When querying battery capacity vis UPS-Link / Smart Protocol what factors does it take into account?

[Anonymous user]

From ftp://ftp.apc.com/restricted/software/ups-link/100/UPS-Link_Protocol_Specification.pdf, what factors does the "f" / battery capacity command query take into account?

In the documentation for the Smart-UPS : http://www.apc.com/salestools/AKOV-699KQK/AKOV-699KQK_R0_EN.pdf

it mentions that :

Smart-UPS is continuously calculating the remaining runtime based on the battery capacity, age, load, etc.

Does the battery capacity reported by the UPS firmware automatically take into account age, load, etc. or is that a higher level function of the PowerChute Business software? 

Does Back-UPS Pro 1300/1500 account for these factors (in particular age) also or is that a feature only of Smart-UPS / PowerChute Business?

[voidstar_apc]

Battery estimation is done by the UPS, not PowerChute.

There are three variables to consider: battery capacity, battery state-of-charge, and estimated runtime.

APC previously used the term 'capacity' to refer to state-of-charge where 100% refers to the point where the battery is full. As the battery ages, full means less and less total charge. I don't think there's much smarts here... until a battery cell fails, the state-of-charge of a lead-acid battery can be determined by its voltage.

Estimating the runtime (ie, total amount of charge it can supply at a given load / load) is where the secret sauce is, and it depends on the specific UPS. This estimate updates during battery discharges and is most accurate during this time. Which works out because that's when its being used to trigger safe shutdowns.

If the battery ages faster than predicted but still passes self tests and isn't being discharged, the estimate may be too optimistic. So you can replace batteries when they fail self-test, but the more risk averse also replace batteries after 3-5 years regardless. Newer SmartUPSes have a battery lifetime warning to help with this.

[Anonymous user]

Thank you for the detailed answer!

So it sounds like a better metric to monitor is runtime.  I am writing a custom software application and would like to use the UPS-Link / Smart Protocol commands to monitor the system.  It sounds like the "j" / Run Time Remaining command would give a better estimate based on battery capacity / load.  Our team is worried about what happens when the battery ages.  As a matter of policy, I guess we could conservatively replace every 3 years, however, is it good enough to rely on the Run Time Remaining value reported by firmware?  Basically if it goes below a certain level, the application won't run unless the UPS is replaced.

With respect to the Run Time Remaining command, in order to get an accurate estimate, do we need to upgrade to SmartUPS or is the Back-UPS value also considered as reliable even as the battery ages.

[voidstar_apc · Accepted solution]

I think you want to look at:

1) Is the battery bad? (ie, failed the bi-weekly self-test)

2) Is runtime below the minimum needed to shut down attached loads safely?

and optionally:

3) Has the battery reached the end of its reliable service life and should be proactively replaced?

Self-test is designed the flag the battery early so you have time to replace it.

I don't think SmartUPS vs BackUPS matters too much. In both cases, the only time the UPS is measuring the actual runtime of the battery is during a battery discharge, otherwise it's an estimate. If power is good and the UPS rarely goes to battery, then the runtime estimate can diverge from the true runtime. If the UPS is lightly loaded, then self-test may not sufficiently stress the battery and proactive replacement makes a lot of sense.