I have honestly searched before posting, but most conversations I found were about the automated meridian flip while tracking, whereas my case is about slewing to targets.

In simple terms the situation is: Northern hemisphere, GEM with an OTA on the West SOP pointing East.
My expectation is that as long as my targets are on the East of the meridian, there is no need for flipping the OTA to the other side. Is that not the case?

The log I am attaching shows 5 minutes where I start Ekos, perform an initial sync, and on the second slew towards West but well before the meridian it tries to perform a flip and approach the target from across the meridian. At the time of the test at my location the Southern meridian was at around RA 21h 25m.

• [2023-08-18T23:22:29.078 JST INFO ][org.kde.kstars.ekos.align] - "Syncing to RA (00h 41m 33s) DEC ( 56° 40' 51\")" — initial sync to Schedar
• [2023-08-18T23:23:26.520 JST DEBG ][org.kde.kstars.indi] - ISD:Telescope sending coords RA: "00h 44m 02s" ( 0.734017 ) DE: " 41° 23' 48\"" ( 41.3968 ) — successful slew to M31, which was just 3m towards East, but I wanted to make sure GoTo is working, which is confirmed by an immediately following solve/sync
• [2023-08-18T23:24:00.940 JST INFO ][org.kde.kstars.ekos.align] - "Solution coordinates: RA (00h 43m 42s) DEC ( 41° 23' 42\") Telescope Coordinates: RA (00h 44m 03s) DEC ( 41° 23' 48\") Target Coordinates: RA (00h 44m 02s) DEC ( 41° 23' 48\")"
• [2023-08-18T23:25:10.382 JST DEBG ][org.kde.kstars.indi] - ISD:Telescope sending coords RA: "22h 22m 01s" ( 22.367 ) DE: " 46° 39' 18\"" ( 46.655 ) — this is the slew towards West, but although the target is well before the meridian, the mount starts performing the flip, which I had to abort in order to save the equipment.

I am seeing exactly the same behavior with the mount simulator: it almost looks like it (whatever <em>it</em> is) thinks the meridian is somewhere just East of the 00h RA line.
Interestingly, if I "scroll" time in Kstars to a point where 00h RA is West of the meridian, the telescope simulator doesn't even bother flipping over the pole to cross the meridian. But once I ask it to slew past the 00h RA further West, it goes all the way around.

Here is a short video clip showing how the mount simulator flips around a line just East of 00h RA instead of the meridian.


Am I missing something basic here? Any hint would be greatly appreciated!

Is it not the mount that initiates the meridian flip and decided when it does it ?


Sent from my iPhone using Tapatalk

Kirill,
I have a Losmandy G11 with Gemini 2 and your understanding is not how it works for my set up.
The mount controls what it does and needs to be configured to determine when and how a meridian flip will happen to ensure that nothing collides with the tripod or pier, which will generally be different for every setup.
You also need to configure Kstars/INDI to compliment the mount settings.
All Kstars does is issue a slew command and the mount responds by determining whether it will go direct or flip.
There is a setting on the Ktrars mount tab that dictates when it should issue a slew command to force a flip when tracking, it slews to the current location, but if the mount does not flip to reach the current position, it will be a problem. Kstars needs to issue the slew command in the range where the mount will flip, but if if does so too early or too late nothing will happen.

Paul

Thank you for chipping in. I have updated my original post at the top to show a video of the simulator exhibiting the behavior I am questioning.
You will see that

• the first small slew West goes the regular way
• the second slew West (still a bit East of 00h RA and <strong>well</strong> East of the meridian) makes it flip
• and after it flips, it knows that it is on the East side of pier now and casually crosses the meridian without any further flips

So yes, it does seem like there is some setting somewhere that is making it behave this way, but what setting is that and where?

To add to my confusion, today (just 8h after I asked the original question) the simulator is behaving the way I expect it to: flips around the meridian itself and not some line East or West of it. It does not depend on whether the Auto MF is enabled or not in the Ekos Mount tab, I have just checked both cases.
But even now, there still seems to be some buffer zone around the meridian, when slewing into which from the East the mount flips. It's just that right now I am observing that zone to be around the meridian, and yesterday it seemed to be around 00h RA (as in my original video clip).

