[hydrogen7800]

Expert here. I'm a longtime lurker on HN, but registered just for this comment. I design aerospace pressure vessels for a living. As showerst alluded to, there are often things called "propellant management devices" or PMDs. There are a few reasons these might be used. In zero G, you need the propellant sump to remain wetted with propellant; it would be bad to ingest the pressurant/ullage gas into your engines. Some PMDs allow the pressurant and propellant to occupy the same volume, and will use surface tension devices inside the tank to direct liquid to the sump. These can be screens, vanes, channels, etc. In other cases, the pressurant and propellant are kept separate by a bladder. These bladders can be rubber, or even metal. The propellant mass is usually a large proportion of a space vehicle's mass, and you can't have that much sloshing around when you need dynamic control. A metal diaphragm keeps the propellant more or less static, and its center of mass in a predictable location.

As for the pressurant gas dissolving into the propellant in non-PMD tanks, I don't know enough about that. I imagine the solubility of He (it is usually helium) in these propellants is either accounted for, or negligible.

Edit: P.S. Software engineering is mostly foreign to me and much of HN content is over my head, but I like he level of discourse here. So, when a topic came up that I could actually contribute to, I jumped.

[kortex]

Nice writeup! Former chemist here. Helium is indeed poorly soluble in most liquids. It's actually used to sparge (purge by bubbling through) solvents to drive out other dissolved gasses.

This is because Helium has very weak intermolecular forces due to its electronic symmetry. For that same reason, it's also as close to an ideal gas, giving you the most pressure/volume bang for your mass buck (only hydrogen is better, and that's bad for oxidizer tanks for obvious reasons).

But this also limits the ability to cryogenically condense helium, which would improve storage density. But you really don't need much in turbopump fed engines.

[mLuby]

Love how much expertise there is here; thanks for adding yours!

When you say "metal bladder/diaphragm" is that a sliding wall with propellant on one side and pressurant on the other? (I can't imagine how the seals in that would work.) Or do you mean the metal actually deforms in place?

[hydrogen7800]

Yeah, the metal deforms. They are single use devices. Imagine a spherical metal tank, with a thin metal hemisphere welded inside. The hemisphere would invert as the propellant is expelled. Hard engineering and manufacturing problems here. I think there have been piston tanks (which would be reusable with sliding elastomer seals. Not many elastomers play well with hypergolics), but they would be very weight inefficient.

[MertsA]

Neat, that's one subject that's always bugged me as every article or video out there just handwaves away settling of propellant tanks in orbit with "ullage thrusters". No one ever mentions how you get a nice continuous feed to your ullage thrusters though.

[hydrogen7800]

Ullage thrusters are used exactly as you described, for tanks that do not have PMDs. For example, the large second or third stage main propellant tanks which need to be lit in zero G. On Saturn V, the ullage motors themselves were solid rocket boosters mounted to the aft end of the stage, and thus did not themselves need PMDs, nor access to the main liquid propellants. On some videos of stage separation, you can see these solids firing from the point of view of the jettisoned stage facing forward. (In fact, I think there were also retrograde solid thrusters at the fore end of the stages to aid in separation.) I suppose you could call ullage solid motors PMDs, too.

For an excellent deep dive into the systems, operations, and functions of the Saturn V, get a copy of "How Apollo Flew to the Moon" by W. David Woods.

[jacksonkmarley]

Awesome detailed answer! Much appreciated.