Here's an update on this thread, at least the core issue with it.
My 9A has an IO360 with Bendix injection and no engine-driven fuel pump, using the SDS dual electric fuel pump module through an Andair duplex fuel valve. Previously I had Airflow Performance fuel pumps and practiced running a tank dry a few times during Phase I, and it recovered just fine.
Recently I changed the fuel pumps from the AFP to the SDS module, including the Borla fuel pressure regulator for return back to the tanks, and today was a good day to repeat the test for actually running the tank dry. I climbed to 9,000 above a local airport with very little traffic and an 8800' runway (KBPG) and ran my left tank dry with 10 gallons still in the right tank, remaining within easy glide range of the airport. I saw the fuel pressure start to wobble from it's normal 40 psi and got the Dynon "low fuel pressure" warning at 25 and kept going - fuel pressure would wobble up and down as it caught slugs of liquid and vapor, and then about 25-30 seconds into it the pressure dropped below 15 and the engine surged for a few seconds as pressure fell through 5-6 psi, and the engine went cold. I left it as it was for another 15-20 seconds to make sure I was pulling a good slug of air into the system (since that seems to be the failure mode everyone is afraid of) and then switched tanks. Nothing at all on the fuel pressure for a good two-count, then at 3 seconds suddenly full pressure and full engine power was restored, zero hiccups past that point. This was all done using one single pump with the other pump in reserve. I do have an AFP purge valve installed at the injector divider, but did not touch it for this test.
A couple things out of this test - even if you've got your head up and locked, and accidentally arrive at an empty tank, you've got 20-ish seconds after the low-fuel-pressure warning with this setup to change tanks before the pressure drops enough to notice any difference in the engine. Second, recovery with the fuel pumps, even if they are slugged with vapor, occurs very rapidly and should not be feared. With that in mind I will not be shy about running one tank completely dry in order to maximize range if the situation called for it.