After the great “Brockville” fireball of 2020dec07, 19:30EST, we checked allsky1pi as well to see if it recorded the event that allsky2 (UWO) captured.

It did not.

Hmm… wonder why not?

After days of troubleshooting we came across an interesting anomaly.

The allsky1pi camera performs autoexposure from early dusk through dark and into early dawn. Looking at Sunday’s data, we had approx 540 images at 161MB, and during the dark cycle these maxed out at 60 seconds.

Interestingly enough, the config file for the software lists exposures and maxexposures at 90 seconds. hmmm.. looking more carefully at each image, they are in fact timestamped 90 seconds apart, not 60.


It seems that the camera is exposing up to a max of 60 seconds and then waiting for 30 seconds to fulfill the configuration file’s 90 sec entry.

Checking the camera stats, it says it can do up to 1000seconds. So.. a possible problem explained but not fixed.

We changed the configuration to 60seconds for exposure and maxexposure and reviewing the last two nights, it seems to be working. Timestamps are now only 60 seconds apart for 60 sec exposures and there are approx 820 images now at 270MB.

So the most likely answer as to why allsky1 did NOT capture the event.. it happened in the30 second deadtime between exposures.

OK.. that is now fixed! no more missing time!

Still have to figure out what is responsible for limiting it to only 60 seconds? The camera hardware itself in some way? The autoexposure has some limiting factor configured somewhere else? Gremlins?