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:

1. ASIAIR correctly:

   • stopped guiding

   • stopped tracking

   • entered Meridian Flip waiting phase

2. After the configured delay, ASIAIR:

   • started Meridian Flip

   • initiated AutoCenter (plate solving + corrective slews)

3. 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:

1. Mount performs GOTO and confirms pier side change

2. 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

1. Pier side validation before AutoCenter

   • verify mount pier side changed before applying corrections

2. Convergence check

   • if error increases after N corrections → abort AutoCenter

3. Safety limit

   • detect large divergence (>2–3° increasing) and stop slewing

4. 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.