Skip to content

[bug] EV and Pixii charging anomalities #1975

Open

Description

@segran2

What happened?

EV and pixii charging works much better now after the last couple of weeks of beta updates, but still not entirely stable.

03:03–03:12 | The EV drew 0 W despite a request for 9.66 kW. Charging resumed at approximately 4.72 kW, limited by load balancing.

03:27–03:28 | FTW’s control loop stalled. The safety mechanism attempted to return the devices to autonomous operation; Pixii control was temporarily blocked and subsequently recovered.

04:18–05:12 | The EV showed no power response for approximately 54 minutes. At the start of the incident, Pixii was simultaneously charging at 6.95 kW, while the EV’s pilot current was 0 A.

05:45–05:55 | The EV drew 0 W despite a request for 10.35 kW. The warning cleared when the charging request changed to 0 W; this does not demonstrate that charging resumed.

N.B The 5:45 issue is no issue as the car was charged to 80%. The problem caused by FTW can't get fresh SoC from the car in real time.

Steps to reproduce

Run FTW according FTW plan.

Expected behavior

Charging to be balanced between Easee and Pixii within safe limits and without unexplained interruptions in Easee charging, 0W pilot a.s.o.

Version

v0.159.0-beta.1 + diagnostic patches to catch errors

Deployment

Native 0.x (install.sh)

Hardware / drivers in use

Pixii, Easee, SMA

Relevant logs

Anything else?

ftw-support-20261106-135957.zip

FTW-EV-Pixii-relevanta-incidenter-2026-11-06.zip

Activity

  1. segran2 commented on Nov 12, 2026

    @segran2
    ContributorAuthor

    Follow-up from the night of 12 November 2026 (CET, UTC+1) on FTW v0.163.0-beta.1 with local Easee pilot diagnostics.

    • 03:47–03:52: FTW requested 15 A per phase (10.35 kW), Easee signalled 13 A, and the car drew about 8.91 kW. FTW recorded a load-balancing limit at informational severity. Similar limits were recorded later in the session.
    • 03:26–03:28: Easee reported a 12 A dynamic setting with a 0 A pilot, 0 A allocated on all three phases and 0 W draw. Three resume calls were acknowledged. Pixii was near 0 kW during the retries and the highest measured grid phase was about 6.6 A. The detailed timeline is in #1948.
    • From 04:35: Measured EV power went to 0 W while Easee still reported a positive pilot of 7–12 A in several journal snapshots. FTW alternated between no power response and vehicle declined. The available measurements do not establish why the vehicle stopped drawing power.
    • Control loop: No matching stall or watchdog warning appeared in the FTW service journal for the checked window, 03:00–05:00. That observation is limited to this window.

    The positive-pilot stop remains open for investigation. The next occurrence would benefit from timestamped pilot mode, the Easee reason code, and fresh vehicle SoC where available.

    Sources: FTW control journal, Easee driver logs, Pixii phase-current and battery-power series, and FTW service journal.

  2. segran2 commented on Nov 13, 2026

    @segran2
    ContributorAuthor

    Follow-up from 13 November 2026 (CET) on the Easee installation:

    • 18:00: FTW requested 15 A and Easee acknowledged the resume commands. Easee reported a 6 A pilot and 6 A allocated per phase, while measured EV power stayed at 0 W throughout the two-minute trace. The owner suspects the car had reached its 80% charge limit. There is no contemporaneous SoC measurement in the trace to verify that.
    • 18:29: FTW requested 14 A. Easee again reported a 6 A pilot and 6 A allocated per phase. Measured power became positive after about 6.6 seconds and reached 4.17 kW. Startup traces
    • 18:39: A separate test on the exact Easee Cloud 1.5.0 candidate for device-drivers PR Map port 80 -> 8080 on RPi image + Documentation reshuffle (#199) #200 showed a 15 A request, 6 A pilot and 4.18 kW measured. The driver identified load balancing. Hardware validation

    These runs show that the car can start with the 6 A pilot on this installation. The cause of the earlier stop despite a positive pilot and the separate FTW control-loop stall remain open. The prolonged zero-pilot starts are tracked in #1948.

  3. segran2 commented on Nov 18, 2026

    @segran2
    ContributorAuthor

    Follow-up after reviewing #1948

    Easee Support says Equalizer can take up to 15 minutes to raise the allowed current again after a reduction. That is consistent with the shorter zero-pilot restart delays. The 54-minute no-draw period reported here had a documented 0 A pilot at the beginning. We lack enough observations to attribute the full interval to Equalizer recovery. The next delayed start with zero pilot is tracked in #1948, where the new diagnostics record Pixii phase currents, FTW's command and current ceiling, and timestamped Easee allocation, pilot and power. Easee Support's explanation

    Two observations still need follow-up here:

    • No draw despite a positive pilot. On 12 November, several snapshots showed 7–12 A pilot with 0 W. The 13 November test confirmed that the car can start at 6 A on this installation. On recurrence, capture the pilot mode and reason code alongside fresh vehicle SoC and the vehicle's charge limit when available. The original 05:45 case, when the owner confirmed the car had reached 80%, remains relevant to retry and feedback behavior when FTW lacks fresh SoC.
    • FTW control-loop stall and temporary Pixii control block. Treat this as a separate event. There was no matching stall or watchdog warning in the later checked window, and the original stall remains unexplained.

    Merged PRs #2008 and #2010 improve the pilot and charger-reason feedback. The pending timestamp handling change in device-drivers #183 helps interpret the order of Easee observations in future traces.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions