×

INDI Library v2.0.7 is Released (01 Apr 2024)

Bi-monthly release with minor bug fixes and improvements

ASI usb3 camera retries 3 times on exposure and fails - git builds

  • Posts: 535
  • Thank you received: 109
Tonight, I am experiencing this error with the latest builds from git. My main camera, a USB3 camera, 1600MM will not take an image. The 290MM plugged into the hub on the 1600 works just fine.

So far I have tried shutting down completely, unplugging and re-plugging the 1600, starting only the 1600, and looking at logs that don't say much about what is going on.

This was also mentioned earlier in a thread for the 294 not able to connect, but as the error is different, I decided a new post was better.
This is the problem where the camera will not capture the exposure. It tries 3 times and gives up.

These are the build dates, so will try older packages before some recent merges.

libasi-1.9.0.git-20210319185434.fc33.x86_64
indi-asi-1.9.0.git-20210319185514.fc33.x86_64

Jim

[2021-03-19T20:18:28.341 CDT WARN ][   org.kde.kstars.ekos.capture] - Skipping flat/dark cap since it is not connected.
[2021-03-19T20:18:28.341 CDT INFO ][   org.kde.kstars.ekos.capture] - "Capturing 1.000-second Luminance image..."
[2021-03-19T20:18:28.352 CDT DEBG ][           org.kde.kstars.indi] - ZWO CCD ASI1600MM Pro : "[DEBUG] StartExposure->setexp : 1.000s "
[2021-03-19T20:18:33.266 CDT DEBG ][           org.kde.kstars.indi] - ZWO CCD ASI1600MM Pro : "[DEBUG] ASIGetExpStatus failed. Restarting exposure... "
[2021-03-19T20:18:33.267 CDT DEBG ][           org.kde.kstars.indi] - ZWO CCD ASI1600MM Pro : "[DEBUG] StartExposure->setexp : 1.000s "
[2021-03-19T20:18:34.570 CDT DEBG ][           org.kde.kstars.indi] - ZWO CCD ASI1600MM Pro : "[DEBUG] ASIGetExpStatus failed. Restarting exposure... "
[2021-03-19T20:18:34.571 CDT DEBG ][           org.kde.kstars.indi] - ZWO CCD ASI1600MM Pro : "[DEBUG] StartExposure->setexp : 1.000s "
[2021-03-19T20:18:35.875 CDT INFO ][           org.kde.kstars.indi] - ZWO CCD ASI1600MM Pro :  "[ERROR] Exposure failed after 3 attempts. "
[2021-03-19T20:18:35.878 CDT INFO ][   org.kde.kstars.ekos.capture] - "Capture failed. Check INDI Control Panel for details."
[2021-03-19T20:18:35.878 CDT INFO ][   org.kde.kstars.ekos.capture] - "Restarting capture attempt #1"
[2021-03-19T20:18:35.879 CDT INFO ][   org.kde.kstars.ekos.capture] - "Capturing 1.000-second Luminance image..."
[2021-03-19T20:18:35.891 CDT DEBG ][           org.kde.kstars.indi] - ZWO CCD ASI1600MM Pro : "[DEBUG] StartExposure->setexp : 1.000s "
[2021-03-19T20:18:37.188 CDT INFO ][           org.kde.kstars.indi] - ZWO CCD ASI1600MM Pro :  "[ERROR] Exposure failed after 4 attempts. "
[2021-03-19T20:18:37.189 CDT INFO ][   org.kde.kstars.ekos.capture] - "Capture failed. Check INDI Control Panel for details."
[2021-03-19T20:18:37.190 CDT INFO ][   org.kde.kstars.ekos.capture] - "Restarting capture attempt #2"
[2021-03-19T20:18:37.190 CDT INFO ][   org.kde.kstars.ekos.capture] - "Capturing 1.000-second Luminance image..."
[2021-03-19T20:18:37.240 CDT DEBG ][           org.kde.kstars.indi] - ZWO CCD ASI1600MM Pro : "[DEBUG] StartExposure->setexp : 1.000s "
[2021-03-19T20:18:38.544 CDT INFO ][           org.kde.kstars.indi] - ZWO CCD ASI1600MM Pro :  "[ERROR] Exposure failed after 5 attempts. "
[2021-03-19T20:18:38.545 CDT INFO ][   org.kde.kstars.ekos.capture] - "Capture failed. Check INDI Control Panel for details."
[2021-03-19T20:18:38.546 CDT INFO ][   org.kde.kstars.ekos.capture] - "CCD capture aborted"
[2021-03-19T20:18:38.548 CDT DEBG ][   org.kde.kstars.ekos.capture] - setMeridianFlipStage:  "MF_READY"
[2021-03-19T20:18:38.556 CDT DEBG ][           org.kde.kstars.indi] - ZWO CCD ASI1600MM Pro : "[DEBUG] Aborting camera exposure... "

