Testing 12V batteries

[Oakmaster]

Every year at work they change the batteries for the back-up system in the Servers at work.  out of 20 or so I'll get 4 or 5 that are still good.

I check the voltage, and see if they will take a charge using my CETK Charger (which has a 'reconditioning' function) if they take and hold a charge then I load test them (usually with a 20W bulb) then re-check the voltages. i then put them on a 12V tyre pump and see if they will blow up a tyre (car tyre) 

With my 'current' batch (did you see what I did there?)  I got them down to ten number with a range of 12.8 to 13Volts, 4 passed easilly blowing up a tyre. the other 6 despite showing a good voltage just crashed when I did the bulb test (showed as 0.01V with the 20W bulb across the terminals)  I havnt encountered this before - advice and explanations would be very welcome.

They are quite small, not quite as big as a motorbike, golf cart, or a jump starter battery, but using 4 toghether you can use it as a jump starter.  Also very handy in the camper as the lesiure battery, individually with the 12V tyre pump (wheel barrows, mowers) and jumpstarting mowers.

below is one Ive bodged-up with copper pipe 'horns' for the jump leads, a 50A Anderson connector, and cigarette lighter socket.

(I'll admit I'm a bit 'tight' - and if I took my time into account - it would have been way cheaper to buy a jump start pack)

Thank you 

[David Sparkes]

Given your test results, I'd be suspicious about them actually performing as required in their Server UPS function.
I'd be asking how long they are expected to run the servers for.
Are they actually expected to run the servers, to maintain sevice, or are they just to provide the ability the power an automatic controlled shutdown, without losing data, immediately the mains goes off?

Dependant on the answers I'd be suggesting an in-service test is performed to test if they can actually provide the service required. Make sure a failed test doesn't leave you with a totally collapsed system.

Also consider that the in-service charging system isn't working very well (low voltage leading to insufficient charging).

Regards.

[FridgeFreezer]

What David said - if they're failing that badly they should've been swapped out much much sooner or the whole setup is worse than worthless.

[mickeyw]

I've always replaced UPS batteries on a 2 year schedule. They're meant to do more than that, but this is playing safe, or so I thought.

I've often been able to use the SLA units removed for camping purposes, and rarely had one that has actually gone bad.

[FridgeFreezer]

BT always replaced theirs on a conservative schedule, and if they ever actually got leaned on during an outage (as in, discharged beyond their rated capacity) they were changed out immediately. Universal emergency service obligation and unlimited public liability will focus the mind I guess

[Oakmaster]

 

Amongst  the very best things about this forum, is the requirement to craft your question so as to get your facts right

(you would never think that Rigorous Attention to Detail was my motto)

so that you get useful answers to the question you thought you had asked.

The VERY best thing is that you usually get some useful and insightful answers to members everyday queries.

BUT – you have to be very careful to set out your query so as to get the answer to the question you really wanted answered.

I really take on board (and am grateful for) all the responses about the inadequate regime we apparently have with our server UPS, but as it turns out, the ready supply of small 12V lead acids, are not any part of the server UPS, but are the residue from individual users UPS packs that are gradually being phased-out as we all transition to laptops.

So apologies for getting that wrong (today I mainly feasted on Humble Pie – and I don’t mean that great band of the 1970s)

What I’d really like to know, and should have been what I asked clearly in the first place is this:

Why? when I have weeded out all the batteries that wont hold a charge (only using a volt meter as the gauge after a re-charge and a couple of months ‘resting’) and only going forwards using the ones who seem to maintain their voltage do I have total failure when I apply even the lightest load to those who would have indicated they would be good?

Thank you, if you are still reading this.

[Escape]

It's not unusual for old batteries to maintain voltage, but fail when current needs to flow. The lead plates get damaged over time. You only need a little bit of material for no-load voltage, but a current draw will highlight the problem. The battery tester I use is basically a big resistor in parallel with a voltage gauge. You first measure no load voltage, if that's OK you flip a switch to add the resistor to the circuit and measure again while a large current is flowing. Often, especially if recently charged, the voltage will be OK until the load is added and then drop dramatically. Much like you are seeing. A lot of info on batteries on https://batteryuniversity.com/learn/article/how_to_restore_and_prolong_lead_acid_batteries

Filip

[TSD]

As Escape says, it's not an unusual failure, but I think it's seen more often now that AGM batteries are more common (as your UPS batteries will be). I think it's not that the AGM batteries are less reliable, but rather they work so well that the gradual degradation isn't noticed until the battery completely fails to deliver on it's job.

 

A while back the Ibex failed to start on first crank one evening. Cranked as normal, but didn't catch on the first few compressions as it usually does. On the second turn of the key it failed to turn over at all. After a bit of investigation I built a proper automated battery test rig, and was surprised to find the Optima battery could only deliver 14Ah, against the spec of 55Ah. Not surprising, it was a few years old, but I had no idea the battery was degraded at all - in normal use the tdi cranked over perfectly well and started in all weathers right up to the day it couldn't. Being an AGM battery with very low internal resistance, it could still easily deliver the required cranking power for a short burst, it just couldn't sustain the effort. My impression is that wet plate batteries often failed with rising internal resistance as one of the symptoms, so slow cranking and marginal starting were more obvious symptoms as the battery aged.

 

I tried a few of the recovery regimes that are out there, desulphation etc. as a lot of smart chargers implement. Over the course of a few weeks, I got the battery back to about 34Ah, then I threw it away. A new replacement Optima was close, but didn't quite meet it's advertised spec. Interestingly the 7 year old orignal battery in my D3 when I bought it was still delivering well over 90% of it's advertised capacity.

 

My UPS  runs a load test cycle regularly, and flags up if the batteries are failing to deliver a useable runtime. It also seems to flag the batteries as failed automatically after a couple of years, regardless of condition. I change them when their time is up and I reckon the failure rate in batteries I've taken out is around 50%, though I only remember once having to change them because the UPS failed the load test run.

 

[FridgeFreezer]

An example of TSD's battery test rig in action on the "house" AGM battery in the ambulance - I tested it when I fitted it, and I tested it again when I removed it - same 100Ah AGM battery in both graphs:

You can see that measuring the voltage with no load (the peaks) there's very little between the two, but one is holding about 100Ah of charge and the other is managing maybe 20Ah - same battery, just a few years down the line. You can see how much more it drops under load, and how it falls off a cliff pretty quickly. The red graph trucks on for ~3x longer just gradually going down until the tester cuts out at ~11.5v.

[Oakmaster]

3 hours ago, Escape said:

. A lot of info on batteries on https://batteryuniversity.com/learn/article/how_to_restore_and_prolong_lead_acid_batteries

This contains a lot of useful info - thanks