Scanner Mirror Flatness Testing

Hi All,

In another thread, I told you that I was going to find some time to do some flatness testing of the mirrors on common scanners. Since the other thread mentioned specifically the DT-40 PRO and Cambridge 6215, and since I happened to have both of them here, I did some flatness testing on the mirrors from these specific scanners.

Just so all of you don't think that I am just pulling things out of my a**, below you can see a picture of the test setup. This is a professional ZYGO model PTI interferometer, made specifically for mirror flatness testing. You can see the DT-40 PRO scanner on the tilt able, with mirror being examined in the monitor. There is also a printer (not shown) that was used to print the interferograms so I can scan them in and paste them in as well (shown below).

Attachment 159

Below you will see the pictures (interferograms) that were taken of the mirrors. In a nutshell, the Cambridge mirrors are pretty flat, and the mirrors from the DT-40 PRO are definitely not flat, having a complex surface structure (sort of like mountains with a valley).

Attachment 160

**DISCLAIMER** Note that I only tested two mirrors from each type of . In another post I mentioned that testing one or two parts out of hundreds does not tell you much about all of the other pieces. BUT, if we tested 10 mirrors from each scanner and saw roughly the same results, then we can draw a conclusion that the mirrors supplied with Cambridge scanners are flatter than mirrors supplied by DT-40 PRO scanners. Intuitively, I would say that this is probably the case.

DISCUSSION

Now, lets discuss why I am bringing this up. In another post or two, the reflectivity of a scanner mirror was being hotly debated. It seems that people were splitting hairs over 95% reflective, or 97% reflective, or 93% reflective, or 80% reflective, etc. Let me tell you something that I hope all of you can understand. The human eye is a highly logarithmic instrument. It can be demonstrated that if you have two laser systems side by side -- WITH ALL OTHER THINGS BEING TRULY EQUAL (and they must be equal for the experiment to work), you would have to reduce the laser power by a factor of 4 or 5 before human vision interprets the dimmer laser as being "half as bright". This means that the difference between 80% mirror reflectivity and 97% reflectivity is nearly imperceptible to the human eye.

BUT, if the scanner mirrors are not flat -- if they are concave or convex, they will make the beam larger (probably not smaller because it would most likely converge and then diverge) at the target surface (the projection screen). A larger beam will appear to the eye as far dimmer to the eye, owing to lambertian effects and the inverse-square law.

What's more, if the mirror is wiggling as the scanner is scanning, the lines will not be straight, but instead will have "ringing" in them. This "ringing" fortunately shows up only in certain parts of imagery, so it most often does not corrupt the entire image. Nevertheless THESE EFFECTS ARE HIGHLY VISIBLE!!! Much more so than a few percentage points of reflectivity... Those present at FLEM saw some of this, especially noticeable in scanners that were not tuned correctly.

Anyway, I hope this helps illustrate why thin mirrors are bad. In the past, I always commented how thin mirrors are no good for scanning. But now we have additional data (at least a few data points) that show how mirrors can be so thin that they aren't even flat standing still, much less scanning...

Best regards.

William Benner