Sealed lead acid battery maintenance?
in [Design]
|
Prev: robots
Next: C-multiplier again + LTspice note
From: John Tserkezis on 29 May 2010 18:49 Robert Baer wrote: > As a "point" of reference, i have an old UPS by APC and replaced the > batery only once so far in the appx 15 years of "use". > Most of the time it just sits there - maybe on trickle charge, who knows. > But on rare occasion when the power goes out and i am online, it keeps > all running for a careful save and shutdown. :-) Just because you *can* do that, doesn't mean you *should* do that. The recommendation is two years, to ensure the battery doesn't magically fail soon afterwards. In fact, in those conditions, they're probably almost guaranteed to last double or more of that. Problem is, "almost" doesn't cut it in a mission critical application, thus the two year replacement recommendation. This is why some UPSs have an automatic scheduled "test", where they shut down the server but leave it powered on, go to battery backup, ensure it lasts pre-determined amount of time and power up again. Otherwise you have no way of knowing a battery has failed. I should qualify my quoted constraints assume normal domestic SLAs. Higher grade UPSs use higher grade batteries that can last many years longer before the recommended replacement.
From: Martin Brown on 30 May 2010 06:29 On 29/05/2010 23:49, John Tserkezis wrote: > Robert Baer wrote: > >> As a "point" of reference, i have an old UPS by APC and replaced the >> batery only once so far in the appx 15 years of "use". >> Most of the time it just sits there - maybe on trickle charge, who knows. >> But on rare occasion when the power goes out and i am online, it keeps >> all running for a careful save and shutdown. > > :-) > > Just because you *can* do that, doesn't mean you *should* do that. And it is at least worth the occasional visual inspection of the batteries for obvious signs of swelling and distortion of the case. > The recommendation is two years, to ensure the battery doesn't > magically fail soon afterwards. In fact, in those conditions, they're > probably almost guaranteed to last double or more of that. > > Problem is, "almost" doesn't cut it in a mission critical application, > thus the two year replacement recommendation. UK replacement is annual test and replace on fail or 5 years old whichever happens sooner. Mine live in a fairly cool environment and would still hold a decent charge when retired. I have seen some 8 year old ex UPS cells that I took to the recycle centre in a bucket because I was unsure if the plastic cases would hold out. They were very obviously mangled shapes with big bulges but still not leaking. > This is why some UPSs have an automatic scheduled "test", where they > shut down the server but leave it powered on, go to battery backup, > ensure it lasts pre-determined amount of time and power up again. > Otherwise you have no way of knowing a battery has failed. Exactly. No point in having an emergency back up that doesn't actually work when you need it. Same applies to emergency lighting. > > I should qualify my quoted constraints assume normal domestic SLAs. > Higher grade UPSs use higher grade batteries that can last many years > longer before the recommended replacement. Even so it makes sense to check them annually. You don't really want a failed SLA spewing corrosive electrolyte over your UPS. Regards, Martin Brown
From: John Tserkezis on 30 May 2010 10:08 Martin Brown wrote: > And it is at least worth the occasional visual inspection of the > batteries for obvious signs of swelling and distortion of the case. Unfortunately, that's not an effective enough test. Most fail at end of life looking like new. So a visual test alone is not enough. Some of the large commercial scale units perform an internal battery test with a limited load, with the marketing blurb claiming it's good enough to detect a dud battery even while the unit is live and running, never having to go offline for the test. > UK replacement is annual test and replace on fail or 5 years old > whichever happens sooner. Mine live in a fairly cool environment and > would still hold a decent charge when retired. I have seen some 8 year > old ex UPS cells that I took to the recycle centre in a bucket because I > was unsure if the plastic cases would hold out. They were very obviously > mangled shapes with big bulges but still not leaking. We were so poor, we had to used other's disposed of batteries as our new ones. We went so far as trying to recover open circuit SLA cells. We dumped 10-15+ volts into the things (single 2v cells!), till they took a charge, then they somehow magically worked. For a while anyway. That's on good quality (albeit old) cells. I've never really had any luck reviving cheapies. > Even so it makes sense to check them annually. You don't really want a > failed SLA spewing corrosive electrolyte over your UPS. True. I've seen batteries go short, overheat, dry out and become basically useless. A real chore to get the buggers out because they had warped so much. No-one noticed because the power had never gone out in that time (many years) till someone thought it would be a good idea to actually look at it...
From: legg on 30 May 2010 14:09 On 28 May 2010 05:51:02 GMT, John Doe <jdoe(a)usenetlove.invalid> wrote: >I would like to learn something about proper care of sealed lead >acid (SLA) batteries. Is there any good authoritative science on >how SLA batteries should be cared for? I am searching... I suppose >battery makers/distributors would be one of the best sources of >information about that? > >Specifically... I wonder why instructions that come with SLA >batteries so strongly emphasize charging the battery immediately >after use, without specifying how much use. Are they concerned >that the battery might eventually self discharge too deeply? Or >does it degrade an SLA battery to rest it in any state but fully >charged? > >Thanks. Check the Yuasa website for VRLA operation and maintenance recommendations: http://www.yuasabatteries.com/literature.php http://www.yuasabatteries.com/pdfs/TechMan.pdf There used to be some good stuff from Exide, but I've lost links to the electronic format manuals. RL
From: Grant on 30 May 2010 17:24
On Mon, 31 May 2010 00:08:26 +1000, John Tserkezis <jt(a)techniciansyndrome.org.invalid> wrote: >Martin Brown wrote: > >> And it is at least worth the occasional visual inspection of the >> batteries for obvious signs of swelling and distortion of the case. > > Unfortunately, that's not an effective enough test. Most fail at end >of life looking like new. So a visual test alone is not enough. > > Some of the large commercial scale units perform an internal battery >test with a limited load, with the marketing blurb claiming it's good >enough to detect a dud battery even while the unit is live and running, >never having to go offline for the test. The only decent test is to take the battery down to minimum voltage on known load (for example, C/4) and charge it up again, measuring what came out vs what went back in. I scored a couple big 100AH batteries recently, one was 102%, the other 96% of capacity -- previous owner just took them in to supplier and asked for new replacement without having them tested ;) > >> UK replacement is annual test and replace on fail or 5 years old >> whichever happens sooner. Mine live in a fairly cool environment and >> would still hold a decent charge when retired. I have seen some 8 year >> old ex UPS cells that I took to the recycle centre in a bucket because I >> was unsure if the plastic cases would hold out. They were very obviously >> mangled shapes with big bulges but still not leaking. > > We were so poor, we had to used other's disposed of batteries as our >new ones. Sometimes you can get lucky > > We went so far as trying to recover open circuit SLA cells. We dumped >10-15+ volts into the things (single 2v cells!), till they took a >charge, then they somehow magically worked. For a while anyway. > > That's on good quality (albeit old) cells. I've never really had any >luck reviving cheapies. I've read that recovery rate is ~30%, it's a poor gamble on SLAs gone high impedance. The AC spiking over DC charging protocol is supposed to be better at recovering sulphated cells. > >> Even so it makes sense to check them annually. You don't really want a >> failed SLA spewing corrosive electrolyte over your UPS. > > True. I've seen batteries go short, overheat, dry out and become >basically useless. A real chore to get the buggers out because they had >warped so much. > > No-one noticed because the power had never gone out in that time (many >years) till someone thought it would be a good idea to actually look at >it... Admiral's test, pull the plug without warning and see what happens :) The occasional load test is probably the best way, otherwise replace on a schedule if your budget allows. Many places seem to have bottomless pockets for breakdowns but nothing for scheduled maintenance. Grant. -- http://bugs.id.au/ |