Title: APS Formation Sensor
1APS Formation Sensor
- Optics
- Dennis Charles Evans
- 15 March 2002
2Optical Studies
- Optical Design
- Centroid Error Modeling
- Summary
- Back-up Information
3Optical Characteristics
- Telecentric Design (Modified Aerial Mapping
Camera) - Changes in length do not change plate scale
- Aperture 100 mm (selected)
- Optical Speed f/5 (selected for ease of
fabrication) - Focal Length 500 mm (result)
- Optical Design Half Angle FOV 5.0 degrees
(larger angles possible) - Plate Scale Needed for Point Design
- Based on deployed element distribution
- 5 km at 50 km (source) 50 mm at 500 mm (focal
plane) - 40960.013 53.248 mm typical detector
- Half angle 2.9 degrees
- The tracker could handle a wider distributed
source by using a larger detector or series of
detectors. - Glasses Schott SK4, F 5, LF 5
- May have to replace with radiation hard glasses
- Design is internally baffled only
- An external sunshade may be needed for a
particular instrument
4Optical Characteristics
- Tolerance Analysis (not done, not expected to be
a problem) - Design was selected to be insensitive to changes
in focus. - Thermal characteristics similar to CHyMERA.
- Thermal changes in index of refraction change
focus more than mechanical expansion of lens
holder. - A matched lens holder coefficient of expansion
will not make an thermally insensitive design. - CHyMERA required near zero CTE (GFRP) because of
thermal gradients perpendicular to optical axis
(banana effect) - Commercial Tilts Decenters acceptable
- Implications
- Proportional Thermal Controller (3 degree C
range) - Would like Lenses near ambient and Detector near
0 ?C - for simplicity detector housing should be above
freezing
5Optical Characteristics
- Ghost images will result from reflections at lens
and filter surfaces - Filter ghosts will be centroid aligned with
beacon image - Filter ghost will be 0.004 of image or lower
- Lens ghosts will be generally diffuse
- Reference to other designs implies lens ghosts
will be 0.001 to 0.0001 of image or lower. - Flat Field Monitoring is provided by illuminating
a diffuser plate on the back of the closed
aperture door - Wavelength filters are needed for resolving
multiple beacons on same daughter or daughters
very close together. - depending on stability, beacons could be turned
off and on for identification and angular
separation
6Telecentric Aerial Photo Lens
SK4 3.57 g/cc
F5 3.47 g/cc
LF5 3.22 g/cc
SK4
F5
SK4
SK4
SK4
LF5
100 mm diameter Telecentric Stop
7Defocused Spots for Centroiding
8AutoCAD 3D Model
9Basic Mounting Internal Baffle Layout
Defining 100 mm diameter Telecentric Stop
Pseudo Stop Aperture
Detector
10Centroid Error Modeling
- Uniform Defocused Image Model
- Photoelectron Random Normal Statistics
11Centroiding Model Results
- Electro-Optical system throughput was calculated
for 0.001 sec cycle. - 3 x 3 and 9 x 9 pixel centroiding with 40 000
PE/px, integrated for 10 cycles (0.01 sec),
results in noise statistics about 5 x better than
the 0.012 arc-sec error needed. - For 1000 cycle averaging, the centroid error is
0.0005 arcsecond equivalent to 0.150 mm at 50
km, much better than the basic requirement. - A significant noise margin exists for the point
design! The margin on noise can be greatly
improved if needed. - Noise is reduced most effectively by increasing
the number of PE/px. - Well depth is being improved for APS arrays.
CCD arrays have an order-of-magnitude deeper
wells at present.
12Centroiding Noise Model
13Centroiding Error Sensitivity
14J40 000PE/px, 9x9, 1 Cycle Average
15J40 000PE/px, 9x9, 3 Cycle Average
16J40 000PE/px, 9x9, 10 Cycle Average
17J40 000PE/px, 9x9, 1000 Cycle Average
18Wide Flux Design Range
- Electro-Optical system throughput was evaluated
at 0.001 second cycle rate, not 1 second. - System time constant is likely to be in the
seconds to minutes range. - Position requirement of 3 mm at 50 km was 0.012
arcsecond - Integration of Centroid for 1000 cycles gives a
position error of 0.0001 pixel width (1/22 of
design goal) - Centroid error is 0.0005 arcsecond equivalent to
0.150 mm at 50 km. - Implication
- System has more capability than required
- Could be reduced in size by an order-of-magnitude
in volume
19Optical Performance Summary
- Thermally stable Telecentric Optical Design
- analytically attractive
- flight proven design concept
- Position error due to photoelectron statistical
noise is well below the 0.012 arcseconds needed
for this design - If 3mm error at 50km is all that is needed, the
size of the sensor system might be reduced
significantly, possibly by a factor of 10 in mass.
20Back-up Information
21Aerial-02.ZMX Prescription
22Lens Element Mass
- ELEMENT VOLUME DATA
- Values are only accurate for plane and spherical
surfaces. - Element volumes are computed by assuming edges
are squared up - to the larger of the front and back radial
aperture. - Volume cc Density
g/cc Mass g - Element surf 2 to 3 372.375669
3.570000 1329.381137 - Element surf 4 to 5 1111.283091
3.570000 3967.280633 - Element surf 6 to 7 670.546997
3.470000 2326.798080 - Element surf 11 to 12 786.881150
3.220000 2533.757302 - Element surf 13 to 14 832.146110
3.570000 2970.761614 - Element surf 15 to 16 249.365462
3.570000 890.234699 - Total Mass
14018.213466
23Photo Electron Noise Modeling
- Photo Electron Noise is related to the square
root of the signal. - The square root of the signal is approximately 1
sigma deviation. - The noise is modeled as a Normal (Gaussian)
distribution with a one sigma standard deviation
equal to the square root of the signal. - Procedure for generating normal distribution
noise - Computer language generates pseudo random number
from 1 to 1000 and scales to 0.001 to
1.000 - Function converts number to random normal
distribution.
24Normal Distribution
25Random Normal Signal Distribution
26Centroiding Model
- 9 TrackerFocalPlane 40000 Centroid is at
12,8 39572 - ArrayJI
- A1 A2
- 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 - 0 0 0 40128 39672 39884 39948 39964 39804 39752
40060 40300 0 0 0 0 0 0 0 0 - 0 0 0 40128 39844 40032 39504 39880 40380 40004
40156 39992 0 0 0 0 0 0 0 0 - 0 0 0 40236 39604 39792 39800 40052 39876 40116
39948 40264 0 0 0 0 0 0 0 0 - 0 0 0 39820 40208 40092 40572 40092 39816 40028
40124 39808 0 0 0 0 0 0 0 0 - 0 0 0 39976 39960 39892 39944 39572 39844 40252
40024 40140 0 0 0 0 0 0 0 0 - 0 0 0 39936 40280 40100 39872 40124 40220 39836
39740 39696 0 0 0 0 0 0 0 0 - 0 0 0 39940 40420 39924 39832 39980 39524 39944
40144 39880 0 0 0 0 0 0 0 0 - 0 0 0 39936 40168 39976 40228 40040 40104 39840
40068 39704 0 0 0 0 0 0 0 0 - 0 0 0 40028 39776 40008 40128 39904 40080 39840
40112 40124 0 0 0 0 0 0 0 0
A1 A2 etc.
27Centroiding Model Numerical Sample
28Centroiding Model Numerical Sample
29A Design with Filter Lens Ghosts
30Typical Reflection Ghosts
31Typical Lens Filter Reflection Ghosts