After a small tweak, I was able to get my idea of RIMS+loop to work basically as I envisioned it in a water test. A ball valve controlling the loop flow rate more linearly (compared to the butterfly valve used in my first revision) proved to be essential.
I’ve made some progress with the design changes I detailed in post 1 of this series. I was able to implement the design, but in testing, it failed to give me the result I was hoping for. It has informed a revised design that I hold hope for working, however.