Low runtime / new batteries discharging fast - SUA3000RM2U-BR

[Anonymous user]

Hello,

I have a SUA3000RM2U-BR for ~5 years. Last week I replaced 8 batteries for new ones, exactly the same brand and specs of the original (Long, 12v, 5ah). After three days charging I decided to test and realized the runtime was extremely low (~3 min).

Through extensive reading I came to know of some programs, such as apctest and APCfix. Through apctest I managed to change the battery dates and via APCfix, the runtime was fixed. At this point, according to the forum information, I though my problem was solving. However, although the runtime was ~90min, removing the power cable resulted in quick battery discharging (~3-4 minutes, 10% load).

Maybe changing the battery constants would be my last resort, but I don't know how to do that with my UPS model (SUA3000RM2U-BR).

The UPS is on a Ubuntu server. APCfix was run using wine and apctest installed with apt-get.

I don't know what else to do, as there is no APC support anywhere close to my location.

Any help is highly appreciated.

Regards.

[BillP]

Hello Thiago,

Did you install an RBC cartridge or did you replace the battery cells?

Thanks,

Shanon

[Anonymous user]

Hello Shanon,

I am on a low budget and it was hard to find the Rbc. Hence, I replaced only the batteries. I was careful enough to get the same brand and specs.

Thiago

[BillP]

Hello Thiago,

We always recommend replacing the battery with a genuine APC RBC, as we can not speak to the exact specifications of any replacement battery cells. Plus, there is a specific way that the cells would have to be installed, to make sure you have the same voltage and Ah. We have seen this type of behavior when third party batteries are used. 

I am not able to support replacing the cells, so I am not able to give you a wiring schematic of the RBC43, which is the correct replacement battery cartridge.

Thanks,

Shanon 

[BillP]

Test the voltage at the batteries and compare it to the voltage displayed in whatever program you use to monitor the UPS as I had this and it turned out the UPS read 7 volts high so thought it was fully charged when the batteries were flat. The new batteries gave an overcharge warning that never dawned on the APC support staff about this problem so I was sent 2 new sets before I diagnosed this myself.

[voidstar_apc]

FYI, apcfix probably changed the battery constant which is why the UPS thought it had 90 minutes of runtime. When you removed the power cable, the UPS monitors the discharge of the battery and adjusts its runtime estimate to be more accurate. Unfortunately 3-4 minute runtime at 10% load probably means something's wrong with the batteries. Assuming your pack is assembled correctly, perhaps the battery vendor let some of them sit on the shelf for too long. Good luck.

[Anonymous user]

Thanks, Gareth.

APCfix show battery voltage ˜54 (I am not at work now to tell you the precise value).

I don't know if I understood you correctly. How can I measure the voltage at the batteries with the battery module on the UPS ?

The problem you diagnosed can be fixed?

[Anonymous user]

Hello voidstar.

Thanks for the information. Is there any way to find out if (and which one) there is a bad battery without purchasing another set of 8 batteries?

I am pretty confident it was properly assembled -- we replaced one by one to preserve the correct wiring.

[Anonymous user]

Hi Shanon,

I live/work in a smaller city in Brazil, at least 300km far from the nearest APC authorized support.

Being on a low budget and far from APC techs, I tried to do the best I could. I opened the unit and photographed the batteries and wiring. Then I went online and purchased the same brand and specs.

BTW, the unit had the original batteries, which were Long, not APC.

Given the weight, cost and the 5 years of usage, it is probably not worth paying for transportation of such heavy equipment to an authorized APC location..

Anyway, thanks for your input.

[voidstar_apc]

I'm not very familiar with diagnosing battery packs for bad cells since APC only sells full packs. Maybe someone else here has advice on the subject.

Gareth's suggestion is a good one. You'd typically use a multimeter or DVM to check the voltage.

