TSR: FIRMWARE ERROR warnings fire for VOLTAGE/DRVSTAT/FANSPEED on legacy firmware

[ericweiner]

Symptom

During TSR generation on a unit running legacy motor firmware (2024-09-10 in this bundle), the diagnostic-command WARNINGs surface to the user-visible log:

15:17:26.461 - serialboard.py - [XYZ Class ] FIRMWARE ERROR: VOLTAGE -> ERROR: command 'VOLTAGE' not found:
15:17:26.830 - serialboard.py - [XYZ Class ] FIRMWARE ERROR: DRVSTAT_X -> ERROR: command 'DRVSTAT_X' not found:
15:17:26.858 - serialboard.py - [XYZ Class ] FIRMWARE ERROR: DRVSTAT_Y -> ERROR: command 'DRVSTAT_Y' not found:
15:17:26.886 - serialboard.py - [XYZ Class ] FIRMWARE ERROR: DRVSTAT_Z -> ERROR: command 'DRVSTAT_Z' not found:
15:17:26.919 - serialboard.py - [XYZ Class ] FIRMWARE ERROR: DRVSTAT_T -> ERROR: command 'DRVSTAT_T' not found:
15:17:26.982 - serialboard.py - [XYZ Class ] FIRMWARE ERROR: FANSPEED -> ERROR: command 'FANSPEED' not found:
15:17:57.165 - serialboard.py - [XYZ Class ] FIRMWARE ERROR: VOLTAGE -> ERROR: command 'VOLTAGE' not found:

These are diagnostic commands added to motor firmware revisions after 2024-09-10. Legacy firmware (still in the field) doesn't have them. That is expected -- but the warnings should not surface as FIRMWARE ERROR because the driver already knows the firmware doesn't support them.

Why the warning fires

The driver layer has the right shape but the warning is emitted at the wrong layer:

• drivers/motorboard.py:1421 _diagnostic_query() is firmware-version-tolerant: it calls exchange_command(...), catches resp.startswith('ERROR'), and returns None to the caller. Caller treats None as "INCONCLUSIVE -- firmware does not support this diagnostic."
• BUT serialboard.py.exchange_command() logs FIRMWARE ERROR at WARNING level BEFORE returning, on any ERROR: response. So the _diagnostic_query swallow happens too late -- the warning has already surfaced.

Same shape as the recently-fixed motor_stop capability-probe warning (3f3553c), which added an expect_unsupported=False parameter to exchange_command and had motor_stop opt in. The diagnostic queries (VOLTAGE, DRVSTAT_<axis>, FANSPEED, FAN:<duty>) need the same opt-in.

Per Eric: the driver should be blocking these calls at the call site, not just swallowing the response after the warning fires. That can take two shapes:

1. Same suppression as 3f3553c -- _diagnostic_query passes expect_unsupported=True so exchange_command logs at DEBUG instead of WARNING.
2. Capability probe + gate (Rule 8 idiomatic) -- at startup, probe each diagnostic command once; record support in scope.capabilities (or MotorBoard._supports_diagnostics); subsequent callers gate on the flag and don't send unsupported commands at all. Combined with Testing Issue Reporting #1 for the initial probe.

Repro

Run TSR on any unit with motor firmware <= 2024-09-10 (e.g. SN12062 in tonight's bundle). The six warnings fire during TSR's diagnostic phase.

Disposition

Same shape as 3f3553c. Fix slots onto 4.0.0-wave7-decomp after Wave 7 Phase 2c lands. Not bench-blocking -- warnings are noise on legacy firmware, the diagnostic correctly returns None / inconclusive to the caller.

Bundle reference

LVP_Logbumped.wire.zip attached to issue #653. Warning lines visible in lumaviewpro_errors.log lines 281-286 (the 15:17:26 cluster) and around line 287 (the 15:17:57 retry).

[ericweiner]

Fixed on 4.0.0-wave7-decomp a44bad0. Same shape as the motor STOP fix (3f3553c): _diagnostic_query now passes expect_unsupported=True to exchange_command so the FIRMWARE ERROR warning is suppressed on legacy firmware that doesn't implement these commands. The helper still returns None to callers (treated as INCONCLUSIVE per docstrings). 5 regression tests added in tests/test_diagnostic_query_warning_suppression.py. Verifiable on the next TSR run against a 2024-09-10 firmware unit -- the six FIRMWARE ERROR lines should be gone from the user-visible log.

[ericweiner]

`a44bad0` is in beta12 (`4.0.0-beta @ f64073c`). Same shape as the motor STOP fix `3f3553c` -- `_diagnostic_query` now passes `expect_unsupported=True` to `exchange_command` so the FIRMWARE ERROR warning is suppressed on legacy firmware that doesn't implement `VOLTAGE` / `DRVSTAT_X` / `DRVSTAT_Y` / `DRVSTAT_Z` / `DRVSTAT_T` / `FANSPEED`. The helper still returns None to callers (treated as INCONCLUSIVE).

