Hi,

I would like to report a Meridian Flip issue that appears to be related to a mismatch between ASIAIR internal state and the actual mount state (pier side).

Setup

Mount: Sky-Watcher (EQMOD)
Control: ASIAIR
Guiding: enabled (PHD2)
Plate solving: enabled (AutoCenter ON)
Meridian Flip settings:
- Stop tracking: 5 minutes before meridian
- Flip after meridian: 5 minutes after meridian

Observed behavior

During an imaging session:

ASIAIR correctly:
- stopped guiding
- stopped tracking
- entered Meridian Flip waiting phase
After the configured delay, ASIAIR:
- started Meridian Flip
- initiated AutoCenter (plate solving + corrective slews)
However:
- the mount did not appear to be in the correct post-flip geometry
- plate solving succeeded, but reported an initial offset of ~2.7°
- subsequent AutoCenter attempts increased the error instead of reducing it (up to 3.1°)

This suggests that:

the correction vectors were computed assuming a different pier side than the actual one

Critical failure

Because the correction direction was wrong:

each AutoCenter correction moved the mount farther from the target
the system continued trying to converge but diverged instead

When I attempted a manual recovery (GoTo to force flip):

the mount slewed further in the wrong direction
eventually reached the tripod legs (mechanical limit)

Even a “Go Home” did not restore correct behavior immediately.

Full recovery required:

manual reposition to home
restart of the system
fresh initialization

Supporting observations from logs

Before flip:
- PHD2 reports one pier side (e.g. West)
During flip:
- ASIAIR behaves as if flip has occurred (AutoCenter active)
After full manual reset:
- PHD2 reports opposite pier side (e.g. East)

This strongly suggests that:

the physical pier side change did not occur when ASIAIR assumed it did
but ASIAIR proceeded with post-flip logic anyway

Expected behavior

After Meridian Flip trigger:

Mount performs GOTO and confirms pier side change
Only after that:
- AutoCenter should start
- correction vectors should converge (error decreasing)

Actual behavior

AutoCenter starts even if mount state is inconsistent
correction diverges
no safeguard stops the process
leads to potentially unsafe slews

Potential improvement suggestions

Pier side validation before AutoCenter
- verify mount pier side changed before applying corrections
Convergence check
- if error increases after N corrections → abort AutoCenter
Safety limit
- detect large divergence (>2–3° increasing) and stop slewing
Optional debug info
- expose pier side and mount state during flip in logs/UI

Notes

Plate solving itself worked correctly
Issue seems related to state synchronization, not solving accuracy
Behavior is reproducible in principle if mount state becomes inconsistent near meridian

If needed, I can provide:

full Autorun logs
PHD2 logs
timestamps of the flip sequence

Thanks for your work on ASIAIR—this is a critical workflow area and I hope this report helps improve robustness and safety.

Meridian Flip inconsistency: ASIAIR starts AutoCenter but mount not updated?