Battery calibration for true runtime with larger batteries?

[Anonymous user]

Hi folks,

I purchased a used APC 3000 RM (5U rackmount) off eBay without
batteries. The standard batteries for this unit are, I believe
basically four 18AH batteries, which are wired together (in serial) to
essentially create two 24v batteries which then get plugged into the
UPS. Since the UPS did not come with batteries and I was looking for
some more runtime anyway, I purchased four deep cycle marine batteries
which are 75AH each; this should in theory give me about 4x the
runtime.


When I wired everything up, and with just a small load like a light
bulb or a fan, the UPS was showing 0% load and the estimated run time
was 293 minutes. From reading in this forum and other places online,
I knew I should do a battery calibration for the correct run time. At
first I did a software calibration (using apctest from the apcupsd
suite), but that did not change the run time. After allowing it to
charge back up to 100%, I then figured I should do a non-software
calibration as was described elsewhere in these forums. I put various
items on the UPS until it got up to a load of around 43%, at which
time it gave a runtime of just under 1/2 hour. I pulled the mains and
waited until the UPS ran down and shut itself off, while monitoring
the values. The runtime estimate decreased in a fairly linear fashion
for a while, but then once it hit 5 minutes it sat there for a long
time, went down to 3 minutes and sat there for a while too before
finally shutting off. At a 43% load, I got a runtime of around 1.5
hours. I removed all the loads, reconnected the mains and allowed it
to charge up.

Unfortunately, with just the small load, it is still giving me the
same 293 minute runtime as before so it appears the calibration didn't
work. Is my large battery capacity simply beyond the UPS's ability to
comprehend? Is it somewhat hardwired for the idea that its batteries
will max out around 18AH?

I notice that there might be an eeprom option to tell the UPS that
it has additional battery packs connected. Should I change that value
until I get a runtime which is more in line with the actual capability
of my batteries? The only thing with that is I believe any additional
batteries would normally be plugged into the 48v connector in the
back, but of course I have nothing plugged in there; I don't know if
the UPS actually knows that or not, and if it does, if telling the UPS
it has extra batteries when there is nothing plugged in there might
cause any errors or problems...

Many thanks in advance for any help or advice anyone can provide!


-Jon-

Message was edited by: JonS

[BillP]

hi,

i have same problem, please tell me the process,

and i have APC SUA1000I and i attached the 90AH two betteries externaly please tell me the counts of bettery calibration by using hyperterminal

what should i change with this (07 B5 13 BC 0A) that my apc gives proper backup and charging please im thankfull to u. im disturb for 2 years

[BillP]

hi,

i have same problem, please tell me the process,

and i have APC SUA1000I and i attached the 90AH two betteries externaly please tell me the counts of bettery calibration by using hyperterminal

what should i change with this (07 B5 13 BC 0A) that my apc gives proper backup and charging please im thankfull to u. im disturb for 2 years

[BillP]

If needed, you can adjust the external battery count easily with a basic terminal emulator.  (Such as HyperTerm or PuTTY in windoze)  This will increase runtime by multiples of the original.  (so if runtime is 20 min, setting external battery count to 1 will make it 40, 2 will be 60, etc.)

Here's how:

1. Hook up a computer to the UPS via a APC "smart" serial cable. If you don't have a serial port on your computer, you can use a USB to serial converter.  You must use an APC cable, as the UPS uses a non-standard pinout.
2. Start up a terminal emulator such as HyperTerm or PuTTY (with serial support).
   Port settings are: 2400 baud, 8 data bits, 1 stop bit, no parity, flow control is Xon/Xoff. When using HyperTerm, make sure FIFO buffer is NOT set.
3. When connected, you will NOT be greeted by any prompt!  Typing in Shift + Y should show the letters SM in the terminal window, this will confirm your connection.
4. Press > to see the current battery pack count. If no battery packs were installed, it should say 0.
   Pressing + will add one pack to the count, - will subtract the count by one. If you have multiple packs, press + several times to add them all. When finished, press > again to check the count.
5. When finished, you can simply close the window (or end the connection in HyperTerm).

-Phil

[BillP]

This is an interesting topic.

