220422-1357 Astromania binoviewer reverse collimation test

I observe that the BV is at the core optically simple, and collimation should be easy to test; after all it is all just flat optical surfaces (mirrors and/or prisms). A ray coming into the nosepiece should be divided into two *perfectly parallel* rays coming out the eyepiece side. Conversely, perfectly parallel rays coming into the eyepiece barrels should be merged into a single ray coming out of the nosepiece.

My theory and simple test is illustrated in the image. I simply view through the BV in reverse: no eyepieces, no telescope! Only the prisms or mirrors within the BV are affecting the image. I hold them like binoculars, and point them at anything, looking into the *nosepiece* (with one eye). There I see a merged image, 1:1 magnification. Theoretically, if the BV collimation is good, distant objects should look perfectly aligned (i.e. should not look doubled).

You would think that the apparent vertical doubling in these pictures would indicate poor collimation in the BV, resulting in vertical misalignment of the stereo image when the BV is used in the telescope. But in the case of this particular BV, the result at the telescope was a good stereoscopic image with no apparent misalignment. This has me puzzled. If you have any insights, please reach out.