BR1500G battery level falls quickly

[Anonymous user]

BR1500G + BR24BPG. Have 4 of this combo and one is behaving oddly.

Powerchute 3.0.2 personal reports 865 watts.

Disconnect power from wall.

Reports 240 minutes left and then at about 10 minutes the battery indicator

falls like a rock and (as powerchute reports after restoring power) it shut

the pack down at 10 min. 6 seconds.

I just attached laptop  to another such and it reports 865 yet this second one

does not suffer the fall-like-a-rock problem.

What is going on here?  How do I fix this?

[BillP]

Hi David,

Can you provide the serial numbers of the BR1500G and BR24BPG which are affected? What type of load are you attempting to support? Is it a similar load for all 4 configuration?Is this a new install ? When did you first notice the short runtime?

Secret Squirrel

[Anonymous user]

The serial numbers being on the bottom of the units, and the UPS being in critical duty here, I'm going to have to figure out how to deal with getting those numbers...shortly.  As soon as I get them I'll post here. And then I'll put S/Ns on the front where I can see them...

I just realized that the BR1500G was purchased mid 2014 while the BR24BPG was Nov 2017.

I  ordered a new BR1500G battery  Sept. 3 (from apc via amazon) as 2014 is clearly out of date, though powerchute reports

no issue with it. It won't arrive for a few days. I trust I can use the packaging to return the old battery.

Recently (some, at least) auto-parts shops here have stopped accepting non-automotive batteries for recycling.

[Anonymous user]

OK. Took photo with phone, awkwardly.  Readable in the photo.

BR1500G  4B1423P00905

BR24BPG  7A1735L25066

[Anonymous user]

I failed to mention that we notice the problem because with ohmconnect.com

we get paid a small amount to shut off all power to the house for an hour

on request. Once or twice a week.

You asked some questions I  failed to notice. Here are answers.

"What type of load are you attempting to support?"

about 15W of router and raspberry pi .  That's what just lasted 10 minutes.

It used to easily deal with an additonal 50W of an idle desktop computer.

"Is it a similar load for all 4 configuration?Is this a new install ?"

It used to deal with the 60W easily for an hour and even some few minutes

of 160W.

One sometimes deals with 400W for a short while but that's unsustainable.

The others show a sensible minutes-left and it drops sensibly

over time based on the load. With an intentionally low load

(each one sees a different situation) the minutes left starts at  990,

180, or 350.

We restore power to the house at 60 minutes so none should run out of power!

"When did you first notice the short runtime?"

A couple weeks ago (maybe 3+) at the end of an hour

the idle computer and router+ stuff were...off, as

the ups had shut itself down.  Only

on the day I posted here had I really done an experiment to

truly understand how it went (how fast and how strangely the

minutes-left looked)..

And, too, at that time (sept 1?) I rememberedI actually had a Windows

machine available so I installed PowerChute on it to learn what the batteries said.

I also tried apctest at that time. It reported ok with "apcaccess status"

though apc test ... I forget if that really worked. If you wish I could

try the linux tools again.

[Anonymous user]

Here is something from the machine on that ups:

q3 526: apcaccess status
APC      : 001,036,0854
DATE     : 2018-09-01 20:06:32 -0700 
HOSTNAME : q3
VERSION  : 3.14.14 (31 May 2016) debian
UPSNAME  : q3apc
CABLE    : USB Cable
DRIVER   : USB UPS Driver
UPSMODE  : Stand Alone
STARTTIME: 2018-09-01 20:05:03 -0700 
MODEL    : Back-UPS RS 1500G
STATUS   : ONLINE
LINEV    : 121.0 Volts
LOADPCT  : 8.0 Percent
BCHARGE  : 100.0 Percent
TIMELEFT : 398.8 Minutes
MBATTCHG : 5 Percent
MINTIMEL : 3 Minutes
MAXTIME  : 0 Seconds
SENSE    : Low
LOTRANS  : 78.0 Volts
HITRANS  : 150.0 Volts
ALARMDEL : No alarm
BATTV    : 27.0 Volts
LASTXFER : No transfers since turnon
NUMXFERS : 0
TONBATT  : 0 Seconds
CUMONBATT: 0 Seconds
XOFFBATT : N/A
SELFTEST : NO
STATFLAG : 0x05000008
SERIALNO : 4B1423P00905 
BATTDATE : 2014-06-02
NOMINV   : 120 Volts
NOMBATTV : 24.0 Volts
NOMPOWER : 865 Watts
FIRMWARE : 865.L5 .D USB FW:L5
END APC  : 2018-09-01 20:06:54 -0700 

[BillP]

Hi David,

I believe what your running into is a run-time limiter which is programmed into the unit. Anything below a certain %, will cause the unit to enter the run-time limiter mode and turn the unit/load off to prevent an unnecessary discharge of the internal and external batteries. This run-time limiter is between 15-20w I believe. If you were to increase the load, I believe you will see a much more accurate run-time.

