I have tried to model Peter's latest acquisition, but it proved to be quite tough. According to Roger Ceragioli's essay on APO's, Taylor's original version used Schott "ordinary/regular" glasses type O.543, O.164 and O.374. Later O.164 got replaced by O.569 because of rapid deterioration of the original middle short flint element.
Any of those are quite hard to find any information on - try hard as I could, virtually no information could be found that could be used to analyze this triplet.
In the end, it was Peter himself that provided data on some of these glasses, enough to get me started.
Using data provided, combination of O.543, O.569 and O.374 and Peter's measured radii and thicknesses would get close, but nowhere near the satisfactory performance. Then I allowed R6 (which is shallowest) to vary, and then also BFL to vary as well a little. That improved things. The last variation was to allow for some tiny conic/aspherization on R1 (which I think was common in those days. It would also explain zones Peter saw). I think melt variations for Peter's refractor would account for differences in radii and BFL.
The final result is I think quite good for a lens that uses no Fluorite type glass. Its secondary spectrum (C to F) amounts to better than 1/10,000 of the focal length, which puts it squarely into a "real APO" territory. Even with much wider wavelength spread (440 to 700nm) it is still very well corrected. It is also completely flat field for all practical purposes, so any eyepiece will only be limited by its own aberrations. I think a comparison with a more modern (and faster) APO would prove quite interesting.
Edited by bratislav, 06 April 2020 - 11:44 AM.