Well, I tried a third TS140P, and flattened out the base. It averaged 76C for Prime95, using the same settings in that as before. The main difference is that it's using only one fan (A TY143) and it's on a slower fan curve than the one I used to test the D15, so that's a bit of a handicap for it to achieve the same temps (at a lower noise value) than the D15.
It also takes up less room, and the fan clips are far easier to work with. I'll probably clip on a second fan and see how it does. Good heatsink when it makes good contact.