[MarkC_apc]

Replacement and Testing of the RBC43 battery pack is quite easy.  It appears you have replaced the batteries correctly or the SUA would not have even started.  Testing required a DC voltmeter.  Diagnosis without is not possible.  A fully charged pack that has "rested" for a few hours should read around 51 volts DC.  When the SUA is operating on mains, it should read around 54 volts DC (float charging mode).  When the SUA is operating on the batteries (mains disconnected), it can read between 47 and 50 volts.  If the voltage quickly drops to below 50 volts when the mains are disconnected with little or no load, one or more of the new batteries is likely bad.  If you are good with batteries (wearing rubber gloves and NO JEWELRY), you can determine which is bad by opening up the battery cage, hooking up the blue Andersons outside of the SUA, applying a light load on the battery pack (without mains) and measuring individual battery voltages.  The bad battery will drop in voltage much more quickly than the rest of the good batteries, so you have to take continuous readings quickly.  This takes some skill and is dangerous due to DC voltages are exposed at every terminal!!  Expect 12.5 volts to drop to around 12 volts each as the batteries are slowly depleted.  If they are all good, they will drop together equally.

Again, you must know what you are doing with batteries to not cause arcing, or even a dangerous shock (50 VDC is dangerous!!).  Otherwise, get all the batteries replaced under warrantee.  

Note that "calibration" of any new set of batteries is required.  You must put at least a 30% load on the unit from full charge to low battery alarm.  However, with only a few minutes of run time, this does not seem to be the issue.  Maybe Schneider techs can comment on this.

[Anonymous user]

​Hello Mark​ and others​,

Thanks for the detailed response. Based on what you said, I performed a test today. I removed all the equipment from the SUA (load 0%) and unplugged it from the mains. While charged and on the mains, the battery voltage was ~54. After removing the SUA from the mains, it dropped to ~51 as you said.

Curiously, even with load zero, the batteries were discharged in less than 2 hours. However, the battery voltage did not drop considerably: 50.76 at 82% charge; 50 at 70% charge; 48.6 at 52%; 47.25 at 34%; 46.5 at 24%; 44 at 3%. Unexpectedly, after reaching 0% and turning off, I managed to turn it on again and it showed a level of 10% and the battery voltage was back up to 50.2. I tested some minor load (a small fan) and the UPS was not supplying energy to it, which was also confirmed by the monitoring software. It went off again and when I turned it on, the charge was 2​0%  and went up to 28%​!​ At some point, even the second led corresponding to battery levels was on. 

It went through multiple rounds of automatically turning off and when I manually turned on it had charge.​ Actually these cycles have not ended yet, but I decided to post back here to see if I can get some light.

​A side note, when it was close to zero (before​ my​ turn off​/on weird cycle​), APCfix reported a "message from UPS: Line failure"​. ​Now the status tab shows only "Ready, Return of line power" and "SmartBoost mode" boxes checked.

Thiago​

[MarkC_apc]

So you were able to get the batteries to 100% on the LEDs and to run for several hours - even though at "no load"?  Not sure what you meant by "cycling" but it is very hard on the batteries to completely discharge them - and then to keep them at a low state of charge.  The voltages you took indicate a very low state of charge (SOC) on the batteries.  I would not recommend going to these low voltages.   2 hours at no load would indicate that the batteries do have some charge capabilities since these SUA units do have a significant idle draw.  If you can do a load test by applying about 800 to 1000 watts (hair dryer or space heater) and see if it takes 15 to 20 minutes to deplete the batteries.  When you do this, it should recalibrate the battery run times also.  If you get significantly less run time, then likely one of more of the "new" batteries are bad.  You must recharge the batteries for several hours  (overnight) before doing such a test.

[BillP]

It sounds very much like my situation. To test the battery voltage you normally test the leads on the battery pack at the front and times it by 2 for a rough estimate (it is normally spot on if you times by 2) I did this then as I was going to check for bad capacitors I measured the voltage on the mainboard where the +/- connections are and it was what I expected. 25 volts at one pack 50 at the board and 56.5 shown via the software (UPS and NMC were the same) Based on this I knew the higher than true reading meant it was only offering a maintenance charge to a nearly flat battery and with it being a SMT series I could do nothing to adjust it unlike the older models that had a pot you could turn.

[BillP]

Just to add a flat UPS will come back up at a few % charge as lead acid batteries recover a little once the loads removed but will drop straight off if a loads applied.

[Anonymous user]

Hi Mark,

Thank you for the tips. I will do this tomorrow and post back here the results.

[Anonymous user]

Thanks, Gareth. I am sorry, but I am far from an expert on this. So, is there a solution for this problem?