I had originally though that the UPS would continue to run based upon the charge condition of the batteries.  The UPS would run until a preset charge level was reached based upon voltage measurements that the UPS conducted during the period where no input AC was present.  I also thought that the internal software would continue to update, or predict, runtime based upon running the UPS on batteries until minimum charge level was reached.  Based on my assumptions, if correct, a battery pack that had four times the capacity of the OEM 18AH pack would run for ~4 times as long.

Maybe I over estimated the intelligence that the infernal UPS software possesses?

[voidstar_apc]

BlueSky wrote:

This is an interesting topic.

I had originally though that the UPS would continue to run based upon the charge condition of the batteries.  The UPS would run until a preset charge level was reached based upon voltage measurements that the UPS conducted during the period where no input AC was present.  I also thought that the internal software would continue to update, or predict, runtime based upon running the UPS on batteries until minimum charge level was reached.  Based on my assumptions, if correct, a battery pack that had four times the capacity of the OEM 18AH pack would run for ~4 times as long.

Sounds about right. Calibrating the runtime prediction helps attached PCs manage their time on battery and shut down before the UPS runs out of battery. If the UPS thinks the battery is worse than it really is, then the PCs may shut down prematurely. The PCs can also command the UPS to wait for the PCs to shut down and then turn itself off (or in APCUPSd parlance, "killpower").

[Anonymous user]

Sorry, I was having some problems posting my question. The forum appears to not print any line that starts with spaces (like indenting the first sentence of a paragraph...)

[Anonymous user]

.

[BillP]

Hi,

There is a [KB-article|http://jpaa-en.apc.com/cgi-bin/jpaa_en.cfg/php/enduser/std_adp.php?p_faqid=1497&p_created=1010390400... ] which describes very little if any improvement in run time on older models even after a dry calibration.

Have seen this myself in a 2200 unit, although it still provides good endurance but doesn't report a realistic run time even after the dry calibration effort.

Hope this helps.

[TheNotoriousKMP_apc]

Jon,

I will private message you with information tomorrow pertaining to Hyperterminal and checking the count.

However, there can be no guarantee with your runtime with a makeshift setup of batteries in there. Just because they are marine-grade and SHOULD give you 4x the runtime, the UPS will never recognize that, so you will get the exact runtime, but we can check that too.

[Anonymous user · Accepted solution]

Hi folks,

I purchased a used APC 3000 RM (5U rackmount) off eBay without
batteries. The standard batteries for this unit are, I believe
basically four 18AH batteries, which are wired together (in serial) to
essentially create two 24v batteries which then get plugged into the
UPS. Since the UPS did not come with batteries and I was looking for
some more runtime anyway, I purchased four deep cycle marine batteries
which are 75AH each; this should in theory give me about 4x the
runtime.


When I wired everything up, and with just a small load like a light
bulb or a fan, the UPS was showing 0% load and the estimated run time
was 293 minutes. From reading in this forum and other places online,
I knew I should do a battery calibration for the correct run time. At
first I did a software calibration (using apctest from the apcupsd
suite), but that did not change the run time. After allowing it to
charge back up to 100%, I then figured I should do a non-software
calibration as was described elsewhere in these forums. I put various
items on the UPS until it got up to a load of around 43%, at which
time it gave a runtime of just under 1/2 hour. I pulled the mains and
waited until the UPS ran down and shut itself off, while monitoring
the values. The runtime estimate decreased in a fairly linear fashion
for a while, but then once it hit 5 minutes it sat there for a long
time, went down to 3 minutes and sat there for a while too before
finally shutting off. At a 43% load, I got a runtime of around 1.5
hours. I removed all the loads, reconnected the mains and allowed it
to charge up.

Unfortunately, with just the small load, it is still giving me the
same 293 minute runtime as before so it appears the calibration didn't
work. Is my large battery capacity simply beyond the UPS's ability to
comprehend? Is it somewhat hardwired for the idea that its batteries
will max out around 18AH?

I notice that there might be an eeprom option to tell the UPS that
it has additional battery packs connected. Should I change that value
until I get a runtime which is more in line with the actual capability
of my batteries? The only thing with that is I believe any additional
batteries would normally be plugged into the 48v connector in the
back, but of course I have nothing plugged in there; I don't know if
the UPS actually knows that or not, and if it does, if telling the UPS
it has extra batteries when there is nothing plugged in there might
cause any errors or problems...

Many thanks in advance for any help or advice anyone can provide!


-Jon-

Message was edited by: JonS