Please compare the video above with this one:


I found a simulator setting called Flip Posn. in the Simulation tab in the Indi control panel. When set to a positive number, it seems to be shifting the flip zone to the West of the meridian, but when it is set to 0, the "zone" begins a little to the East of the meridian as on the video above.
There is no such setting on my real mount's driver control panel though.

If you point to zenith, where do you end up?

If I do it here, the FOV is centered 1 RA minute (15 arc minutes) east of the meridian line. So if your experiment relied on the meridian line, the question is: Is that line drawn correctly?
Did you check hour angle at the position where you slewed to?

Hi Peter, thank you for the attention to this case.

First of all, how do you point the telescope to zenith? If I press Z on the keyboard, the map view moves to zenith, but nothing is selected as a target that I could tell the telescope to slew to.

But so far everything, including previous responses here point to the fact that KStars indeed has nothing to do with any flippings of the mount. Instead, it does just 3 things:

1. it tells the mount the observing site coordinates
2. it tells the mount the local sidereal time
3. it tells the mount the slew target coordinates

The mount then decides itself how to perform the slew. Keeping this in mind I am trying to understand what could cause the observed behavior and why I observed it with both my real mount and the simulator after that (also why the simulator started behaving "normally" the following morning).

The logs look like both the location and LST were set to the mount correctly. In fact, LST is being assigned multiple times along the way, such is the Temma protocol, apparently.

But your question of whether the meridian line is drawn correctly is a good one. Check this out:

The RA coordinate of the meridian line (below in the status bar) and the ST in the time info box above mismatch by more than a minute. And I am close to the maximum zoom possible in KStars here.

However, originally my problems with the mount started when I was trying to slew it to my photography target, which was still a couple of hours East of the meridian.
It is also the case in the video attached to my first post here. The slew target is more than 2h before the meridian, yet the simulator (!) goes around.

The mystery has been solved.

Ultimately the central problem manifesting itself in various weird behaviors was that the system time on the INDI server host computer was not being set correctly.

In my setup of Raspberry Pi running INDI and controlling all hardware + laptop running KStars as the control terminal it was either of the following situations:

1. "KStars updates all devices" option was selected in the Time & Location Updates configuration of KStars, and indeed the laptop date & time were being correctly assigned as current in both KStars and the mount driver in the INDI control panel, but the system time on the Raspberry Pi was still incorrect
2. "GPS updates KStars" option was selected, and indeed whenever I had a good GPS fix, the correct date & time were being assigned to both KStars and the mount driver, but the system time on the Raspberry Pi was still incorrect

In relation to this very thread, where I was originally questioning the seemingly weird mount behavior as it was flipping around lines other than the meridian, the underlying reason can be summarized as follows:

• Given a slew (go-to) command, the mount decides whether to perform a flip or not based on whether the target coordinates are on the other side of where the mount "thinks" the meridian is or not
• The Temma protocol with which my Takahashi mount is controlled assumes that the current Local Sidereal Time be explicitly assigned to the mount immediately before every slew command
• To calculate the current LST the INDI Temma driver uses a libnova function that relies on the current OS time

So in effect, incorrect OS time was resulting in incorrect LST being assigned to the mount by the driver, the mount calculating the meridian using incorrect input and flipping around lines corresponding to results of those calculations.

I managed to resolve the issue completely in the situation with using the GPS module and making sure the indi_gpsd binary is allowed to update system time, the details are here: indilib.org/forum/general/13613-gps-nmea...date-time.html#95481.
Another method to set system time based on GPS readings has been proposed here by using a Python script that queries GPSD, but that script would need to be given elevated priviliges.

For the case when "KStars updates all devices" I have verified that the OS date & time on Raspberry Pi are not synchronized and stay incorrect, and unlike the case above, there are no error messages anywhere, so I am not sure KStars is even trying to set that time. I guess if needed, I would just manually set the time on Raspberry Pi using the date command.