[BillP]

I doubt it if the true voltage it lower than the UPS shows, some older had an adjustment inside but you would need to be qualified to open up the UPS as it is very dangerous. APC offer no support for this and I have never seen inside any UPS other than the one I own.

[Anonymous user]

Ok, but then, in your opinion, the problem is with one or more batteries or with the UPS?

[MarkC_apc]

Be aware that if the battery packs are not being charged by the UPS (batteries disconnected or the mains disconnected from the UPS) , they will measure around 50-51 volts (nominal - 48) if fully charged - down to 48 if low on charge.  If the mains are connected, the UPS charge unit is always activated and will bring the voltages (eventually, but usually fairly quickly) up to full "float" or around 55 volts.  The SUA RM units with the RCB43 battery pack will show the full voltages ("48") on the blue Andersons.  The extended run units (XL) have two large "24" volt battery systems inside the unit that are in series, so that you can find both "24" and "48" volt wiring inside the UPS itself. These have "grey" connectors that should measure "24" volts.

[Anonymous user]

Thanks, Mark. As per my description, do you have an idea of what is going on?

Today, after fully charging the batteries, I disconnected them and measured the voltage one by one. All the same value and apparently ok without load.

[voidstar_apc]

1) Is the battery voltage reported by the UPS substantially higher than what your volt-meter shows? Your measurements seem to indicate the battery is being correctly charged so this is probably not the issue.

2) Does the amount of time the UPS operates match the estimated runtime displayed? Regardless of its estimate, the UPS runs until the battery is drained or it is commanded to turn off. The estimate is important for doing safe shutdown with PowerChute. So if the two match, then the estimate is well calibrated.

3) During discharge with at least 30% load, did any of the batteries discharge unevenly, or was the voltage uniform across all of them? If a battery is weak and drops in voltage compared to the other batteries, then it's bad. If all the batteries are good, then maybe they just have less amp-hour capacity than the originals.

[MarkC_apc]

As Voidstar suggested, you have to put load on the batteries to see if one of more is "weak".  Try that and let us know.

[Anonymous user]

Hi guys.

I am performing some tests today. One of the 8 batteries failed with a 1 milli amp-lamp and a 35 milli-amp cooler. I though I had solved the problem, but replacing it with an older battery (which is still in a fair condition) has not solved the problem, as the fully charged system was keeping 135W load for less than 5 minutes. So, there might be other doomed batteries there.

A friend is bringing a stronger lamp this afternoon and we will re-test the other 7 batteries. The nightmare continues.

I will post back later.

[Anonymous user]

Hi all,

I have just tested the remaining 7 batteries with a 4amp car headlamp and 2 of them had a quick drop in voltage. I replaced them with batteries from the old pack (after testing them using the same method).

Now I will return these three defective batteries to the seller and ask for replacements.

I tested them with ~200W load and it took 20 minutes to go to 22%, when I put it back on the mains. It is still not good, but I guess this can be explained by the 3 old batteries (used for 5 years) that I used to complete the pack. Besides, these were in a desk for seven days without charge. Is it reasonable to assume this?

I will repeat this test tomorrow morning, after 12+ hours charging, but at least the 3 min runtime phase is gone. 

Is it bad to use 3 old batteries with the new ones while I don't receive the replacement parts?

Thank you all for the support. It was definitely of great help !

Thiago

[MarkC_apc · Accepted solution]

For now, put four of the best batteries in one "48" volt section to operate the unit - that is better than having even one "weak" battery in each of the 2 parallel battery sections.  With that many "new" batteries being weak, I'd find another supplier and return all of them - it would appear they were not actually new, or left uncharged for months.  Think of batteries like a chain - one weak one and the whole chain breaks quickly!

[Anonymous user]

Thanks Mark. I've just posted them back to the seller and will stay with the old pack until the new ones (from another seller) arrive. Again, all of you guys were of great help, which I highly appreciate.

Regards.

[MarkC_apc]

Again, if you have 4 "good" batteries from your old stash, connect them up and make sure the unit passes the battery test - which will be more reliable than the two sets, both with "weak" batteries in the series.

[Anonymous user]

Thanks, Mark. I have done that. Now waiting for the new set.

[Anonymous user]

Hello all.

Just a feedback: got my new batteries, ran a calibration and it is all fine now.

Thank you all for the help.