On: NASA lays out how SpaceX will refuel Starships in low-Earth orbit

> If Starships actually use pressure difference, then it's not quite free flow, and geometrical positions aren't that important

That's how I'm reading it, analogous to a pressure-fed engine [1]. A separate, expendable gas supply pressurises the donor tank to force the propellant out and over.

[1] https://en.wikipedia.org/wiki/Pressure-fed_engine

Not sure if it's too similar to pressure-fed engine, that one works with a clear acceleration direction, which really helps to keep fuel at the local bottom.

> A separate, expendable gas supply pressurises the donor tank to force the propellant out and over.

But what if the fuel is a mix of liquid bubbles and the gas freely floating in the tank?

(BTW, on the different point, some pumps applicable here could be really simple, like analogous to https://en.wikipedia.org/wiki/Vacuum_ejector#Water_aspirator)

> that one works with a clear acceleration direction, which really helps to keep fuel at the local bottom

I believe the plan is to impart a gentle spin.

> some pumps applicable here could be really simple

The Venturi effect reduces pressure. It's definitely worth exploring. But you'd need to test it to explore where it causes volumetric effects, e.g. by exacerbating boil off.

> I believe the plan is to impart a gentle spin.

That would be something new. And for a partially refueled stage, that's a lot of mass to spin. How fast? the speed of spinning could be important too.

> that's a lot of mass to spin. How fast?

No clue! But I think they were targeting mili-g acceleration--could probably be achieved by the boil off alone.

Heating the donor by exposing it to sunlight while shading the receiver should provide some pressure differential for no added power. Gentle rotation would start with the center of gravity wihin the donor, but as the transfer goes on, it'd move to the middle of the two rockets and stop the transfer at that point.