CCD_Rasnik: Principle of the coding of the mask.


A mask is generated by stacking position coded basic building blocks. We will call such a block a B3. The horizontal and vertical codes give the position of the basic building block in the total mask. Counting starts at 0,0 in the lower left corner of the mask. A startbit is used to make recognition and decoding easier. It is the basepoint of the building block. The ones in the binary position code invert the location color. The startbit is also inverted in respect to the basic chessboard. The code for the horizontal position is in the vertical codelines and visa versa.The size of one spot (field in the chessboard) is in the order of 100 microns, which covers roughly 10 times 10 CCD camera pixels.

Coding example

(click to enlarge)

Example of a coded mask, with B3's of 11 by 9.


The codebits of the example above.

These are roughly 10 times 10 B3's. The vertical code is 10 bits long and the horizontal code 8 bits. So theoretically one can stack 1024 B3's vertically and 256 B3's horizontally.

Glass mask

The actual mask is a metal on glas mask, as used for integrated circuit production (or a contact copy of such a mask). The anti-reflective coated side should be pointing to the projecting lens.These masks are produced by commercial firms, taking Gerber files as input. Postscript files, plotted on high resolution machines, turned out to be disappointing. Since postscript interpreters work with floats, rather then integers, it is very hard to define the data in such way that rounding off is done in the proper way.Also tests were performed with gray code locations instead of inverting them. In practice gray has to be defined as fine lines of black and white and the results are unpredictable.

go back to the code page


august 1995