×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

Eqmod telescope pier side

  • Posts: 210
  • Thank you received: 104
I have a problem with the actual implementation of TELESCOPE_PIER_SIDE in the emod driver, as it is based on HourAngle-6h it is wrong when you point below the pole.
For example you start guiding on M101 when it rise on the N-E, pier side is wrongly reported as PIER_EAST. A few hours later when it cross the HA limit (east of Polaris) it suddenly change to the right value PIER_WEST, but the scope is really always pointing in the same direction.
I have implemented dec guiding reversal based on pier side in the Phd2 Indi interface but this problem make it dangerous to use.

In the attached proposed patch I change the way to get pier side, no more form HA, but directly from the DEC movement direction (as defined in Ascom), this way the reported value is always right and this fix my problem.
I also add an internal Unknown state to be sure the first read side is reported to the property.
I make basic testing with the eqmod simulator in both Northern and Southern hemisphere and it look good.

Can someone with a better knowledge of the eqmod driver please review this patch ?
My doubt are :
do I get the dec direction at the right place ?
can this change have any side effect at other place in the driver (alignment, limits,...) ?

Patrick
7 years 1 month ago #14170
Attachments:

Please Log in or Create an account to join the conversation.

I emailed Jean-Luc since he's the expert on this topic but it appears he's a bit busy. I applied the patch and it fixes the issues you described. It doesn't appear to have affect on alignment (it is used for DEC Inversion in Southern hemisphere). I will commit the patch and will do more testing tonight at my observatory if it is clear.

Btw, I wonder how we can perhaps generalize this back to INDI::Telescope? It is specific to GEMs so INDI::Telescope must only enable this for GEMs. Currently there is no selection for mount types (no assumption is made).
7 years 1 month ago #14219

Please Log in or Create an account to join the conversation.

  • Posts: 210
  • Thank you received: 104
Thank you Jasem,
I will also do more testing as soon I have a clear night, at the moment this is rainy for the days to come.

Yes it can be interesting to generalize this property for the other drivers.
This probably make necessary to add a GEM switch option because it is not trivial to find otherwise. For example the LX200 are fork mount but many use this protocol for GEM using compatible box.
Non GEM mount can return a value of PIER_NONE for example.
7 years 1 month ago #14224

Please Log in or Create an account to join the conversation.

Yup, I was thinking along the same lines. But perhaps PIER property gets only defined if the mount type is GEM. Default mount can be UNKNOWN and must be updated to a known type before connecting with the mount (one-time operation). This could still be disruptive to a lot of existing users so have to think carefully about it.
7 years 1 month ago #14257

Please Log in or Create an account to join the conversation.

  • Posts: 226
  • Thank you received: 88

Replied by Jean-Luc on topic Eqmod telescope pier side

Hi,I understood this property as describing the side of the pier the optical tube lies, not where it points to. I thought it may be used in dome drivers, but have no knowledge about them.
I've read the description of the Ascom property, so it is more clear for me now.
I noticed that you use the unaligned coords, I wonder if that will be correct when near the "normal"/"beyond the pole" limit. If the scope is still beyond the pole and the target has passed the limit, the Dec reversal may appear too late. But I don't really know what you do and how.
Otherwise I've searched the use of pier in libindi, and only found it in lx200 pulsar and eqmod.
Jean-Luc.
7 years 1 month ago #14272

Please Log in or Create an account to join the conversation.

Jean-Luc,

With this change, the dome driver assumptions about PIER side are incorrect? I'm sorry but I still do not understand what exactly this property is supposed to convey besides "other end of the tube is on this side of the pier"? Perhaps we need to officially define it in INDI.
7 years 1 month ago #14278

Please Log in or Create an account to join the conversation.

  • Posts: 210
  • Thank you received: 104
Jean-Luc,

I understand the potential need for a dome driver to know the physical position of the telescope on the mount. But as this position is computed using the pointing hour angle it is very easy for any client to compute it using the aligned RA/DEC returned by the driver.