Secret Squirrel 

[Anonymous user]

That sort of limiter ... I would not have thought of.  I'll leave that system on router + idling desktop which will bring the load to 50W. I am not optimistic that will help at all (nor do I see how it can hurt).

The replacement battery for the 4-year-old BR1500G is taking forever to get here (ground ship only). No idea when it will arrive.

One of the other BR1500G + BR24BPG I've turned down to under 20W load (shut off the 700W system and left 3 non-busy Raspberry Pi's running)  and seen the 'minutes left' at 999 decreasing slowly over an hour to the lower 900's and never had the slightest problem.  Which is why I'm not very optimistic a 50W load (Instead of 16W) will matter, but I guess we'll see.  Maybe I'll run an experiment - unplug from wall with the 50W load. I guess I should.

[Anonymous user]

So with router, rpi, desktop idleing incl. monitor, it was 44W load.

At 14:13 I unplugged the wall power. It showed 527 minutes left.

At 14:19 it showed 522 minutes left

At 14:28 it showed 496 minute left.

Then the minutes left dropped quickly: 448,416,385,343,306,210 by 14:29.

At that point I plugged into power again rather than

let power go low enough to turn off the power to the load.

The battery meter on the BR1500G showed

there was maybe 25% left.  If there are 5 white blocks when full that would mean 20% as two blocks are white.

Something appears to be wrong with the ups?.

So... what now? Suggestions?

I have no idea when the replacement battery for the BR1500G will arrive or even if it

was actually shipped.

[Anonymous user]

Summary: still FAILS

A new RBC124 arrived and I unplugged everything

and installed the battery in the BR1500G replacing the 2014

battery  (the BR24BPG was new 2017).   Plugged the load and

the BR24BPG back in.

Plugged UPS into the wall.  Let it charge for about 4 hours (the fan stopped

long before 4 hours). powerchute says battery full.

With  load to 44Watts unplugged wall power at 10:54AM

10:54 says 527 left

10:59 says 522 left

11.04 says 501 left

then at  11:04 says 480 422 385 306 so I plugged back into wall.

Then after being on wall power for 12 minutes powerchute says remaining battery charge is 51%,

which is sort of absurd.

That's not a useful performance.  This is broken.

What do I do now?

[BillP]

Hi David,

I would recommend placing a good 30-40% load on the unit and unplug the input power cord. Allow the unit to completely exhaust itself to the point of turning off. Once off, plug the unit back in to power and allow it to completely charge for 24hrs before reconnecting any equipment.

This process will aid the unit in readjusting it's battery constants to help improve the accuracy of the units' run-time when switching to battery. The unit is not meant to power small loads and therefore it does have a difficult time providing an accurate run-time time.

One last thing to mention is that although the BR24BPG is fairly new (2017), when installed with older batteries the older batteries can cause strain on the newer batteries resulting in premature failure/weakening. It is never recommended to mix old batteries with new.

Secret Squirrel

[Anonymous user]

Followed the procedure and preliminary indications are things are ok, original

setup (with one new battery) now back in place..

The temporary replacement BR1500G (no BR24BPG on it) was just to

filter power while I went through the procedure.  There's been

no time to do the sensible follow-on test yet.  Likely a live test

will occur this week (outside of my control...).

[Anonymous user · Accepted solution]

In the natural course of events I discover that the unit (even after taking the steps requested

above)  still falls down a cliff after 10 minutes.

Earlier remarks about the run-time-limiter make no sense to me as two other (identical except

for system dates) systems happily run for an hour or even two (disconnected from wall power)

with 15W power draw while the remaining-minutes slowly decline from 600 minutes

(more or less, depending on stuff here)   ... in a sensible way.

Unless there is some way to make progress I'm stuck with this poorly-performing

unit.  I certainly don't intend to pay $250 to rebuy the whole thing.