[frahlg]

Problem

v0.155.1-beta.1 (#1804) runs Charge now on one phase for a charger that cannot run three, and the API now reports phases: 1 for it. The slider's top was still wrong. It took its range from max_charge_w, but Core treats a hold as a ceiling and floors it to allowed_steps_w within max_charge_w (clampManualPower → floorChargeW).

A one-phase charger with max_charge_w: 7360 and allowed_steps_w: [0, 1380, 2300, 3680]:

	Slider	Readout at the top	Status after Start	Core's order
master 16c4ad0	6–32 A	32 A · 7.4 kW	32 A requested	3,680 W
this branch	6–16 A	16 A · 3.7 kW	FTW received 16 A	3,680 W

Change

evChargeNowTopW(lp) in web/app.js returns the highest step at or below max_charge_w, or max_charge_w when there is no table. The slider's top, its readout and the idle line ("Starts at up to …") use it. Amps are still W / (V × phases), with the phases the API reports. The bottom of the range and the sent phase_mode do not change.

A three-phase table that goes past the maximum now stops at the step Core can order: table [0, 4140, 6900, 11040] with max_charge_w: 7360 reads 10 A, where master read 11 A.

Evidence

• web/ev-charge-now-range.test.mjs builds the real buildManualChargeSection with a small fake DOM, reads the slider and readout, and clicks Start. On master, 3 of 4 cases fail (max 32 against 16, and 11 against 10). The no-table case passes on both, as it should.
• npm test: 874 pass, 0 fail.
• Browser, local rig (sim-ferroamp, Core on this branch, sim-ocpp easee-charge-up -plug, the loadpoint above, 25 A fuse): Overview → Check charging and battery level → Charge now. Before: slider max 32, "32 A · 7.4 kW", banner "32 A requested". After a reload: max 16, "16 A · 3.7 kW", idle line "Starts at up to 16 A · 3.7 kW". A new Start reads "FTW received 16 A", and /api/loadpoints shows requested_w: 3680, requested_a: 16. No console errors apart from the dev build's version check (503).

Overlap

Draft #1792 (two-phase charging) changes the same maxA line. Whichever lands second rebases: #1792's Math.floor for two phases applies to wToA(evChargeNowTopW(lp)).

Paired follow-ups, not in this PR

The other clients copy the box page's range and need the same change. Each one also has to read allowed_steps_w, which neither parses today.

• ftw-webapp: chargeCurrent in src/lib/format/ev.ts.
• ftw-app (iOS): EV.current(_:) and EV.watts(_:amps:) in appleApp/FTWKit/Sources/FTWKit/Format/EV.swift. Android has no Charge now.

Known limits, left as they are

• Amps between steps. Core floors the hold, so on the table above, 7–9 A orders 6 A and 11–15 A orders 10 A, while the readout shows the slider's amps.
• A table not on whole amps at 230 V. Easee's 11,100 W at three phases reads 16 A, but 16 × 690 = 11,040 W floors to the step below. Master does the same with that maximum.

🤖 Generated with Claude Code

The Codex finding, fixed in d120b27

At the top of the slider, Core now orders the top step exactly. Before the fix, a top step that is not a whole number of amps rounded down, and Core floored the order to the step below. Three phases, 230 V, steps ending 9,700 / 10,400 W:

	Readout at the top	Sent at the top	Core orders
bde379c	15 A · 10.3 kW	10,350 W	9,700 W
d120b27	15 A · 10.4 kW	10,400 W	10,400 W

One step below the top, 14 A sends 9,660 W as before. Core floors that to 6,900 W, and the readout still says 9.7 kW: the between-steps limit this PR already lists. npm test: 877 pass, 0 fail, including three new cases for this example.

Rendered with the real buildManualChargeSection and the box's stylesheets. From the top: the Codex example; the same charger with Charge now running at the top; the one-phase charger; a charger with no step table:

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-22T14:32:01.219900Z	bde379c	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: bde379c139

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Preserve access to the highest step when converting to amps

When the highest table entry is below max_charge_w and rounds down to whole amps, this change makes that previously reachable charging level impossible to request. For example, at three phases and 230 V, a maximum of 11,000 W with steps ending in [9700, 10400] previously submitted 11,040 W, which Core clamped to 10,400 W. The new maximum is 15 A, submitting only 10,350 W; clampManualPower → floorChargeW then selects 9,700 W, and the slider cannot go higher. Preserve the exact top-step wattage in the submitted hold or otherwise ensure the amp conversion does not lower the ceiling below that step, and add a non-whole-amp regression case.

Useful? React with 👍 / 👎.