SESSION 14 · 10/26 · PARTLY WORKED
First light for the 127, with the mount asleep
The 127 Mak's first frames of sky: an 8% crescent at 15 degrees, hand-pointed and untracked, craters along the terminator in a 50-frame PNG run. The AZ-GTi never spoke to the laptop all night - its WiFi had quietly gone back to hotspot mode - so rig-ui could not jog, track or centre anything, and the Moon crossed the sensor in forty seconds.
The 127 Mak on the AZ-GTi, SV205 on the finder and then on the Mak. Laptop inside, scope out. Written up the morning after.
The mount was the night's problem before the night started. rig-ui came up at 20:15 with "no reply from the mount". The AZ-GTi had been on the home wifi the last time it was used; tonight it was broadcasting its own hotspot, SynScan_4a6d, and nothing else. Sky-Watcher's WiFi module drops back to hotspot mode on its own if nothing connects to it for long enough, and that is what had happened. The SynScan app could reach it directly through the hotspot, but I could not get station mode to take again, and the laptop can only be on one network at a time. So for the whole session the page could show me the camera and nothing else: no jog, no track, no tap-to-centre, no search.
Why I could see nothing while the focus score was high
The first live frame was black except for a saturated white blob in the bottom-left corner. rig-ui's focus score sat at about 22, which on the 102 would have meant close to focus. It meant nothing. The score only looks at pixels brighter than level 60, and the only such pixels were a burnt-out bloom - it was measuring the noise around a clipped edge. Later, with the crescent centred in the finder and the limb clipped at 255, the same score read over 1100. Rule from tonight: when the histogram shows clipping, the focus score is not a focus score.
The Moon in the finder
The SV205 was on the finder, and the crescent sat nicely centred with the earthshine showing. rig-ui had no finder set, so it applied the Mak's 0.193"/px to a view that is really about 2.4"/px: the disc was some 780 sensor pixels across, which is 31 arcminutes at the finder's scale and two and a half at the Mak's. Tap-to-centre and the search spiral would have stepped a dozen times too far. With the mount dead it did not matter, but it would have.
First light on the Mak
Moved to the Mak by hand. At 0.193"/px an untracked Moon runs about 80 pixels a second and crosses the sensor in forty, so it was centre, shoot, re-centre. The 27.8 ms live view was clipping the limb; the capture went out at 21 ms.
| time | run | frames | exp | notes |
|---|---|---|---|---|
| 20:51 | 20h51m10s | 50 PNG | 21 ms | 8% crescent, alt 14.9, WSW, Sun at -17.4. Under the "Skywatcher 102" 1300 mm profile - the 127 has no profile yet. |
Stacked the sharpest 25 of 50, aligned on the green channel: 157 pixels of drift over the
run, absorbed cleanly. Craters resolved all along the terminator and the limb is sharp.
Strong atmospheric dispersion at 15 degrees, orange on the limb and blue on the terminator;
the green-channel mono render removes it. The stack and its linear .npy are in the
night's keepers/.
What I got wrong
- I let the mount go idle on its hotspot timer. The WiFi config does not survive being ignored. Connect the app to it soon after powering on, every time.
- I trusted the focus number on a clipped frame. Twice.
- I did not tell rig-ui which camera was the finder. The scale it used was twelve times wrong and nothing on the page said so.
- AstroDMx still thinks the 127 is the 102. The log's focal length, FOV and limits are all for 1300 mm.
Left for next time
- Get the AZ-GTi back on the home wifi. SynScan Pro over the hotspot, Wi-Fi Setting, Modify Station: SSID, password, keep the access point on, Apply, power-cycle. If it will not join, look at the router's 2.4 GHz security mode. Factory reset of the WiFi, if it comes to that, is four hours powered up with nothing connected.
- A hotspot mode for rig-ui and the
scopelauncher, so a mount that has forgotten its network can still be driven from the phone. - Grey out the focus score when the lit region is clipped, and show the clipped percentage next to it.
- A 127 profile in AstroDMx at 1500 mm, and the SV205's scale on it measured rather than assumed.
- The SER stacker cannot read PNG runs. Tonight's stack came from a scratch aligner.
- The Xeon's RAID had been unmounted since its reboot on 10 October, and the site had been showing no nights at all. Mounted again and the archive went through; worth a health check that actually complains.
