Reading a battery health report

CurrentLast checked Handover and completenessElectronics

How to read a battery health figure, what the cycle count adds to it, and which thresholds hold.

The reading sequence

4 steps, starting at handover and completeness and running to confirm the figure with a discharge check. Evidence on this point is read at E1, and 2 of the 4 steps carries a condition that ends the assessment outright.

Step 1

Get the report photographed with the unit's own identifier

Make sure the health figure belongs to the unit in the conversation rather than to a screenshot from somewhere else.

Evidence: E1 — one photo of the battery panel on screen with a device identifier visible in the same frame, plus one of the system information panel

Pass
  • The battery panel and a device identifier appear in a single frame.
  • The panel is the operating system's own screen rather than a third-party overlay.
Query
  • The battery panel is shown while the identifier sits in a separate photo.
  • The panel is legible but too small to read the design capacity field.
Return
  • The screenshot shows no identifier and no system interface around it.
  • The figure comes from a utility that does not state how it derives its number.
Step 2

Read the three fields as one set

Read design capacity, current full charge capacity and cycle count together, because a health percentage on its own hides which of the three moved.

Evidence: E1 — the step 1 photo, read at all three fields with the figures written down

Pass
  • All three fields are legible, and the health percentage can be recomputed from the first two.
  • The design capacity matches the capacity this model is specified with.
Query
  • The health percentage is shown without the design and current capacity fields.
  • The cycle count is given as a rounded figure or a range rather than a number.
Return
  • The design capacity field reads zero, or the panel reports no battery data at all.
  • The design capacity is a different value from the one this model is specified with.
Stop here if: The design capacity field reads zero, or does not match the capacity the model is specified with. The percentage printed beside it is then computed from a figure that cannot be checked against anything. What to do next.
Step 3

Place the two figures in a band

Sort the report into one of three bands using the cycle count and the recomputed health together, rather than the percentage alone.

Evidence: E1 — the three fields written down in step 2, read against the bands below

Pass
  • Cycle count under 300 and recomputed health at 90 per cent or above.
  • Both figures fall inside the same band, so neither contradicts the other.
Query
  • Cycle count between 300 and 600, or recomputed health between 85 and 90 per cent.
  • One figure sits inside a band while the other sits just outside it.
Return
  • Cycle count above 800, or recomputed health below 80 per cent.
  • The unit reports very few cycles for its apparent age while health sits in the bottom band.
Step 4

Confirm the figure with a discharge check

Replace the computed figure with a measured one by watching the charge percentage fall under a stated load.

Evidence: E1 — a video of thirty minutes of screen-on use at a stated brightness, with the percentage before and after both visible and a clock in frame

Pass
  • The percentage falls by no more than the capacity figure predicts over that period.
  • The start percentage, the end percentage and a clock all appear in the same video.
Query
  • The video covers ten minutes rather than thirty, which is too short to separate two bands.
  • The brightness is not stated, so the load cannot be repeated by anyone else.
Return
  • The percentage falls about twice as fast as the capacity figure predicts.
  • The unit shuts down while a percentage is still showing on screen.
Stop here if: The unit shuts down with charge still showing, or falls at about twice the predicted rate. The report and the behaviour disagree, and the behaviour is the measurement rather than the estimate. What to do next.

Why the thresholds sit where they do

A battery percentage is a computed figure, not a measurement: it is derived by dividing one reported capacity by another, and both of those numbers come from the same controller. Reading the three fields together is therefore what makes the figure checkable, because a percentage can be recomputed and a design capacity can be compared against the capacity the model is specified with. The 300 and 800 cycle bands and the 90 and 80 per cent health bands are set from the shape of the curve rather than from a single test: capacity loss is slow for the first few hundred cycles and then steepens, so a unit in the middle band is a normal used unit while a unit in the bottom band has moved past the flat part of the curve. The discharge check exists because the controller's own numbers can drift away from behaviour, and half an hour under a stated load is the shortest test that separates the upper band from the lower one.

Common misreads

2 pairs recorded for this point. Each runs in both directions: 1 finding that looks worse than it is, and 1 that looks better.

Check this point against what you were sent

Insufficient evidence

Nothing decided on battery health report reading yet: 0 of 4 steps have been read.

0 of 4 steps judged

One row per step. Evidence levels: E1 seller photos before dispatch, E2 your photos after arrival, E3 platform public information, E4 third-party reference only. This covers one judging point; the sequence as a whole is scored on the photo scorecard.
StepReadingEvidenceStop met
1. Get the report photographed with the unit's own identifierneeds E1—
2. Read the three fields as one setneeds E1The design capacity field reads zero, or does not match the capacity the model is specified with. The percentage printed beside it is then computed from a figure that cannot be checked against anything.
3. Place the two figures in a bandneeds E1—
4. Confirm the figure with a discharge checkneeds E1The unit shuts down with charge still showing, or falls at about twice the predicted rate. The report and the behaviour disagree, and the behaviour is the measurement rather than the estimate.
Score the whole category

Sources and checks

Last checked · next review due · status current.

Compare current electronics listings

Return to Electronics: pre-purchase judging points, or read how verdicts are defined on the method page.