 volt equivalent? They designed the electronics to adapt to the line frequency in use.

UPS typically use SLA, sealed lead acid, batteries, I suppose because of the cost. Notice it is very hard to get capacity info on most UPS units. They are happy to tell you how many amps or VA they will provide (less often watts), but almost never a WHr rating.

Regular Li batteries would be good enough and much smaller or more energy or some of both. But the cost would rise. But then a replacement SLA is not cheap either. The one from APC is as much as the unit and a Power Sonic battery is still $35! They can last a long time though. I've had one for well over a decade and it's still keeping my devices running through a power failure. Usually my ISP goes out and I can see my modem not being able to get to the rest of the world. Lol!

I wrote them a note letting them know the lack of a 3 pin connector is a problem for me and that they should consider making a unit specifically for the rPi with an external 5 volt PS and physically smaller. I don't expect them to follow up on that though since their core competence is SLA batteries and that would almost demand a Li battery.

there will be an equivalent uk/eu version around somewhere

Couldn't see an equivalent on the UK section of the APC site.

Tell me about it! I ran into that gotcha when shopping for my UPS a few years back. Fortunately I knew about a decent supplier, who I'd been buying SLAs from for a long while: one with real people who know the products and who answer their phones.

They sell Yuasa NP7-12 SLAs for $US 21.60 - quite a reasonable price.

Yuasa is the most consistently reliable SLA type I've found. I use them for starting model engines and as in-flight power in my sailplane as well as in my Riello UPS, though as I haven't replaced those yet, I'm uncertain of the model number since they were pre-installed.

If I was feeling keen and was planning to do anything with my RPi that needed a UPS, I'd just plug a wall rat into mine. My ADSL modem isn't on it as yet: that would take precedence as IIRC BT exchanges have been on UPS systems for years.

On Mon, 14 Aug 2017 21:14:18 -0400, rickman declaimed the following:

Lead acid typically works with cheaper charging circuits, and work well with a continuous float charge -- unlike NiCad/NiMH/Lithium rechargeable with their multistage charging cycles and warnings to remove them from the charger when they are "full". .

Just looked at the photos on APC website ... it's the first UPS I've seen that's light enough to have 'keyholes' in the case to allow wall mounting :-)

I didn't mean "I can't see any UPSes on the UK APC website" I meant "I can't see an equivalent model to that one" :-)

NiCd worked perfectly for this sort of usage. You just had to keep the charge rate down. At a 1% rate or thereabouts NiCd batteries could be left on charge for years at a time: perfect for UPS-type operations.

I used to use 50mAh cells made up to 4 or 5 cell batteries them to run timers on models. The batteries were permanently on in the model (the timer drew 300uA except when it tripped its solenoid (500mA for 15mS) and when at home the charger was always on and plugged into the models. It was continuous current design set at 900uA: take off the 300uA the timer used and the remainder is near as dammit a 1% charge rate.

AFAIK NiCd is the only chemistry that will work this way, so its sorely missed here.

NiMH has a rather too high self-discharge rate for this charge regime to work well with it.

Those times are over. Eneloop's discharge is so low, they can be sold "ready to use" with a decent shelf life, something no NiCd ever managed. According to specs they should do just as well with a continuous trickle charge. Whether it's true, who knows?

Yep, Eneloops are excellent. They're about the only rechargeable that my Pentax K-100 DSLR doesn't destroy in under 5 charges. Too bad there aren't any really small versions, ideally the same size as the Sanyo

I was all set to buy one until I realized the AC outlets didn't have ground pins. That just narrows the focus too much. Directly providing 5 VDC is pretty handy, but some time ago I bought a power pack that has six USB connectors for charging and powering things like the rPi. Unless I end up with more than one location, I don't need those outlets as much.

Don't modern charging systems for Lithium switch off after full charge is reached - this is my understanding for devices like phones and tablets but I guess may not be the case for UPS.

Richard

Indeed so. A UPS is a near perfect fit, as are emergency lighting systems, fire alarms etc.

But are they bright enough to go on watching a fully charged Li battery and to top it up when its self-discharged by 5-10%?

That is what you need for a UPS or an emergency power system.

As others have said, that is easy for lead-acid or NiCd, because you can use a low trickle charge rate and leave it on for years without harming the battery.

Something similar may work for NiMH. I've never seen anything to say whether it does or not. All I know is that the one device I own that uses NiMH and has a built-in charger does not have a trickle charge mode or fully charged detection. OTOH a multi-chemistry charger I use for battery testing and cycling works well for lead-acid and NiCd, but its capacity measurement doesn't seem to work as well with NiMH.