UPS Testing [RESOLVED]

[bluestreak]

Hi!

New to the forum but been a long time fan! 

I have an APC SmartUPS that I was lucky to obtain from a customer site where it had sat dormant for quite some time. To my surprise the unit was at the time in an ok condition and the batteries were still in reasonable order.

However UPS recently died its an APC SUA3000RMI2U so 3000VA line interactive rack unit. It died in a peculiar fashion and instead of a dead battery it just emitted a continuous tone one day and refused to start or do anything useful. I figured it was the batteries but couldn't be sure as according to the manual, APC support, google and everything one I spoke to a dead battery should give a series of beeps and show the dead battery light.

Through tinkering and checking what I could using only a very basic multi-meter, all while being extremely mindful that I only have limited knowledge of the technology, I did eventually stop the continuous tone whenever it was plugged in, batteries attached or not, and it started to emit the right beeps and show a dead battery light. Sigh of relief here as the tone was very annoying and loud.

I also found this video to be very helpful although not about the same unit  https://youtu.be/Fj7e3WGUKO8

Taking the battery tray out its got 8x 12v Pb batteries in a 2x4 arrangement giving a 48v output. I've checked all the cells and found only one of them is bad showing less than 6v. So I thought well could I run the UPS of only a single 1x4 pack? just to get some UPS at least until I can source replacement batteries.

Nope that doesn't work. It states bad batt still .

This makes me wonder about the state of the other batteries are they too dead. Well checking each voltage i get 13.10~volts on each of the 4 I'm using. I ensured they were all charged up tested again with same result. I then left the project not having any money to do much else. The batteries have sat under my desk since then, so about a month not doing anything, I figured if the batteries were dead they would show a big voltage drop. Checked them just now and they are all showing the same 13.10V~ as before. The question now, are these batteries dead or do I have a dead UPS?

My question to the community is how can I test the UPS without buying a load of new batteries and wasting the money. The unit is quite nice and I'd like to re-use it but just don't want to spend money on it if its not going to work.

Does anyone know if there is a way to power the UPS from a powersupply to essentially fool it into thinking it has batteries connected? Thus testing the UPS is going to work with brand new batteries.

I'd appreciate any help

(I'd like a new Online UPS, but new batteries = $200, new UPS = $3000+, so simply cant justify a new UPS at the moment)

Thanks

Bill

[Ian.M]

A fully charged lead acid battery can have a normal terminal voltage with as little as 1% of its plate area still active.  Where it will fail is loss of its rated capacity and ability to deliver significant current.   

Load test the individual batteries.   Its 3000VA rated, with 8 batteries (I believe 4.5AH), and assuming 80% efficiency, each has to be able to deliver close to 40A at full output.

I found a runtime graph for it:

If you wanted to test a single battery at full load, you'd need a 0.3 ohm 500W load resistor.  Four 12V 120W in-cup water heaters off Ebay, wired in parallel should do the job and only cost you about $10.  They *MUST* be immersed in a large bucket of water throughout the test.  A half load test would be two such heaters, and you'd be looking for 10 minutes runtime. If the battery under test cant manage 5 minutes at half load, its FUBARed.   

Once you know how many batteries are still usable, you'll know how many more you need to scrounge to make up a test bank.  If you cant get the test batteries in the original battery pack, you'll have to extend some wiring.  Always assume that UPS battery banks are mains live unless proven otherwise, so be careful if you set up an external battery bank.

[nctnico]

When it comes to APC I'd just get a battery to test it. APC UPSs are notorious for cooking the batteries so it is likely the problem is caused by the battery no matter what the symptoms are.

[Ian.M]

Unfortunately, getting a set of eight batteries to test it is very much a gamble.  If the existing batteries are all nearly dead or worse it may be the only way forward but I'd want to confirm that first.

[tautech]

Got a little 1kW APC here that I was given some time back and IIRC the charging circuitry was overcharging and cooking batteries. Luckily it uses only 2 x 12V batteries for 24V so it was reasonably easy to perform the diagnosis with a  couple of car batteries.
All the passives in the battery monitoring/charging part tested OK and I sourced the IC's needed and they're still in my IC storage trays. 

Must get back to it one rainy day......

[3db]

I agree with Ian.M.
It's worthwhile testing the batteries.
If you decide to use the UPS you will want to check out the batteries in future too.

3DB 

[CJay]

This makes me wonder about the state of the other batteries are they too dead. Well checking each voltage i get 13.10~volts on each of the 4 I'm using. I ensured they were all charged up tested again with same result. I then left the project not having any money to do much else. The batteries have sat under my desk since then, so about a month not doing anything, I figured if the batteries were dead they would show a big voltage drop. Checked them just now and they are all showing the same 13.10V~ as before. The question now, are these batteries dead or do I have a dead UPS?
Thanks

Bill

After repairing *many* UPS' the single most common fault was dead batteries.

If you had one measurably dead battery then the rest are likely to be stuffed as well and you should be replacing them as a complete set.

