How do you test-fire vacuum engines in a vacuum chamber? Why doesn't it fill up almost instantly from the exhaust and spray (deluge) water?

How are vacuum optimized engines tested without disintegrating them? is an excellent question as it's attracted quite a number of informative and diverse answers.

@Uwe's answer informs us that there exist at least one or two low-pressure test chambers where engines can be fired for tens of seconds and the exhaust cleverly managed/diverted to maintain a pressure below 50 Torr for tens of seconds which is pretty amazing!

From NASA Plum Brook Station In-Space Propulsion Facility Test Stand Characterization Hot Fire Test Brian K. Jones, John C. Zang, Hal F. Weaver, Nicholas A. Connelly, and Gerald M. Hill, NASA Glenn Research Center, Plum Brook Station, Sandusky, OH, 44870, United States

Abstract

A test facility modification to enable small scale altitude propulsion testing at the NASA Glenn Research Center’s In-Space Propulsion (ISP) Facility was verified with a hot fire test campaign. As the facility’s primary steam supply system undergoes refurbishment, the alternate facility configuration, known as the “vacuum accumulator” mode, would enable rocket engine testing up to 10,000 lbf thrust. The NASA Johnson Space Center developed the vehicle for the verification test campaign: the Integrated Cryogenic Propulsion Test Article (ICPTA). Constructed primarily from assets of the former Morpheus Project, the ICPTA provided an integrated liquid oxygen (LOX) / liquid methane (LCH4) propulsion system including a 2,800 lbf thrust main engine. The ISP Facility’s vacuum accumulator configuration leveraged the large test volume of the facility and a diffuser insert to maintain altitude conditions. During hot fire, the ICPTA main engine “started” the diffuser insert constructed for the test campaign. As a result, the test chamber upstream of the diffuser insert remained at altitude conditions throughout the hot fire. Upon engine shut down, a backflow deflector mitigated blow back into the test chamber by restricting the mass flow and redirecting it away from the test article. The test campaign successfully characterized the performance of the vacuum accumulator configuration. In addition, it provided an opportunity to collect data for an integrated LOX / LCH4 propulsion system in an altitude and thermal vacuum environment.

I see “vacuum accumulator”, "diffuser" and "backflow deflector" but after a read through I still don't understand how this actually works.

Question: How do you test-fire vacuum engines in a vacuum chamber? Why doesn't it fill up almost instantly from the exhaust and spray (deluge) water?

Figure 1. A cutaway view of the In-Space Propulsion Facility. In the center of the test chamber, the facility’s diffuser provides the interface between the test chamber and the spray chamber. The steam ejectors are located outside of the test building in the yard on the right.

Figure 9. Hot Fire 2.6 test chamber and spray chamber pressures during a 27 second main engine hot fire test in the lowered position with the backflow deflector.

• Useful links ejector and Venturi Effect. But altitude simulated test for rocket engines are not listed in the list of applications.
– AJN
Oct 23 '21 at 7:24
• "the exhaust cleverly managed/diverted to maintain a pressure below 50 Torr for tens of seconds which is pretty amazing" - it's even better than that, the test chamber pressure goes down during the test due to the exhaust working like another steam ejector! Oct 23 '21 at 11:39
• Don't recall there being any "spray (deluge) water" in these chambers. Oct 23 '21 at 12:31
• @OrganicMarble somehow I forgot to add the url to the Plum Brook link from Uwe's answer. I've added it now. From section II A, starting middle of page 4: "Beyond the diffuser, the rocket exhaust enters the spray chamber. The spray chamber’s 67 feet diameter and 120 feet depth hold up to 1.75 million gallons of water, whichs fills the chamber roughly halfway. Four 2,000 horsepower vertical turbine pumps move the water up to an array of spray nozzles at a rate of 56,000 gpm each. The spray nozzles then evenly distribute the water to cool the rocket exhaust." There's more...
– uhoh
Oct 23 '21 at 13:23
• Awesome, thanks! Oct 23 '21 at 13:31

I got a book from 1963, "Wie funktioniert das?", how things work:

Translation:

 Treibdampf                                    high pressure steam inlet
Kühlwasser                                    cooling water inlet
Auspuff                                       exhaust pipe
Saugstuzen                                    pump inlet side
Vakuum                                        vacuum
Kühlwasseraustritt                            cooling water outlet
Zwischenkühler                                intercondenser
fünfstufiges Dampfstrahl-Luftsauger-Aggregat  steam ejector vacuum pump with five stages


Source

For a vacuum of about 1 kPa, a steam ejector with two or three stages will do. Using five stages, a vacuum pressure of 4 Pa is possible. But the upper stages should be much bigger than the lower stages, the volume flow rate increases eight- to tenfold for each step.

It is an application of the Bernoulli's principle and the Venturi effect, the vacuum ejector:

For an engine test stand you need very large steam ejectors to keep up with the engine. The hot steam is much colder than the engine exhaust and will absorb a lot of heat. Of course much more steam than the exhaust of the engine under test is needed.

A steam ejector is build without any moving parts. A large steam generator may hold enough hot water for an engine test for tens of seconds, but heating all that water before the test may take many hours.

• I'm surprised you can get down to 4Pa with a vacuum ejector. I mean, sure you have a (presumably supersonic given the construction) nozzle, but thermal velocity of atoms in a gas has a fair tail.
– TLW
Oct 23 '21 at 19:40
• @TLW A steam ejector with 6 stages may produce 0.4 Pa.
– Uwe
Oct 24 '21 at 9:19
• This is really all so amazing, thank's for the great answer!
– uhoh
Oct 30 '21 at 23:13