Sun Sensors for Attitude Determination

I'm interested in developing a sun sensor (or maybe a horizon sensor)
for a low-cost, amateur, high altitude project. However, I'm having
some difficulty determining the best setup for the sun sensor. The
majority of the sensing will occur over 20km (about 65,000 ft), so the
effects of radiation from the sky can be considered minimal.

I was considering two methods:

The first involves 5 photocells mounted at 90 degree angles to each
other (one pointed vertically), with a light shield mounted below
allowing a half-angle of around 60 degrees (i.e. blocking the light
from the earth, so long as the sensor is 30 degrees from vertical or
less). Measuring the resistance of all 5 photocells should yield
(with the appropriate calculations) the position of the sun wrt the
photocells. One concern I have about this method is the large
variation of performance of photocells (regardless of having the same
part number), and would probably require some careful calibration.

The other method is to have a sheilded field of, say, X by X
photocells and a single opening centered over it. This method is much
more "coarse", expensive, and harder to make, but less suseptable to
non-solar radiation. I'm aware that most "non-amateur" sun sensors
use CCDs instead of photocells, but I believe that is a little beyond
my experience (and budget).

Does anyone else have knowledge of other techniques involved in sun
sensors? Are any photocells or other light sensors better (or more
precise) than others? Also, are there any wavelength filters I should
consider? Any info is greatly appreciated!

Thanks in advance!
Dave

David Harper wrote:

Does anyone else have knowledge of other techniques involved in sun
sensors? Are any photocells or other light sensors better (or more
precise) than others? Also, are there any wavelength filters I should
consider? Any info is greatly appreciated!

Are you familiar with these papers:
http://cubesat.calpoly.edu/_new/refe...nsor_paper.pdf
https://escies.org/public/mnt4/S10.4Fleron.pdf

/steen

In article ,
David Harper wrote:
The other method is to have a sheilded field of, say, X by X
photocells and a single opening centered over it. This method is much
more "coarse", expensive, and harder to make, but less suseptable to
non-solar radiation. I'm aware that most "non-amateur" sun sensors
use CCDs instead of photocells, but I believe that is a little beyond
my experience (and budget).

The new CMOS imager chips are much easier to deal with than CCDs -- for
example, they have on-chip A/D conversion, so the outside interface is all
digital -- and also, if memory serves, quite a bit cheaper. Their imaging
performance isn't as good as CCDs, but for seeing the Sun that's not
important.
--
"Think outside the box -- the box isn't our friend." | Henry Spencer
-- George Herbert |

(David Harper) :

I'm interested in developing a sun sensor (or maybe a horizon sensor)
for a low-cost, amateur, high altitude project. However, I'm having
some difficulty determining the best setup for the sun sensor. The
majority of the sensing will occur over 20km (about 65,000 ft), so the
effects of radiation from the sky can be considered minimal.

I was considering two methods:

The first involves 5 photocells mounted at 90 degree angles to each
other (one pointed vertically), with a light shield mounted below
allowing a half-angle of around 60 degrees (i.e. blocking the light
from the earth, so long as the sensor is 30 degrees from vertical or
less). Measuring the resistance of all 5 photocells should yield
(with the appropriate calculations) the position of the sun wrt the
photocells. One concern I have about this method is the large
variation of performance of photocells (regardless of having the same
part number), and would probably require some careful calibration.

The other method is to have a sheilded field of, say, X by X
photocells and a single opening centered over it. This method is much
more "coarse", expensive, and harder to make, but less suseptable to
non-solar radiation. I'm aware that most "non-amateur" sun sensors
use CCDs instead of photocells, but I believe that is a little beyond
my experience (and budget).

Does anyone else have knowledge of other techniques involved in sun
sensors? Are any photocells or other light sensors better (or more
precise) than others? Also, are there any wavelength filters I should
consider? Any info is greatly appreciated!

Using Google I found http://www.redrok.com/electron.htm this is single axis
but uses cheap leds as the sensors, this suggest that you can make a 3 by 3
or 5 by 5 sensor without breaking the bank.

Also lots more hits on:
http://www.google.ca/search?num=100&...sensor+tracker

--
I make public email sent to me! Hydrogen Peroxide Rockets, OpenBeos,
SerialTransfer 3.0, RAMDISK, BoatBuilding, DIY TabletPC. What happened to
the time? http://webhome.idirect.com/~earlcp

On 21 Nov 2004 10:11:58 -0800, (David Harper) wrote:

snip
Does anyone else have knowledge of other techniques involved in sun
sensors? Are any photocells or other light sensors better (or more
precise) than others?
snip

I'm not sure I understand fully your need, but it did stimulate a thought. I'd
consider the idea of an optical system designed to image the entire sky, but
with a very poor focus so that the sun is spread across 30% or so of the image
plane. (I'm kind of thinking a defocused fish-eye lens.) A quad-detector
placed at this plane will provide four separate signals, the sum of which
provides your reference level. Orienting the quad detector optimally against
the ecliptic for the largest sensitivity, as the sun crosses over, may take some
thought or experiment or both, but I imagine that this can be arranged into a
small self-contained system. Whether or not it provides the precision or
repeatability you need... is another story.

Also, are there any wavelength filters I should
consider?

I'd tend to want to use a narrow part of the (sub)visible band with Si
detectors, perhaps in the 800nm or 880nm (far enough above the bandgap that
there is little temperature dependence of the wavelength response function) only
because they are very good (good NF, stable, cheap, widely available, high
impedance and low leakage against offset voltage, etc.) But I really have no
experience here, at all.

I'm curious about the answers, so I'll read along with you.

Jon

(David Harper) wrote in message
and would probably require some careful calibration.


Genrally, some form of calibration is requied. Since the sun is just
so damm bright there really is little problem in designing one that
works at sea level and then calibrating it on a fine day. But keep in
mind thermal stablity from the start.

Greg