Hi again!

It seems a new odd bug have come up, even with the "compatibility" firmware on the ASI120MC, as was found in this thread to make the ASI120MC work:
www.indilib.org/forum/ccds-dslrs/1656-as...xposure-looping.html

The RPi with the ASI ran fine all night with the ASI as an all-sky camera capturing an image every minute, but when I tried to copy the files from the RPi to my laptop using ssh, the RPi crashed completely. In my own logs, I can see that the capturing script complains about not capturing anything (it does not find the .fits file) just before the RPi crashes and becomes impossible to reach. I am able to reproduce the bug, as it have happened several times this morning. If I copy without indi running, the coping of the files works fine.

The RPi could also just be changing its IP, as I reported it doing with the original firmware in the above thread msg #12873, thus making me unable to contact it. However, in the previous bug, I was able to find its new IP address and log into the RPi. This time I was not able to do this, so it may or may not be related.

I must stress. that everything worked nicely 4-6 months ago with an ASI120MM, where the RPi ran for two weeks without any crashes and I was copying files each day. So this seems like a bug which have been introduced with one of the new versions of indi or some update of rasparian. How can I help fix this?

Best Regards,
Søren

Hi again!

Is it possible to get old indi releases? Just to check if it works with an old release as it did previously, but I can only find a "daily build" for RPi....

Best Regards,
Søren

Some more wierdness, I now seem to be able to completely crash the driver by ordering a short exposure:

pi@observatory ~ $ echo "stop indi_asi_ccd" >indiFIFO
pi@observatory ~ $ echo "start indi_asi_ccd" >indiFIFO
pi@observatory ~ $ indi_setprop "ZWO CCD ASI120MC.CONNECTION.CONNECT=On"
pi@observatory ~ $ indi_getprop | grep ASI | wc -l
104
pi@observatory ~ $ indi_setprop "ZWO CCD ASI120MC.CCD_EXPOSURE.CCD_EXPOSURE_VALUE=0.0001"
pi@observatory ~ $ indi_getprop | grep ASI | wc -l
0

I have tried to reboot the RPi and power-cycle everything and it still does this. Short exposures worked well a few hours ago. Doing a longer exposure, say 0.1s, the driver does not crash but I get no image.

Then I connected the camera to my windows machine and it worked fine, I then reconnected it to the RPi and power-shuffeled everything and now the above sequence works again.

Best Regards,
Søren

I also on rare occations get wierd images, see the attached file.
(the humidity sensor and rain sensors are not working at the moment, in case you were wondering about the 99.9% humidity and rain )

You are taking a 0.0001 exposure. Isn't that bellow the minimum exposure supported by the ASI120?
I think the minimum is 0.001. At least mine is. Anyway the manufacturer states 64 µs (0.00006s) but I never got below 0.001s.
That weird image seems to me that your USB bus is not getting the whole image at the same time.
Have you tried to reduce the USB bandwidth in the indi settings to 80% or lower?

There are no older INDI packages, but you can compile from Git and choose to compile at whatever point in time you like. I suggest you raise these issues in the ZWO forums as it is an SDK related bug and not an INDI issue.

On windows it easily goes below 0.001s and it also does that in indi right now.

It indeed could be a USB traffic issue. AFAIK the ethernet also runs on the same chip as the USB, so this might be the reason why I can bomb the system by loading the ethernet with heavy traffic. The bandwidth is at 40% btw. but I will try to lower it even further. What does the "ASI120MC.CCD_CONTROLS.HighSpeedMode" do, anyone knows?

I am sorry to have raised the issue here, but since it only happens with indi I thought it was an indi problem. I will try to email ASI, ZWO or whatever they are named

Yes you are right I just tried an exposure below 0.001s and it's working.
And yes the RPI ethernet is shared with USB bus, but at 40% it should work, it's already too low.
I remember that was the value I used with my RPi and it worked, so maybe there's another problem

If you dare, you could try to loop some exposures (mine does one exposure pr. minute) and then load your RPi with ethernet traffic to see if you can reproduce the crashing. But at your own risk of casue
After a power-cycle everything seems to work after a crash, except for that one time where I had to connect the camera to my windows machine...

I just tried using ASI120MC for first time with Orion 50mm guide scope. I normally use Lodestar X2 but just wanted to see if I will have problems with it. I did a quick focus and then tried to see if I run into any issues... and to my surprise, zero issues, it's working perfectly and been guiding for the last hour half without issues.

Wonderful image

Is you ASI also connected using a RPi?

Well you're both luckier than me, I can't even connect to ASI120MM on my RPi.
It works on x86_64 but not on my RPi. Like it does not recognize the connected camera.
By the way, it's a B+ with the latest Raspbian and indi 1.3.
Hummmm coincidence or not, the same happens with my SSAG.
I was just trying to duplicate universalmaster's crash but no luck. Back to the PC!