FITS times wrong

Hi all,

I’ve noticed that the dates/times in FITS headers are wrong. For example I took an image at 18:55 local time (UTC+0200 with daylight saving) and the header contained:

DATE-LOC= ‘2019-09-13T20:55:42’ / Local observation date
DATE-OBS= ‘2019-09-13T18:55:42’ / UTC observation date

The image was dated into the future by my timezone.

I’m using beta 226.

I haven’t found where that could be changed in SGP so I theorise a bug.

Thanks.

@Cedric_Raguenaud

I reviewed the code here and it looks correct. Mostly… I did find an issue with how SGPro reads time from QSI cameras. What kind are you using?

Hi Ken,

I use a qhy. But I don’t think they’re aware of the time.

It’s like sgp thinks my system is set to UTC but I live in a +0200 timezone, so it adds 2 hours to the time. I hadn’t seen that with the previous beta. The computer definitively displays the right time.

I’ll dig into this some more tonight.

@Cedric_Raguenaud

The time selection logic for DATE-LOC goes like this:

  1. Check if the camera provides time. If it does, grab and convert it to local time, write it in the FITS header.
  2. If the camera does not provide time, use the current time, get the local time from the machine.

Verified as working… just took a picture at 14:11:

image

You can check to see if the camera is providing time in your logs. After taking a picture, open logs and search for:

DATE-LOC time provided by

You’re totally right. The log says the time is provided by the camera. Logs and fits headers attached. But it’s not something I can set. So it smells like the qhy ascom driver is broken.
sg_logfile_20190914234630.txt (20.5 KB)

What property of the driver is used for time so I can report it?

Thanks.

@Cedric_Raguenaud

The ASCOM property we read is named LastExposureStartTime

www.mainsequencesoftware.com