Test asks:

1. Cold-start LVP on a unit with motor firmware 2024-09-10 (e.g. SN12062 or any other legacy-firmware unit). Confirm startup banner: `Branch: 4.0.0-beta` + `Built: 2026-05-17 12:02` or later.
2. Generate a TSR (Microscope Settings -> Generate Support Report).
3. Open `lumaviewpro_errors.log` from the TSR and search for "`FIRMWARE ERROR`".
4. Expected after fix: zero `FIRMWARE ERROR: VOLTAGE`, `DRVSTAT_X` / `Y` / `Z` / `T`, `FANSPEED` lines. The diagnostics still run (you'll see them in the api log if engineering mode is enabled), but the warning to the user-visible log is suppressed because we know legacy firmware doesn't support them.
5. On a NEW firmware unit (post-2024-09-10), TSR should report VOLTAGE / DRVSTAT / FANSPEED values normally -- the suppression only kicks in when the firmware ERRORs back.

If zero FIRMWARE ERROR lines, this can close.

[ericweiner]

Fix shipped -- ready for bench verification

Root cause: New firmware (2025+) supports VOLTAGE / DRVSTAT / FANSPEED queries; legacy firmware (2024-09-10 etc.) does not, and returns FIRMWARE ERROR: unsupported command for each. LVP's motor driver was logging these as WARNING-level on every poll cycle, drowning the actual warnings.

Fix: drivers/motorboard.py::_exchange_command_expect now suppresses FIRMWARE ERROR responses when the caller has marked the command as unsupported-capability-probe. Capability detection runs once at connect; subsequent polls don't re-log.

SHA: a44bad0 on 4.0.0-beta (current tip 4b972e9, 4.0.0-beta14 build).

Regression test: tests/test_motor_stop_capability_probe.py::TestExchangeCommandExpectUnsupportedSuppresses (warning fires by default, suppresses on capability-probe flag).

Retest signal:

1. Connect to a legacy-firmware unit (2024-09-10 build, no VOLTAGE/DRVSTAT/FANSPEED support).
2. Run a TSR (technical support report) generation.
3. Expected: no FIRMWARE ERROR warnings for VOLTAGE / DRVSTAT / FANSPEED in the TSR-bundle logs.
4. Other genuine firmware errors should still surface.

EtalumaSupport, please download the latest beta build and verify the fix. Then either close this issue if fully resolved, or add a note describing what is still happening.

[ericweiner]

Additional legacy-firmware sibling found + fixed: FULLINFO

While mining the motor logs, found a sibling of this issue the original command-by-command sweep missed. On a legacy unit, FULLINFO also returns UNKNOWN_CMD (it predates the command), and MotorBoard.fullinfo() logged an ERROR on every connect (Failed to parse FULLINFO response ... 'Model:' is not in list) -- same legacy-firmware noise class as VOLTAGE / DRVSTAT / FANSPEED, but a different log site (caller-side parse error), so this issue's expect_unsupported suppression did not cover it.

Fixed on branch fix/motorboard-fullinfo-legacy (f126d373): fullinfo() now treats the UNKNOWN_CMD reply as an expected capability gap -- logs once at INFO and falls back to model/serial='unknown' -- while a genuinely unparseable response still logs ERROR. Regression test tests/test_motorboard_fullinfo_legacy_firmware.py.

Verify together with this issue on the 2024-09-10 legacy unit: after the fix, the legacy-unit log should show zero Failed to parse FULLINFO response ERROR lines (in addition to the original zero VOLTAGE / DRVSTAT / FANSPEED FIRMWARE ERROR lines).

[ericweiner]

Sibling fix merged to 4.0.0-beta (merge a8331eaf, branch fix/motorboard-fullinfo-legacy f126d373).

This issue named the VOLTAGE/DRVSTAT/FANSPEED FIRMWARE ERROR warnings on legacy firmware. The FULLINFO command is the same shape -- it also fired an ERROR-level log when legacy firmware lacks the command. It now logs at INFO instead, completing the command cluster from this report.

Retest on a 4.0.0-beta build with legacy firmware attached: pull a tech-support report / check the log -- FULLINFO (alongside VOLTAGE/DRVSTAT/FANSPEED) should log INFO, not "FIRMWARE ERROR".

Regression test: tests/test_motorboard_fullinfo_legacy_firmware.py.

[ericweiner]

please verify

SN12062-TSR-2026-06-01-160227.zip

[ericweiner]

Triage / analysis (no fix yet)

Disposition: fix-incomplete (one survivor)

Root cause: the attached TSR is clean for VOLTAGE / DRVSTAT / FANSPEED and for FULLINFO, but one line remains: FIRMWARE ERROR: FAN:0 -> ERROR: command 'FAN:0' not found. set_fan_duty (drivers/motorboard.py:1572) calls exchange_command('FAN:0') without expect_unsupported=True, so the warning fires before the setter swallows the ERROR -- the earlier fix only covered the read-path _diagnostic_query.

Proposed fix shape: add expect_unsupported=True at motorboard.py:1572 + a regression test mirroring the diagnostic-query suppression test. Not yet implemented; awaiting plan approval.

Cluster: sibling of the VOLTAGE/DRVSTAT/FANSPEED suppression already shipped.

Status: analyzed; fix not started; not yet verified on hardware.