Atik EFW2 no longer reporting finish of filter change

[RononDex]

Describe the bug When using the Atik EFW2, indi no longer realises that my Atik EFW2 finished moving the filter into place. I can physically verify that it moves the filter into place, but the INDI control panel keep showing a yellow light on the selected filter slot and the sequencer is waiting for the filter change to complete.

To Reproduce Exact steps to reproduce the behavior.

1. Connect to Atik EWF2
2. Change filter through indi

The logs below clearly show that indi is seeing that the filterwheel stopped moving, but the light still stayed yellow.

Expected behavior Completed filter change operation is correctly registered by EKOS / INDI

Screenshots If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: Arch Linux ARM
• Version 2.1.3

Log Files

[2025-04-03T13:13:20.944 CEST INFO ][           org.kde.kstars.indi] - Atik EFW2 :  "[INFO] Setting current filter to slot 2 "
[2025-04-03T13:13:21.097 CEST DEBG ][           org.kde.kstars.indi] - Atik EFW2 : "[DEBUG] iPosition: 0 isMoving: 1 "
[2025-04-03T13:13:22.099 CEST DEBG ][           org.kde.kstars.indi] - Atik EFW2 : "[DEBUG] iPosition: 0 isMoving: 1 "
[2025-04-03T13:13:23.102 CEST DEBG ][           org.kde.kstars.indi] - Atik EFW2 : "[DEBUG] iPosition: 0 isMoving: 1 "
[2025-04-03T13:13:24.105 CEST DEBG ][           org.kde.kstars.indi] - Atik EFW2 : "[DEBUG] iPosition: 0 isMoving: 0 "
[2025-04-03T13:13:25.108 CEST DEBG ][           org.kde.kstars.indi] - Atik EFW2 : "[DEBUG] iPosition: 0 isMoving: 0 "
[2025-04-03T13:13:26.111 CEST DEBG ][           org.kde.kstars.indi] - Atik EFW2 : "[DEBUG] iPosition: 0 isMoving: 0 "
[2025-04-03T13:13:27.114 CEST DEBG ][           org.kde.kstars.indi] - Atik EFW2 : "[DEBUG] iPosition: 0 isMoving: 0 "

[RononDex]

I was just able to confirm that this is a driver regression, when using indi 2.1.2 and indi-3rdparty at commit ac2dbea86e79a9a2cb821bec14cbebd47c353f94 it works properly

[TallFurryMan]

Thanks for this report. In your log excerpt the filter has 0 as the SDK index of the filter, even when moving. I suppose when you select another filter on your wheel this index moves to 1, 2...up to your number of filters minus one? Or it does stay at zero whatever the choice?

[TallFurryMan]

Ah sorry, the copy icon was masking the first line, where you logged that the filter is the second one, thus index 1 for the SDK. I'll ping Atik on this.

[RononDex]

Thanks! Thankfully it was easy for me to downgrade to a working version :)

[knro]

@RononDex Did you build this from source including latest libatik ?

[RononDex]

yes I built it from source (both indi and indi-3rdparty)

[RononDex]

Do you know if this was already fixed in the current indi-3rdparty git repository?

[knro]

No changes made regarding this issue. @TallFurryMan Any ideas if this was handled or noted in the latest libatik update?

[TallFurryMan]

I'm still waiting for Atik here. I sent a first message on Apr 3, ping'd again one week later, and sent a mail to generic support one week ago.

[TallFurryMan]

I received news from Atik. They need the exact camera model you used @RononDex, if you have more information than "EFW2".

They need us to encircle which SDK is causing the issue between indilib 2.1.2 and 2.1.3. They noted 2.1.2 and 2.1.3 embed the same SDK but they need the confirmation. So I'll get back to them with the answers in this ticket.

They also need to know if an Atik camera was connected as well.

Good afternoon Eric,
I’ve been assigned to investigate the issue concerning the EFW2 device and would like to clarify a few details to help identify the cause:
- Which Atik camera model was used to control the EFW2 device?
- Which specific EFW2 device model was involved?
- With Indilib version 2.1.2 and the old Atik SDK (2023.07.14.1456), does the EFW2 device operate correctly?
- With Indilib version 2.1.3 and the old Atik SDK (2023.07.14.1456), does the EFW2 device operate correctly?
- With Indilib version 2.1.3 and the new Atik SDK (2024.11.26.2038), does the EFW2 device operate correctly?
Once I have the confirmation of the models and versions used, I’ll attempt to reproduce the steps to better understand the behaviour.
Kind regards,
Diogo Ferreira | R&D Engineer

EDIT: for whatever reason I had invented a 2.1.4, sorry.