UPDATE:
I reverted back to these dates:

libasi-1.9.0.git-20210315172622.fc33.x86_64
indi-asi-1.9.0.git-20210315172641.fc33.x86_64

and things are working again. There were quite a few updates during the last 4 days, one of these must not be quite right.

Update 3:
I spoke too soon. While some frames are able to download, for the most part, it still is not working right. I found how to turn on more debug logging as well, so attached that below.
 
Last edit: 3 years 4 weeks ago by Jim. Reason: logs and updates
3 years 4 weeks ago #68903
Attachments:

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

  • Posts: 535
  • Thank you received: 109
Final update for the night. I went back to

libasi-1.9.0.git-20210310202412.fc33.x86_64
indi-asi-1.9.0.git-20210310202416.fc33.x86_64

and every exposure is working as it should. I am capturing for the night. I hope the logs and information help to point out the problem.

Jim
3 years 4 weeks ago #68906

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

Could this be related to the newer libasi ?
3 years 4 weeks ago #68916

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

  • Posts: 535
  • Thank you received: 109
Looking at changes in both packages, here are the commit messages:
 
indi-asi:

Refactoring - ASI CCD driver (#359) Paweł Soja* Yesterday 1:20 PM
ASI CCD - Remove 'IS' Functions (#357) Paweł Soja* 3/16/21, 11:31 AM
Dev package versions 1.9.0 (#352) Jim* 3/10/21, 2:18 PM

libasi:

ZWO EAF and EFW SDK updates (#360) Jarno Paananen* Yesterday 1:18 PM
Leave usbfs_memory_mb up to system to decide instead of tampering with it Jasem Mutlaq 3/14/21, 7:53 AM
Dev package versions 1.9.0 (#352) Jim* 3/10/21, 2:18 PM

The version I used last night successfully was the #352, where only the package labels were changed, so at least there are not a lot of commits to weed through, and since it was broken in the 3/15 build, the break happened between 3/10 and 3/15. It looks like that leaves only the libasi change on 3/14.
3 years 4 weeks ago #68921

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

  • Posts: 1119
  • Thank you received: 182


I don't know if this means anything, but Jasem's comment up here

"Leave usbfs_memory_mb up to system to decide instead of tampering with it Jasem Mutlaq 3/14/21, 7:53 AM"

refers to the line I added to asi.rules to get my EFW working again:

indilib.org/forum/ccds-dslrs/9089-indi-c....html?start=12#68923

Jo
 
3 years 4 weeks ago #68931

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

  • Posts: 535
  • Thank you received: 109
Interesting. The line in question is:
ACTION=="add", ATTR{idVendor}=="03c3", RUN+="/bin/sh -c '/bin/echo 256 >/sys/module/usbcore/parameters/usbfs_memory_mb'"

Which could certainly affect capture. I wonder then if it is not set here any longer if it needs to be set on the local system. It seems this setting might need to be set based on camera and memory available, which would be unfortunate.
There is another thread on here where a user changed it from 256M to 512M to help their new hi-res camera work better.