It is more important that the driver return the mechanical pointing state as described in Ascom because this cannot be guessed by the client application. We really need the mechanical state based on unaligned coordinates.

For example, a client application need this value to manage the meridian flip correctly and an auto-guiding application need to know when the optical train as been reversed.

Jasem, I think it is good if INDI use the same definition as Ascom.

The description of Ascom SideOfPier is here:
www.ascom-standards.org/Help/Developer/h...opeV3_SideOfPier.htm

And I attach here the document with the figures that help to understand the problem, so you not have to get the Ascom developer tools.

If more detail is really need for a dome driver we can extent the property to distinguish the four quadrant.
Point B in the attached Ascom document can be named WEST(pointing west) and point C is EAST(pointing East).

Patrick
7 years 1 month ago #14295
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 226
  • Thank you received: 88

Replied by Jean-Luc on topic Eqmod telescope pier side

What I meant is that this property may be used in dome drivers. I run " grep -ri pier" in indi and it only matches in eqmod and lx200 pulsar. So it seems it is not used for now. Otherwise I have no knowledge in writring dome drivers and do not plan to write one. Maybe people running scripts should be warned if they use the property.

The link given by Patrick is the one I read, I also read this one . I realize that I should read them again as I make confusion in the terminology, when looking at the one you attach. Anyway that is indeed a good thing to have a precise definition. I would say that the same could apply to pulse guiding: they are named Guide North/GuideSouth for instance, but are more related to DEAxis +/DEAxis-. The same for the move buttons. Actually I don't know if this is what client software expect.

Jean-Luc.
7 years 1 month ago #14298

Please Log in or Create an account to join the conversation.

  • Posts: 983
  • Thank you received: 375

Gentlemen, I have not taken a look at the patch but have just had an accident with eqmod driver, which is kind of dangerous.
Telescope pointing east of meridian, slewing to a position (significantly) west of meridian - instead of going around, including meridian flip, the telescope just crossed the meridian and gets into collision with the tripod. It has never happened before. Are you sure that the software accounts for a pier side the way it should?
Last edit: 7 years 1 month ago by Radek Kaczorek.
7 years 1 month ago #14383

Please Log in or Create an account to join the conversation.

  • Posts: 226
  • Thank you received: 88

Replied by Jean-Luc on topic Eqmod telescope pier side

Hi,
Thanks for reporting the problem.
I looked again at the patch, it should not make a difference, with respect to the driver itself I mean.
But I don't have a mount by hand here so can't do real tests. I just made a East/West goto test in the simulator, it did a meridian flip.
Now I wonder if there may be some changes in the behavior of clients ?
Jean-Luc.
7 years 1 month ago #14387

Please Log in or Create an account to join the conversation.

  • Posts: 210
  • Thank you received: 104
I get two hours of clear sky tonight before the cloud return, enough for testing mount pointing and meridian flip.
As I have a mobile setup I start as usual by making an automated 15 points model on the whole sky, checked the reported pier side on each position is good.
Then do a lot of goto everywhere switching side or not, near or far from the meridian, below the pole or on the south horizon.
Finally start a sequence on ngc2174 to test auto-flip after meridian one hour later but the cloud do not wish I finish that.
For me everything work as expected.

Patrick
7 years 1 month ago #14660

Please Log in or Create an account to join the conversation.

  • Posts: 969
  • Thank you received: 94

Replied by alacant on topic Eqmod telescope pier side

Hi everyone
indi-eqmod was not mirroring ASCOM in that its guiding was side of pier dependent; DEC pulses were dependent upon side of pier upon initialisation.
A big thanks to Andy Galasso for helping to track the bug, code the fix and to Jasem for his rapid merging of the pull request, now available in master. We hope that this helps others who have had difficulties with the driver.
Details at: github.com/indilib/indi/pull/658
Cheers and clear skies,
Steve
Last edit: 5 years 8 months ago by alacant.
5 years 8 months ago #27818

Please Log in or Create an account to join the conversation.

Time to create page: 0.693 seconds