[TallFurryMan]

@RononDex do you have the ability to build the newer indilib package with the "old" Atik SDK?

[RononDex]

Which Atik camera model was used to control the EFW2 device?

--> I do not have an Atik camera, it the filter wheel is connected via USB to my raspberry pi (over a powered usb hub).

Which specific EFW2 device model was involved?

--> How can I check the device model?

Regarding the other questions, I tested it with indilib 2.1.2 and the old SDK and then it worked. I was not able to get the old indi-3rdparty with the old Atik SDK to compile with indilib 2.1.3 due to mismatch in dependencies. The drivers would instantly crash when trying to start them

[TallFurryMan]

@RononDex Are you able to use your EFW on a Linux with x86_64 architecture? I can rebuild 2.1.3 with the SDK change reverted and provide it to you.

[RononDex]

yes I could connect it to my desktop pc (also running arch linux, but x86_64)

[TallFurryMan]

I built indi 2.1.3 with the previous Atik SDK for you to test:

indi-core: https://drive.google.com/file/d/1g48F4iWFDAf7M05Oy8fF55lJEZD7RmVb/view?usp=sharing indi-3rdparty libraries: https://drive.google.com/file/d/1UttMAezSGuWnA1ZSE2bKP9v5dc8Wu7_B/view?usp=sharing indi-3rdparty drivers: https://drive.google.com/file/d/1JN2B_aOtChpbG6UUAbbDzoyk1Sqwchx0/view?usp=sharing

The environment in which those files were built is Ubuntu 24.04. If that doesn't work, I can use another environment.

Of course these files are only valid for the investigation of this ticket, nothing else.

[RononDex]

I will probably be able to test it in a week or so. My telescope is currently setup outside to use the good weather for imaging :)

[RononDex]

Also my desktop is running on Arch Linux, I am not sure if it can install a .deb file, I might find a way. Is there no way I could compile it myself (have a separate branch I could check out?)

[TallFurryMan]

Sure! https://github.com/TallFurryMan/indi-3rdparty/tree/issue/2214

[knro]

Is this issue resolved? Next INDI release due in 2 days.

[RononDex]

I will see if I can test it with @TallFurryMan branch tomorow

[RononDex]

@TallFurryMan I just tested it using indilib 2.1.3 and your linked branch of indi-3rdparty on a x86_64 machine and there it appears to be working correctly

[MarcelNoordman]

I have this problem on my machine and submitted a support ticket with Jasem. Attached my log file for reference. I am running on a raspberry pi 4

log_08-37-43 (1).txt

[TallFurryMan]

Pinged Atik back with this. The regression is then in one of the 2024 SDKs.

[MarcelNoordman]

I raised a ticket with them last week. This was there reply:

"Since the change that occurred in the Indi driver, although the SDK remains, there are some incompatibilities in the latest version of Indi. Unfortunately, we do not support Indi Driver.

As always, we are already in communication with Indi Driver support, who assured us that an update will be available in the next few days. Let's hope, then, that EFW 2 becomes compatible again.

Hélder Gonçalves | Production and Support Manager"_

[TallFurryMan]

Hm, not a very good communication from them, agreed. I made it clear we were looking into the possibility of a driver regression, but here it's definitely in the SDK. And it's not really a breaking change in the communication between the indi client and the indi driver when we observe the log.

However, if we take too long to come back to Atik, we risk our ticket there to be closed with a dummy reason.

Any chance one of you could test with the ASCOM driver too?

[RononDex]

I do not have any windows pc setup at the moment

[TallFurryMan]

I got news from Atik, they ack the issue is in the SDK and will push a new version soon. No ETA yet.

[RononDex]

That is good news!

[atlendor]

I'm very pleased to have come across this information as I have exactly the same issue! Has there been any progress, does anyone know? I dont want to send Jasem at Stellarmate/Ekos (yet another) support ticket if he/others are already on it...?

[MarcelNoordman]

Any progress on this issue? Do we already have an ETA?