Fig.1

Shows the 4-,2-, and 1 segment capacitors

The surfaces were cleaned with P500 then P1200 , the wires and plates were pre-soldered prior attachment using the Kristall 511

The surfaces were cleaned with P500 then P1200 , the wires and plates were pre-soldered prior attachment using the Kristall 511

... targeting the 4 to 20 pF range. I actually mishit a lot testing the 16µm paper for dielectric . . .

So i varied the sizes and dielectrics and attempted to collate some science out of that

The physical dimensions :

Fig.2

showing the capacitor plate omitting the sharpened edge geometry

And the atats. -- had to recalibrate the Kyoritsu's KEW1018 measurements against the set of commercial capacitors from 15 to 180 pF -- here's what we got :

Fig.3

Shows the Cap.-meter re-calibration logic

Fig.4

Shows the Old and the New calibration curves -- the Red one shows the difference and actually is an optimistic error estimate for the measurement at X-axes plot capacity

Next -- matching the TEST and the THEORY (using the Old calibration formula - it takes an error trial to match the measurement to unknown actual capacity . . . ) :

Fig.5

-- as it might've been guessed -- the larger blades´ surface geometry (used carpet blades) falls far from the ideal - thus the inserted 16 micron dielectric has a lot of free space from it to the razor blade -- the lllliiiillll illustrative graphs show the dielectric-gap for ideal smooth surface and what the actual average of what it came out

Fig.6

(it actually "rolls toward you" close your eyes momentarily - keeping that in mind - if your brain suggest a different solution !) -- shows in relative scale the last assembled 6pF capacitors with double-layer 2× 240 micron dielectric -- these capacitors had the least capacity deviation inside the set -- which is obvious because the µm variations won't have much effect for average plate distance

[Eop]