Hours to Seconds

A Backup Rating in Hours, a Discharge Log in Seconds

Backup power is sold in hours and proved in seconds. The datasheet promises four hours of autonomy at half load; the acceptance test that has to demonstrate it is a logger sampling pack voltage and current once a second until the inverter drops out, and what it hands back is a row count. Somebody then has to say whether 13 140 samples is a pass or a failure, which means putting a rating written in hours and a measurement written in seconds on the same axis.

Conversion factor: 1 h = 3 600 s, so multiply hours by 3 600. The standard 15-minute UPS rating point is 0.25 h, or 900 s; a four-hour telecom autonomy is 14 400 s, and a test that cut off at 13 140 s delivered 3.65 h — nine per cent short of the rating.

Why the Two Units Belong to Different People

Ratings are quoted in hours

Capacity is an ampere-hour figure and autonomy is stated against it, so the specification, the tender and the site drawing all count in hours or in whole minutes.

Evidence arrives in seconds

Battery monitors, electronic loads and inverter event logs timestamp per second, so the proof of a rating is a duration in seconds with no unit conversion done for you.

The C-rate is a time in disguise

A discharge rate of 1C empties the nominal capacity in one hour, so the rate is simply the reciprocal of the autonomy in hours: 4C is a quarter of an hour, C/8 is eight hours.

Fast discharge destroys capacity

A pack pulled hard returns far fewer ampere-hours than its label claims. That is why short autonomies are rated separately instead of being divided out of the 20-hour figure.

Checking a Measured Discharge Against the Rating

The workflow below is what happens on commissioning day, when a load bank is connected and the question is whether the installed pack does what the specification says.

Put the rated autonomy in first

Type the specification figure — 0.25, 0.5, 4 — into the hours field and note the second count you are aiming at. A comma is accepted for the decimal point, so 0,25 pasted from a datasheet works unedited.

Check what discharge rate that implies

Divide 1 by the autonomy in hours for the nominal C-rate. A ten-minute ride-through is roughly 6C, a rate at which a lead-acid pack returns barely half its twenty-hour capacity — the rating has to come from the manufacturer's short-duration table, not from arithmetic.

Reverse it for the logged result

The swap button (↔) turns the page into s → h so the run length the logger recorded reads as the hours the report has to quote. Spaces used as thousands separators in a pasted figure are ignored.

Copy the plain number into the test record

The copy control on each field gives the digits alone, with no unit and no spacing, which is what a commissioning spreadsheet or a monitoring threshold field expects. Ctrl + C inside a field does the same.

A converted number is not a prediction: turning 4 h into 14 400 s says nothing about whether the pack will last that long. Temperature, age and discharge rate all move the real end point, and only a discharge test settles it.

Autonomy Targets and the Discharge Rate They Demand

The autonomy figures that turn up in specifications, each with the second count a logger would record and the nominal discharge rate implied by it. The rate is simply the reciprocal of the hours, which is why short backup times are so much harder on a battery than the ampere-hour label suggests.

Backup scenario	Autonomy (hours)	Autonomy (seconds)	Nominal discharge rate
Ride-through until the generator picks up	≈0.0833 h (5 min)	300 s	≈12C
Graceful shutdown of a server rack	≈0.1667 h (10 min)	600 s	≈6C
Standard UPS rating point	0.25 h	900 s	4C
Extended cover for short outages	0.5 h	1 800 s	2C
One-hour reference discharge	1 h	3 600 s	1C
Telecom cabinet or remote site	4 h	14 400 s	C/4
Healthcare and life-safety loads	8 h	28 800 s	C/8
Capacity rating basis for lead-acid	20 h	72 000 s	C/20

The bottom row is the one that causes trouble. A pack labelled 100 Ah is usually labelled at the twenty-hour rate, meaning 5 A for 72 000 s. Draw 50 A from it and the naive answer of two hours is badly wrong: with a Peukert exponent of 1.25 the pack lasts about 1.12 h, roughly 4 050 s instead of 7 200. Autonomy at high rates has to be read from the manufacturer's short-duration table or measured, never divided out of the headline capacity.

What This Pair Does on Commissioning Day

Rating and log in the same window

Both fields stay live as you type, so a specification in hours and a discharge duration in seconds can be compared without pausing the test to open a calculator.

Turn a logged cut-off into report hours

Swapping the pair converts the raw second count from a battery monitor into the hours a commissioning certificate or an asset record is written in.

Minutes for the ride-through numbers

Every time unit in the app is searchable on both sides, so a five- or ten-minute ride-through can be entered as minutes rather than as an awkward fraction of an hour.

Short transfer windows stay readable

Results keep up to eight decimals, so a 10 ms static-switch transfer expressed in hours does not collapse to zero when it sits beside a four-hour autonomy.

Questions From the Battery Sizing Review

What does a C-rate tell me about how many seconds the pack holds up?

Directly, if you read it as a reciprocal. A rate of 1C empties the nominal capacity in one hour, so 2C is half an hour or 1 800 s, 4C is 900 s and C/8 is 28 800 s. The catch is that the nominal capacity was measured slowly, and a pack cannot deliver the same ampere-hours when it is being pulled hard. Treat the C-rate as the time you would get from a perfect battery and the manufacturer's discharge table as the time you will actually get.

Why did the discharge test stop 1 260 seconds short of the rated four hours?

Three usual suspects, and they stack. Temperature is the first: capacity falls away below the 25 °C the rating assumes, and a cold plant room can cost a tenth of the run. Age is the second — a battery is generally considered end-of-life once it delivers 80 % of rated capacity, which turns 14 400 s into 11 520 s while the label still says four hours. The third is the end-of-discharge voltage the test used; stopping at a higher cut-off point ends the run earlier and leaves usable energy in the pack. A shortfall of 1 260 s, 8.75 % of the target, is well within what a mildly aged string does on a cool day.

Does doubling the ampere-hours double the autonomy?

Only if the discharge rate stays where it was. Doubling capacity while the load is unchanged halves the effective C-rate, which not only doubles the run time but recovers some of the capacity that high-rate discharge was throwing away — so the gain is usually a little better than double. Doubling the load instead of the capacity is the mirror image and much worse than halving: at a Peukert exponent of 1.25, going from 25 A to 50 A on a 100 Ah pack takes the run from about 9 630 s to about 4 050 s, a factor of 2.4 rather than 2.

Why do monitors log in seconds when nobody specifies backup that way?

Because the interesting parts of a discharge are short. The voltage dip at the moment the load is applied, the recovery, and the knee at the end where terminal voltage falls off a cliff all happen inside a few seconds, and a one-second sample interval is the coarsest that still captures them. A four-hour test at 1 Hz produces 14 400 rows, which is trivial to store and gives the analysis room to find the knee. The hours only appear again when the result is written up.

How much autonomy do we need before the generator carries the load?

Enough to cover start, stabilise and transfer, plus a retry. A standby set typically reaches usable output in ten to thirty seconds, but the sizing figure people actually install is five to ten minutes — 300 to 600 s — because the first start attempt can fail, the transfer switch has its own timers, and someone may need to be phoned. Where there is no generator at all, the target changes shape completely: the battery has to outlast a graceful shutdown of everything on it, which is a fixed number of minutes derived from the slowest service, not from the outage.