Trebor1
•
Regular Member
•
Posts: 401
Re: A 3D-Printed Ball Lens Mount
ProfHankD wrote:
I purchased a 30mm "clear crystal ball" and made a 3D-printed mount for it to be used with a Canon PowerShot ELPH180 (or similar camera):


The mount has three printed parts. The mount fits the ball in the front, which is then locked in place by screwing the lock part over it. The whole rig is then connected to the PowerShot by screwing-in the printed 1/4-20 bolt to the camera tripod socket. It looks very cute....
I purchased a 30mm "clear crystal ball" for $4 on eBay, shipped from the USA. They start around $2 from China. Anyway, the one I got has multiple nicks, a couple of significant inclusions, and isn't all that close to being spherical.
The image quality is unfortunately as poor as expected. A door peephole produces better image quality. However, you can judge for yourself:
This is a flat poster -- there just isn't enough DoF for imaging the ball


Obvious distortion from this ball not being circular
Maybe you need bigger balls? It appears there are a number of higher quality K9 (Borosilicate crown glass) balls from various suppliers, in the 60mm-120mm diameter range. There is even one available from Rollei, for additional kudos/poseur value but the original (Chinese) manufacturer may well be the same?
https://www.rollei.com/products/photo-accessories/lensballs/
Although a properly designed fisheye is always going to outperform a simple DIY ball lens, how would a simple, for fun lens, perform? When based on:
1. The R W Wood camera, front pinhole with water filled space behind, or instead using a hemispherical ball lens, facing the sensor as in the Franke widefield camera.
2. A design based on the Tisse camera: convex-plano front lens element, pinhole or aperture stop, rear hemispherical ball element. Might be some advantage in filling the pinhole/aperture stop with canada balsam (or similar), with a refractive index, close to that of glass, in order to minimise any adverse effects from total internal reflection?
See page 1109, in the following link for design schematics, (b) looks interesting:
https://hal-iogs.archives-ouvertes.fr/hal-00700221/document