LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   General (https://www.linuxquestions.org/questions/general-10/)
-   -   APC Smart-UPS, not that smart? (https://www.linuxquestions.org/questions/general-10/apc-smart-ups-not-that-smart-4175505275/)

trafikpolisen 05-17-2014 12:03 AM

APC Smart-UPS, not that smart?
 
Continue to the last part if you don't feel like reading everything:)

I recently got my hands on an APC Smart-UPS 1000XL that the factory I work at was going to throw out. I previously have a Smart-UPS 1000 (which i acquired the same way), but with lower capacity batteries, that has served me fairly well for about three years. Batteries in the newly acquired UPS were dead so I took the chance that there was nothing else wrong with it except for the batteries and bought some new ones.

Installed the new batteries and the UPS came to life. However, apcupsd reported Low bat and an expected runtime of about 1 minute with the batteries fully charged and under 50% load. After a quick search on google i found that the UPS might need calibration which involved a deep discharge of the batteries. I did a "Manual runtime calibration" according to instructions from the apcupsd's manual pages and after the batteries had recharged apcupsd reported 12 minutes runtime, which quickly fell to just a few minutes when the mains was pulled.

As much as I hate to deep discharge new lead-acid batteries i decided to do the procedure again and measure the time things ran before the UPS shut down completely. When i pulled the plug it took maybe two minutes before the battery-bar was down do one led and blinking. By this point the UPS would have sent a shutdown command to the computers if the USB-cable had been connected. It continued to run for a total of 31 minutes before the batteries were out, this with a home server, a bunch of network stuff, a fairly decent computer with monitor and a 42" LCD-TV hooked up to it. The batteries are now recharging and it will be interesting to see what the calculated runtime will now be.

Now the interesting question: The UPS monitors the load, battery voltage and presumably charge current, so why couldn't the capacity and remaining runtime be calculated in real-time based on the rate at which the battery voltage drops, instead of being based on the capacity the microprocessor (incorrectly) think the batteries have and the load? That way there would be no need to calibrate and no problem when replacing the batteries.

Ser Olmy 05-17-2014 05:31 AM

This is an APC classic. Several (all?) APC UPS models store a battery parameter in NVRAM indicating the capacity of the battery, and this parameter gets gradually adjusted downwards as the batteries age. This prevents the UPS from reporting overly optimistic runtime estimates with older batteries.

The problem is, sometimes this parameter doesn't get reset when you replace the batteries. I know it didn't when I replaced the batteries in my APC Smart-UPS 700, even though it's connected to a Windows computer running the official, proprietary APC software, and I followed the exact, official replacement procedure.

The solution is to use a terminal emulator and send the relevant reset commands directly to the UPS. To do this, you will have to know:

a) The proper "battery constant" value for the UPS model in question (9A for the Smart-UPS 1000XL, according to my notes)

b) The exact procedure for resetting the constant.

If you have a serial cable and are willing to try resetting the constant manually, I can provide the gory details.

trafikpolisen 05-17-2014 10:30 AM

Thanks for your reply!

Things appear to be MUCH better after the last manual calibration. Now this afternoon with the batteries completely recharged apcupsd reported 40 minutes of runtime with about 43 percent load attached. But I still wasn't too comfortable with pulling the plug. However, after ten minutes on batteries with the same load apcupsd reported 27 minutes left and 69% battery charge. I then put it back on mains. Sure, it maybe dropped a little bit too much, but it now feels "within normal". I wonder what two "unnecessary" deep cycles have done to the capacity of my new batteries though..

So, a calibration once or twice a year should keep this battery parameter from dropping "out of range"?

I did have the UPS hooked up to a computer (or rather a vm) with the proprietary software to turn off the "on battery" beep warning. There really should be a "Reset battery data" option in that software. No wonder companies are throwing out UPS's with bad batteries instead of just replacing them.

I read this http://www.apcupsd.com/manual/manual...ttery-constant about resetting the battery constant as you mention, but it felt a bit scary. Even though i got the UPS for free I don't feel like bricking it as i have put 175$ on the new batteries.
But if I should need to reset the battery constant, what kind of serial cable do I need? Is it a standard one, or some specially wired proprietary?

Emerson 05-17-2014 10:52 AM

Deep discharge won't damage your battery if charged immediately. Leaving battery discharged will cause damage.

Ser Olmy 05-17-2014 03:17 PM

APC uses a proprietary wiring scheme for "smart" communication. A standard serial cable would only provide "failed/on battery" and "shutdown" signaling by (ab)using the TxD/RxD lines.

The USB connection is little more than a glorified serial console, and according to the article in your link, the program apctest provides access to the UPS console, presumably regardless of cable. You may want to try that.

The procedure in the article is the exact same I have in my notes, and I can confirm that it works. I can't remember whether one's supposed to press Enter after issuing the commands, but that should be obvious once you've typed in the first command, "Y", as the UPS is supposed to respond immediately.

I seem to remember that my UPS emitted a beep or click every time "+" or "-" was entered to adjust the battery constant. If your UPS does the same, there's no reason to be worried.

maples 05-27-2014 09:34 PM

Quote:

Originally Posted by Emerson (Post 5172663)
Deep discharge won't damage your battery if charged immediately. Leaving battery discharged will cause damage.

I've heard and read a lot about batteries. If you keep them charged all of the time, they will go bad. If you discharge them below [insert random number] percent, they will go bad. If you need to store them, it's bad if they're empty, but it's also bad if they have a charge.

Basically, your battery is going to die. Maybe some things are less bad than others, I don't know. But from what I know, there is nothing you can do to make them last more than a few years, maybe a decade if you're lucky.

Emerson 05-27-2014 09:53 PM

Quote:

If you keep them charged all of the time, they will go bad.
Only if the battery is connected to a regular charger which will eventually overcharge. Fully charged lead-acid battery will not go bad when constantly charged with trickle charger or some other charger which limits the current once the battery is fully charged.


All times are GMT -5. The time now is 11:47 AM.