[BAA Comets] Camera time-stamps for astrometry

[Graham Relf]

Listening to Roger's talk on Saturday encouraging us all to measure our 
comet images I suddenly realised that I did not know for sure when my 
camera creates the time-stamp for each image. I had assumed it was at 
the end of the exposure when the file is created for the data to be 
saved into. I have just done an experiment which reveals that my 
assumption was wrong.
My Canon DSLR in fact records the date and time of the beginning of the 
exposure.
For comets I always stack either 32s or 64s exposures (the latter 
controlled by APT software). So in my case, for reporting astrometric 
results for each image I need to ADD half the exposure time to the 
camera's time-stamp (rather than subtracting, as I had assumed).
I determined this by first taking a 30s (manual) exposure then turning 
the dial as quickly as possible to 1" and taking another one. The 
interval between the time stamps was 38s. So it had taken me 6s to 
adjust the dial (30" by the dial is in fact a 32s exposure. I have 
previously timed that. That's nearly a 7% difference but camera makers 
stick to conventional dial marks despite shutters being more accurately 
controllable now).
So do you know what happens in your camera? Does it store the start time 
or the finish time? You need to know that for accurate astrometry.
Are we aiming for 1s accuracy?
I wonder how many MPC reports have the time wrong by up to a minute or 
so, depending on exposure length?
Or do they correct the time knowing what each camera does? I think that 
is very unlikely but if so my reasoning above would be redundant. Can 
anyone tell me about this?

Graham