1. Solomon William KAMUGASA
Development and validation of an absolute Frequency Scanning Interferometry (FSI) network TECHNICAL PROGRESS
Solomon William KAMUGASA
2nd PACMAN Supervisory Board, CERN
25/09/2015

2. Project goals
Realise absolute distance measurements with FSI between 2 points of interest.
Absolute Multiline
To provide a portable alternative to CMM that can cope with larger measurement volumes.
3D coordinates
High accuracy
Reliability
Role in PACMAN
Fiducialisation
Scale for µ-triangulation
Methodology:
Simulations
Development of suitable prototypes
Calibration, comparison tests, validation, extrapolation

3. Absolute distance realisation
First prototype
Why a sphere?
Fibre tip w.r.t sphere centre can be determined accurately
Absolute distances in different directions from same point
Use existing reflector supports
CMM & µ-triangulation measurable
Properties
Aluminium (not very stable)
1.5 inch diameter
7-9µm sphericity
Collimator glued in place

4. Global Relative Redundancy (GRR)
For uniform network strength, local reliability ZI should be close to GRR
For well controlled (reliable) network
ZI ≥ 25%
If GRR < 25%, some ZI values < 25%
4 FSI stations insufficient for a reliable network

5. 5 station pyramid geometry
Best 4 station geometry
Square geometry
PDOP of 1.5 in both cases
Square planar arrangement: Self calibration impossible
Tetrahedral arrangement:
Stations need to be closer to object than in square arrangement
More sensitive to displacement of target
Tetrahedral geometry
5 station pyramid geometry
Pyramid geometry
Target positioned so as to form platonic octahedron
Benefits of square geometry (PDOP, less sensitive to displacement)
Self calibration
PDOP=1.38

6. Simulations Performed in LGC++ 1000 simulations A priori stdev 5 µm
30 cm
40 cm
Measurement object side elevation y x z
Target
Station
2 m
Performed in LGC++
1000 simulations
A priori stdev 5 µm
Confidence level 99%
Power of the test 90%
5 stations
10 targets
Stations arranged so that optimum target position is object COG
All targets in on plane (same y coordinates)
4 stations 1 m from object COG, 5th station 2.4 m from COG (in y)

7. Some simulation results
Stations
Additional targets
Station x,y,z stdev [µm]
Target x,y,z stdev [µm]
NABLA [µm]
Global reliability
Fixed
N/A
3-5
25-35
1.5733
Free
14-330
4-88
59-160
5.7738
5 (interstation)
1-2
3-4
28-47
1.4702 5
(large volume)
4-9
4-6
28-38
5.4277
4-15
38-227
(Small volume)
3-11
28-39
4.3576
3-8
35-114
Relates to additional targets
N=2 Retro-reflector Study
N=2 desirable (wide viewing angle)
Encouraging results with QDaedalus (it’s early days)
Measurable with FSI
(ongoing optimisation studies)

8. Training, Secondment & dissemination
Latest Training
Measurement uncertainty estimation in engineering applications, CERN (sept)
Effective Article and Report Writing, EPFL (Oct-Nov)
3 month secondment (2014)
Etalon AG, Braunschweig, DE
1 week technical visit (Nov 15)
Liberec TU, Czech metrology institute, Liberec, CZ
Past Training
Programming
French
Health & safety
Presentations
Team building
Optics lab
AXEL
Dissemination
Swiss National Report on Geodetic activities
IWAA 2014
LVMC 2014
ETHZ 2014
PACMAN 2015
CLIC 2015
Conference
EPMC, Manchester, UK (Nov 2015)
Presentation
Poster
Possibly paper (Measurement Science and Technology)

9. Next steps 1. Produce prototype in ceramic
more stable, measureable by QDaedalus, accurate machining
Status: Order placed
2. Acquire 0.5inch N=2 targets
Compare performance with SMR
Calibrate (compare theoretical & experimental offsets)
3. Finalise simulations
Precision, reliability, cost, feasibility
4. Build automated prototype
Remove error due to touch, speed up data collection

10. Thank you for your attention.