[time-nuts] finding time astronomically , Part 2
jfor at quikus.com
Mon Jan 23 20:59:32 EST 2012
You might be able to track the reddish stars both night and day. If you
put a dark red filter in front of the sensor, it will get rid of much of
the sky. The sky and stars are very different optically, the former is an
area source the latter a point source. The energy from the sky varies
directly with the area being observed, the latter does not.
> So, to summarize the chain so far..
> You need to solve two problems:
> What's my camera orientation with respect to the stars.
> Where is the Sun (or something else) as it moves across the field.
> Conceptually, if I have my camera fixed and look at stars over some
> hours, they'll follow a path that's an arc (think of pictures pointed to
> north star). That will give me the orientation of my sensor with
> reference to the celestial pole, and the instantaneous positions of the
> stars gives me rotation around that axis.
> But that's not sufficient to tell me what time it is, just how I'm
> oriented relative to the stars.
> So then, I look for something that moves, and by occultation or some
> other means, I can tell what time it is. (I suppose this is basically
> what the celestial nav method of lunars does, but, of course, the moon
> has to be visible)
> But, given that 1 second time accuracy requires 0.004 degree kind of
> measurements, that's tough with a wide field of view camera with
> megapixel kinds of resolution.
> And, it's going to be hard to detect stars with a small sensor, because
> they're not very bright. I was fooling with my old iPhone 3G, and it can
> see Jupiter pretty easily, and maybe Sirius, but you're not going to see
> even 0 magnitude stars.
> However, maybe a small inexpensive reflector to increase the aperture
> and a webcam would do. You could replace optical perfection with
> calibration, etc. (I suppose that's what Chris was doing with the
> camera lenses).
> There's a whole FOV aim point tradeoff here.
> Going with "sun only" schemes.. you get solar noon (and you apply the
> equation of time in some other way) by fitting a curve to light
> intensity vs time.
> Aligning with vertical can be done with a plumb bob or equivalent, and
> then a slit/photodiode can work, with curve fitting. Is this something
> that is "arduino-able"? (at least the data collection.. the reduction
> might be done with post processing)
> How do you align the slit vertically, relative to the sensor? (to the
> required seconds of arc)
> I guess I should go look at some descriptions of zenith sun detectors.
> it's probably obvious once you know.
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to
> and follow the instructions there.
More information about the time-nuts