Load test the batteries, from memory the UPS you have uses either 12V 7AH batteries or 12V 4AH so something like a 50W headlight bulb would be suitable, monitor the voltage across the terminals of the battery (don't burn yourself or anything else with the bulb.) a dead cell will cause the voltage to drop immediately,  poor cell/s will cause it to droop quickly.

If it droops then charge the battery and test again, also monitor the time it takes to recover close to the off load terminal voltage, a good battery will recover quickly, a poor one will be much slower and may not recover at all.

APC UPS have a battery test mode, if you hold the 'On' button it will perform a battery calibration, this will, if you've replaced the batteries, allow the UPS to 'characterise' the new pack and it will immediately stop the whistling/beeping but don't have anything critical on the output as it will drop it if there's a problem with the UPS/pack.

[bluestreak]

Hi All,

Thanks for all the responses I've read through them all and I'm going to try a load test the batteries with a 12v 55w lamp I have.

How long and what kind of voltage drop should I expect? it doesn't have to be great but enough so the UPS will likely fire up.

Thanks again!

[tautech]

Hi All,

Thanks for all the responses I've read through them all and I'm going to try a load test the batteries with a 12v 55w lamp I have.

How long and what kind of voltage drop should I expect? it doesn't have to be great but enough so the UPS will likely fire up.

Thanks again!

Going on what CJay said I want them to hold at least 10V for a minute or so, with that result you'd expect them to at least fire up the UPS but if any are tired they might just fall over while doing so.
Maybe a top up after the bulb test. 

[Ian.M]

The batteries should be able to hold >10V running a 55W bulb for 50 minutes or so when new.  If they cant hold 10V for 5 minutes they are NOT in usable condition.

However using a 55W bulb as a test load is  fairly unrealistic as it will only draw a very small fraction of the max load the UPS can draw.   That means it wont pick up a small increase of internal resistance, thats still large enough to cause excessive voltage drop when the UPS applies a brief test load, which its likely to do during startup.  That's why I suggested a heavier test load.

[bluestreak]

Ok I have run each battery with the lamp between 10 and 20 minutes and have four of them that didn't drop lower than 11v.

Hopefully these four batteries will be enough to get the UPS to run through the start up test on their own. Its just so I can be sure if I buy new batteries its not a waste. I'm not expecting them to last more and a couple of minutes.

[Ian.M]

Does the manual say it will run on one bank of batteries only?  If not you'll have to scrounge some small lead acid batteries.

You should fully recharge all the batteries you tested before you try them in the UPS.   Make sure the charger is suitable for SLA batteries - if all you've got is an old dumb car battery charger, don't use it unattended, stick the 55W bulb in series to limit the maximum current, and manually terminate the charge depending on the voltage across the battery measured with a DMM.

You could also float them all for 24H on a bench PSU set for 14.5V, and a current limit of about 2A, with a silicon diode to each battery for discharge protection.  If you are going to run it longer than 24H, check none of the batteries are over 13.8V.

[bluestreak]

I have a LiPro balancing charger that has the settings for PB batteries. Charging them all up now at 2A ok on the 12v setting. It's getting to 14.7V and then automatically managing the charge (It'll charge almost anything).



The whole pack is a set of 8 batteries but in 2 rows of 4 in series so outputs 48v. The manual doest say anything about it as it like expects you to replace the pack as a whole. I think this should work with only 1 row though as the voltage output is still the same. It then has a a single Anderson type connector on it to the UPS. I'll test it tomorrow once the batteries are charged up.

[Ian.M]

I thought they might have separate +ve connectors to the UPS for redundancy.  If you are only using one string, then you really need to check them on a heavier load than a 55W bulb.   Do you have any 12V heating gadgets >100W you could try?

[madires]

That SmartUPS should also run without any batteries, since it supports battery replacement while in operation. So I'd guess there's something else wrong. BTW, the old SUA models got a hidden service menu which allows to adjust the float voltage of the batteries. It seems that APC loves to set the float voltage higher than recommended, especially when selling overpriced battery packs 

[CJay]

The batteries should be able to hold >10V running a 55W bulb for 50 minutes or so when new.  If they cant hold 10V for 5 minutes they are NOT in usable condition.

However using a 55W bulb as a test load is  fairly unrealistic as it will only draw a very small fraction of the max load the UPS can draw.   That means it wont pick up a small increase of internal resistance, thats still large enough to cause excessive voltage drop when the UPS applies a brief test load, which its likely to do during startup.  That's why I suggested a heavier test load.

You are of course correct in that they should be able to support a heavier load but given the run times expected and the capacity of the batteries in use, a 5 amp load will serve to weed out the gross failures and in my experience a UPS with batteries that 'pass' my simplistic test will run the UPS well enough witha small load to allow the OP to decide if it's worth buying a new set of batteries.

