BBP,
Good analysis w/ good technical explanation.
I've done more than a few searches for studies of efficiency differences (COP) between R-22 & R-290, and R-290 comes out ahead in every study I've found.
Hacking a R-22 system straight across to R-290 works fine. The metering device may not be optimum, if you are using cap tubes. If you have an adjustable orifice device, like an adjustable TXV, you should be able to dial it right in.
Your study implies quite a few parameters that are different.
There are some free HX sizing programs in the Internet that will take into account the multi-variant factors, and you can dial right in the optimum components you need.
As I recall,
THIS_ONE did a pretty good job.
-AC