Hi all, in an NSC system, I need to obtain a precise x and y luminance value on the color detector, I know the transmittance curves of the material to be crossed and consequently created the glass, I set a merit function where I search for the desired x and y values. I have created a radial source to which I would like to set the x and y emission values according to CIE1931 as variables. In the past, to find the scatter values to apply to a given HPA material, I used the parameters of the multi configuration editor. Is it possible to use the same methodology?
Thanks a lot to everyone.
Best answer by MichaelH
Hi Luca,
There is a very convoluted way of setting the Tri X and Tri Y values as variables but you have to use a mix of the MCE and 2 ZPL macros. The thought process is as follows:
Use a dummy object to control the TriX and TriY values (use X Position and Y Position); this dummy surface will be the variables the MFE will optimize
Use the MCE with NPOS 1/1/4 & NPOS 1/1/5 to read in the X Position & Y Position in the MCE; these cells will have ZPL solve on them...the ZPL solve will:
read in the X/Y Position using NPOS
use SETNSCPROPERTY with 181/182 to set the Tri X & Tri Y on the Source Radial
return the Tri X & Tri Y value with a SOLVERETURN (this can be any value but you have to have a SOLVERETURN so might as well be the actual Tri X/Y value)
Use the MFE with NPXV & NPYV to target the desired Tri X and Tri Y value
The properties of the objects can be called in the multi-configuration editor using the NPRO keyword:
Unfortunately, the source properties are not available there.
In one of our examples, we had different sources with different wavelengths and we were optimizing on the source power. I don’t know if that could be an option for you.
There is a very convoluted way of setting the Tri X and Tri Y values as variables but you have to use a mix of the MCE and 2 ZPL macros. The thought process is as follows:
Use a dummy object to control the TriX and TriY values (use X Position and Y Position); this dummy surface will be the variables the MFE will optimize
Use the MCE with NPOS 1/1/4 & NPOS 1/1/5 to read in the X Position & Y Position in the MCE; these cells will have ZPL solve on them...the ZPL solve will:
read in the X/Y Position using NPOS
use SETNSCPROPERTY with 181/182 to set the Tri X & Tri Y on the Source Radial
return the Tri X & Tri Y value with a SOLVERETURN (this can be any value but you have to have a SOLVERETURN so might as well be the actual Tri X/Y value)
Use the MFE with NPXV & NPYV to target the desired Tri X and Tri Y value
Excellent I would say, I can integrate this workaround into my system and make it look for the appropriate values of emission x and y to obtain the chromatic coordinates of luminance on my target.
We use 3 different kinds of cookies. You can choose which cookies you want to accept. We need basic cookies to make this site work, therefore these are the minimum you can select. Learn more about our cookies.