[CJay]

That SmartUPS should also run without any batteries, since it supports battery replacement while in operation. So I'd guess there's something else wrong. BTW, the old SUA models got a hidden service menu which allows to adjust the float voltage of the batteries. It seems that APC loves to set the float voltage higher than recommended, especially when selling overpriced battery packs 

It will, the batteries are hot swappable but it will drop like a sack of s**t if it tries to self test faulty cells or you're unlucky enough to have a power outage during replqacement.

I've also seen them fail a new set of cells on installation but pass them after a power down.

[FlyingHacker]

I have seen numerous APC and CyberPower UPSes (we have about 40-50 of them) show all kinds of weird error codes and exhibit unexpected behavior due to the dead or weak batteries. I have wasted time taking a couple apart only to find that the real issue was the battery(ies) that tested "OK."

So, as others have said, you really have to start there. It is possible nothing else is wrong with it.

I find http://www.atbatt.com/ usually has the cheapest prices on the individual batteries. (Obviously you just buy the batteries and not the whole tray or combined unit from APC at inflated prices.) Though I did once find a deal on New Egg for like $8 a pop with free shipping!

[bluestreak]

Hi All,

so the testing of the batteries gave me 4 'ok' units. OK being that they lasted long enough to power a lamp for while. Not the best and most accurate test but enough to satisfy the batteries weren't a complete lost cause.

My future advise would be to have a larger load as many others have suggested to get a better result.

For me this has proved to be a worthwhile exercise though as I loaded the pack up with only 4 batteries in a single row. This gives the pack (sorry forgot to mention it was removable earlier) the right voltage the unit requires. I plugged the pack in and it registered the batteries. Powered on once and I immeadiately get bad batt warning light+beep. Switched it off using the buttons and then held the on button again. It fired up, ran a batt test and hey presto! the unit is now showing as working and a quick test with a flick of the wall switch shows it as holding a charge even when I have a server on it (I was using a dual PSU unit so I didnt kill it ).

I wasn't expecting it to work so was happy it has. I'm now in a better position to buy replacement batteries before using the unit completely. I have plugged it back in with a very small load (Small NAS), as I figured a small load may as well benefit from the UPS with a very limited capacity than no UPS at all while I order the replacement batteries. Had some rotten weather recently to so its very timely.

Thank you everyone for the help my question has been answered 

[GreyWoolfe]

It seems that a lot of people deride APC but I have 3, 12 year old SmartUPS 1500 UPS's in my office.  These are decommissioned assets from my company.  All they have ever needed was batteries and the company still pays for batteries for 2 of them.  They work as designed when Faulty (Florida) Power and Light makes the lights flicker.  I also have a couple of spare units and plenty of hardware to build battery packs.  I typically have to replace the batteries after about 4 years.

[FlyingHacker]

The older APC units were fantastic. About five years ago their quality went down. We had two or three units fail within a year or three. Then the quality seems to have gone back up in recent years.

When the APC quality went down we tried some CyberPower units. They were decent, but have trouble working with output from our building generator. So we went back to APC.

We still have some smaller APC units that are twenty years old and still work fine. One kept a NetBSD server machine up for just under ten YEARS (no reboot for 9years 11 months 4 days). Finally the battery in it died and I didn't know it. There was a power hiccup and it went down. I even used it to move the machine powered on 12miles in the back of an SUV. (Daisy chained another UPS for the ride.) That was at the three year uptime mark.

[CJay]

It seems that a lot of people deride APC but I have 3, 12 year old SmartUPS 1500 UPS's in my office.  These are decommissioned assets from my company.  All they have ever needed was batteries and the company still pays for batteries for 2 of them.  They work as designed when Faulty (Florida) Power and Light makes the lights flicker.  I also have a couple of spare units and plenty of hardware to build battery packs.  I typically have to replace the batteries after about 4 years.

I've got similar somewhere, APC for all their problems were/are a pretty good solution, there are better and there are worse out there but you need to search to find better and swallow the extra expense.

I've rarely had a gross failure on an APC that was anything other than batteries and I've replaced *hundreds* of them in my time.

Even on the units that had faults other than batteries I was usually able to pinpoint it to a particular problem external to the UPS (problems with lightning or building/street wiring) and I don't think I've ever seen UPS protected equipment damaged.

[CJay]

Hi All,

so the testing of the batteries gave me 4 'ok' units. OK being that they lasted long enough to power a lamp for while. Not the best and most accurate test but enough to satisfy the batteries weren't a complete lost cause.

Excellent, pleased you've got to the bottom of it.

My future advise would be to have a larger load as many others have suggested to get a better result.
Thank you everyone for the help my question has been answered 

Maybe if the batteries were going to be re-used 'in production' but as a rough guide the load test you did is perfectly adequate to get you to the point where you can make a value judgement, I don't think I'd ever rely on a set of batteries that had failed previously to run any of my IT (I do have several ex UPS batteries performing other non critical light(ing) duties though)

We got there though, you've got a working system and didn't have to throw away a good piece of equipment