Without this being set in the INDI driver, here is the default for my system:

$ cat /sys/module/usbcore/parameters/usbfs_memory_mb
16

That certainly was my problem last night. Reverting to before this was removed, I can verify that it is set to 256M

www.baslerweb.com/en/sales-support/knowl...d-usb-cameras/29826/
3 years 4 weeks ago #68936

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

  • Posts: 1119
  • Thank you received: 182
Thanks! That is very informative, but as far as I understand, that only refers to the image buffer?

www.flir.com/support-center/iis/machine-...ding-usbfs-on-linux/

In my case, I didn't even get that far, the control panel wouldn't even detect my ASI-EFW. Is the same memory used by the Indi control panel to mount peripherals?

Jo

PS:  I guess the simplest way to check would be to run

sudo sh -c 'echo 256 > /sys/module/usbcore/parameters/usbfs_memory_mb'

and then check whether the missing peripheral (ZWO-EAF or ZWO-EFW or camera) is coming up in the control panel. 

Since in every case I have found reported here a ZWO peripheral was involved, that might suggest that it is indeed the asi.rules file that sets a memory limit that is too low and thus prevents also other non-ASI peripherals from getting mounted in the Indi Control Panel.

Does that make sense?
Last edit: 3 years 4 weeks ago by Jose Corazon.
3 years 4 weeks ago #68938

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

  • Posts: 535
  • Thank you received: 109
Jo,

What did removing that line change for you such that it was deemed successful enough to merge to master?

Funny, searching for that parameter brings up other former topics on here:
indilib.org/forum/general/5125-still-sam...-it-out.html?start=0 for example.

It seems we need to figure out how to automatically allocate this buffer based on equipment list? Based on available memory?

Or if it did not fix your problem, we should probably just revert it for now, as it is causing other issues.
indilib.org/forum/ccds-dslrs/5701-qhy5-e.../43632.html?start=36
Looks like it was needed for QHY too.

Edit 2:
Even for the Pi, it looks like it should be a minimum of 200M: instrumentation.obs.carnegiescience.edu/...e/ZWO/zwoserver.html

Jim
Last edit: 3 years 4 weeks ago by Jim.
3 years 4 weeks ago #68939

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

  • Posts: 1119
  • Thank you received: 182
Hi Jim,

I did not REMOVE that line, I ADDED it back from an asi.rules on a system that was working. It was MISSING on my system when the ASI-EFW was not detected in the Indi Control Panel.

As I wrote in the other thread, I cannot be certain that that was what made the difference in my case, since I also ended up purging indi-bin and indi-full, then reinstalling it by recompiling from sources. So there may be some other changes that were made (and fixed by the complete reinstallation) while I was manually trying to fix my system.

I guess a simple test would be for someone who still has trouble with their peripherals not getting recognized to just run that command

sudo sh -c 'echo 256 > /sys/module/usbcore/parameters/usbfs_memory_mb'

to temporarily increase the memory to 256M, then restart Kstars and see if that fixes the problem of missing peripherals in the Indi Control Panel. What do you think?

Jo
3 years 4 weeks ago #68941

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

  • Posts: 535
  • Thank you received: 109

Jo, I misunderstood. I was under the impression that it was removed by Jasem to address a problem you had. Sorry.

@knro what problem were we looking to fix with the change, and did it make something else better?  It appears that many systems that do not set it high enough will fail.
At this point I think we should revert the removal of the line.

Jim
 
The following user(s) said Thank You: Wouter van Reeven
3 years 4 weeks ago #68942

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

  • Posts: 22
  • Thank you received: 4
I have the same failure with my as1600mm with indi and asi driver compiled from the latest git.
Capture fails.

$ cat /sys/module/usbcore/parameters/usbfs_memory_mb
16

$ echo 1024 /sys/module/usbcore/parameters/usbfs_memory_mb
solves the problem
The following user(s) said Thank You: Jose Corazon
3 years 4 weeks ago #68944

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

  • Posts: 535
  • Thank you received: 109
I put a PR up to revert to setting 256 again.

Jim
 
3 years 4 weeks ago #68957

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

Time to create page: 0.393 seconds