Centroid Gain Selection on AOACQ tool: What does it do?
- Centroid gain is a 608 element vector keyword called 'wscngn'
- The 'typical' value for all 608 entries in wscngn is 0.35165
- This value is computed from seeing, object size, and 'wscngn0'
(a value read from the wscngn.dat file which is 0.35165)
- The formula used for this computation is:
- When seeing and object_size are both zero (the typical case)
the 0.35165 value is unchanged by this formula.
- The AOACQ tool provides these choices for object size (the value
in parentheses is what is used in the above fwhm equation):
For a few bodies in the list above, the actual size range (taken
from JPL-Horizons between 2001 and 2010 when observable from
Manua Kea) is given in square brackets following the aoacq value.
- Neptune (2.4*0.7) [2.20 - 2.24]
- Uranus (4.0*0.7) [3.35 - 3.48]
- Titan (1.0*0.7) [0.69 - 0.77]
- Io (1.0*0.7) [0.78 - 1.00]
- AGN (1.0)
- LGS (1.5)
- Star (0.0)
- The values in parantheses in #6 above
(which were taken directly from the aoacq.pro source code)
are of the form X*Y where X
is a rough approximation of size expressed in arcseconds and Y is
a multiplier for converting subtended angle to fwhm (always 0.7).
- What effect does it have? Check out these before and after
images link to neptune images
- Why is it needed? The spot size affects the quad-cell centroid
Here is a graphic depiction
- How critical is it? Given the discrepancy in the size values
used from the pull-down (parens in #6 above) from the actual sizes
(square brackets in #6 above),
possibly this value needs to be good to only within 10 or 20%.
This may be a good issue for future engineering.
For bodies like Io, which can vary in size by 20%, should
we be setting this value to a size taken from Horizons?
- This table gives some typical
values for seeing set to either
zero or 0.75. Note that the default value following a reboot
(0.35) is for 'star' and seeing=0. However, acqtool comes up with
the seing pull-down defaulted to 0.75. So just going from uranus
then back to star, will not restore centroid gain to 0.35, but
rather will increase it to 0.53 (unless you specifically change
seeing to zero).
- More info regarding centroid gain and the wscngn keyword
appears at this link.