The "replica method" has been around in metal analysis and processing. Film is used in order to take an impression of a surface. Then, it is analysed for structures. In optical production, compound moulding seems advantageous in roughness determination in inaccessible spots or on large components (astrological optics). In preliminary trials, different glass samples with different roughness levels were manufactured. Imprints were taken from these samples (based on DIN 54150 „Abdruckverfahren für die Oberflächenprüfung"). The objective of these feasibility tests was to determine the limits of this method (smallest roughness determinable / highest roughness). The roughness of the imprint was compared with the roughness of the glass samples. By comparing the results, the uncertainty of the measuring method was determined.
The spectrum for the trials ranged from rough grind (0.8 μm rms), over finishing grind (0.6 μm rms) to polishing (0.1 μm rms).