I can see one glaring problem here. Primaluce is confusing the terms “mechanical distance” and “optical distance” in their description, which makes their entire explanation super confusing. So, let’s first get one thing straightened out here. The backfocus (called the back working distance [BWD] in optics) is defined by the distance from the rear of the baffle nut to the focal plane. That is simply a mechanical distance between two planes. The term “optical distance” simply refers to the equivalent mechanical thickness needed for the required light path including internal optical components. This term appears to be most specific to amateur astronomy and I personally wish that everyone in the industry would dump it. It is simply too confusing for almost everyone and it can be easy replaced with some very minor adjustments due to optical widows. It is very simple to say that the BWD of the reducer is 105 mm but if we include a 3 mm filter and (say) a 3 mm camera window, the required spacing will be 107 mm. That’s the distance that will be required from the rear of the baffle nut to the sensor, taking into account the two optical windows in the system. How hard is that?
Be that as it may, the problem that Primaluce has is that when they stack up all of their gear, the “optical thickness” of the assembled stack is greater than the 105 mm mechanical BWD they have with the reducer—by 10.8 mm. Note, that has nothing to do with any notion of “adding backfocus” by screwing things together. That concept is even more confusing than computing optical thicknesses!
They are solving the problem by simply screwing the rear of the baffle nut far enough inside the focuser until the sensor is in the correct position. They can do that because of the depth of the threads on the baffle nut and the size of the aperture through the focuser. They are simply overlapping parts to get it all squeezed into the required space.
At the end of the day, I personally never trust any of this stuff until I make a scaled drawing of all of the components showing the virtual and real focal points relative to where the sensor sits. Here’s an example from a CDK17 project I did a while back—and this particular drawing doesn’t even show all of the true ray paths through the window! They are simply represented by the offset center line and the displaced focal point. This is the best way to avoid mistakes, which is really important if you have to order expensive custom machined parts.
📷 image.png
Good luck with it!
John