(note: these pages refer to the gonio-photometer built by me around 1990 at Fraunhofer ISE, Freiburg. For the newer 2004 gonio-photometer pgII, developed by me at pab Ltd, please check the pab Ltd website.)
My diploma thesis 1989-1990 in Physics consisted of designing, building and operating an out-of-plane gonio-photometer
for BRTF measurements at Fraunhofer ISE from scratch.
This included shaping and fitting models for the Radiance simulation program,
used in day-lighting.
Many of sample of new window materials, mostly translucent insulation types, have been measured with it over the years and for multiple
European research projects.
After checking existing concepts and some thoughts on my own, the initial design featured two parabolic mirrors (light house mirrors)
with halogen lamps, full 4 axis control using high-end 5-phase stepper motors and a detector rail system that covered
a quarter of a sphere.
This setup is shown below (all images taken by me at FhG-ISE, except first photo showing testing the stepper motors at home).
Part of the first parabolic mirror is visible on the left side, around is a card-box with black surface inside that shielded stray light. The
second mirror is on the right hand side. The detector is mounted on the carriage on the linear rails and was tilted towards the source by
adjusting the gear drive by stepper motor. Barely visible is the drive system for the vertical axis, the basic arrangement of bearing,
motor and timing belt that I came up with and that proofed so sound over the years, that around 20 years later it was reused in a more
elaborate version with my pgII gonio-photometer.
In retrospect, the main drawback of the first design had been the use of the steel ''backbone'' (towards the wall) that limited a 360deg
movement, and the inflationary use of adjustment screws. It was really laborious to align, compared with the pgII.
Note the honeycomb sample on the table in front.
| | | | |
Together with the two axis sample holder, the design ensured that every incident and outgoing angle could be measured, while still using a commercial (COTS) linear rail system for the second detector axis. The latter is a nice, economic and reliable way to move the detector precisely, but from a physics point of view, the varying detector distance introduces errors that are hard to compensate for. All axis had full computer control, using 5-phase stepper motors.
The data processing saw the first version of my mountain program, using the ray-tracing software "rayshade" to do the visualisations.
Here are some scanned pages of my diploma work, including first fits of material models to measured BRTF data.
 |  |  |  |  |  | 
|
The large illuminated area of the sample causes converging beams from sample to detector with different outgoing angles (see original images from 1990 below).
| far-field detector at infinity | near-field detector at finite distance |
 | 
|
To handle this, a CCD camera (compact head and separate electronics) for near-field photometry had been added to the gonio-photometer already in 1990. In image below, the camera head and optics are mounted on the left side of the gear disc, with the camera electronics in an extra housing. However, the overall dynamic range, with video and frame-grabber, turned out to be considerably lower (1:40) than the data-sheet for the Thompson-CCD specified (1:5000).
 |
To handle this, a numerical model for the BRTF and the near field situation was implemented and its parameters fitted. The
Levenberg-Marquardt fit included summing up the distributions from the patches. Results had been quite reasonably, taken that the model for
the cone shaped BRTF had been relatively simple.
Results below include the original plots (HP-GL generated by Disspla), de-dusted, annotated and converted to SVG format for web use.
 |  |  |
| "straw" like material | laminated strips wa3101 | glass tubes |
 |  | 
|
| legacy data not yet available |  | legacy data not yet available |
Activities revolved around "transparent insulation" material, a topic which was very much "en-vogue" in the nineties.
The central idea is to have a high radiation energy flux from the outside to the inside, while keeping thermal convection and conductance in
the reverse direction to a minimum.
Different production methods and materials had been tested.
The last image shows an example of fairly advanced glass tubes with very thin walls and approx 6mm diameter.
All tube-based TI material shows a cone-shaped BRTF, which may cause glare of the material is used in bright sunlight and within view.
The width and shape of the scattering depends on material and production method.
The data example for sample wa3101 shows one of the early BRTF (incident angle 30deg), measured on Saturday January 6 1990, 17:33.
This screen dump used a modern version of mountain, which shows a lot more detail then the rayshade
generated visualisation in the diploma work (see image in scanned version above). By today's standards, the angular resolution of
this particular peak had been too low.
An algorithm for automaticly refined angular scans had been added later.
Since BRTF of this kind of material is non-standard and complex, it provided a very nice introduction into the art of BRTF measurements.
next: further work work based on this gonio-photometer (Phd, etc)
