Large upper stage auxiliary propulsion system

[nicp]

Not so long ago a large Chinese upper stage was left to decay and land wherever it wanted to. Not for the first time I think.

Much more recently Starship flight 3 lost attitude control and had it actually been in orbit that would have done the same. Whether this was loss of ullage gas, electrical or computer failure I don’t think is known yet. Frankly after some weeks of thought that looks like a dog’s dinner to me.

People tend to not appreciate 100 tons of flaming metal dropping out of the sky on them.

Saturn had the APS system on the upper stage for roll control and ullage burns. This used storable propellants.

Now I know “the best part is no part” but no attitude control does not seem best to me.  A simple and reliable and most importantly a redundant attitude control system for any huge upper stage.
It’s added complexity I know but that much mass of derelict just heading where it wants is not clever.
Thoughts?

[edzieba]

Not so long ago a large Chinese upper stage was left to decay and land wherever it wanted to. Not for the first time I think.

Much more recently Starship flight 3 lost attitude control and had it actually been in orbit that would have done the same. Whether this was loss of ullage gas, electrical or computer failure I don’t think is known yet. Frankly after some weeks of thought that looks like a dog’s dinner to me.

People tend to not appreciate 100 tons of flaming metal dropping out of the sky on them.

Saturn had the APS system on the upper stage for roll control and ullage burns. This used storable propellants.

Now I know “the best part is no part” but no attitude control does not seem best to me.  A simple and reliable and most importantly a redundant attitude control system for any huge upper stage.
It’s added complexity I know but that much mass of derelict just heading where it wants is not clever.
Thoughts?

Only necessary for the small fraction of large upper stages that reach orbit, rather than the majority that are on suborbital trajectories with designated drop zones (including the recent Starship launch, the handwringing being only out of ignorance of the publicly filed launch details). The CZ-5B leaves an upper stage in orbit because it has no third stage and injects he payload directly (payload does perform insertion burn itself), the regular CZ-5 does not, as like the majority of large upper stages it has further stages to perform final orbit insertion.
RCS also would do nothing for stages in higher orbits, e.g. the Centaurs that have burst and produced debris within their disposal orbits (post-passivation, so RCS would already have been vented anyway).

This seems like a solution looking for a problem.

[uhuznaa]

Not so long ago a large Chinese upper stage was left to decay and land wherever it wanted to. Not for the first time I think.

Much more recently Starship flight 3 lost attitude control and had it actually been in orbit that would have done the same. Whether this was loss of ullage gas, electrical or computer failure I don’t think is known yet. Frankly after some weeks of thought that looks like a dog’s dinner to me.

People tend to not appreciate 100 tons of flaming metal dropping out of the sky on them.

Saturn had the APS system on the upper stage for roll control and ullage burns. This used storable propellants.

Now I know “the best part is no part” but no attitude control does not seem best to me.  A simple and reliable and most importantly a redundant attitude control system for any huge upper stage.
It’s added complexity I know but that much mass of derelict just heading where it wants is not clever.
Thoughts?

Only necessary for the small fraction of large upper stages that reach orbit, rather than the majority that are on suborbital trajectories with designated drop zones (including the recent Starship launch, the handwringing being only out of ignorance of the publicly filed launch details). The CZ-5B leaves an upper stage in orbit because it has no third stage and injects he payload directly (payload does perform insertion burn itself), the regular CZ-5 does not, as like the majority of large upper stages it has further stages to perform final orbit insertion.
RCS also would do nothing for stages in higher orbits, e.g. the Centaurs that have burst and produced debris within their disposal orbits (post-passivation, so RCS would already have been vented anyway).

This seems like a solution looking for a problem.

In higher orbits this isn't an immediate problem since they don't decay within a long time. Leaving stages stranded there is still not a good thing (since they WILL be a source of debris over time) but this isn't a pressing problem.

But in lower orbits (like where dozens of Starship tankers will go and also all Starships launching Starlink satellites) this definitely is a problem, especially when you plan to launch lots and lots of them and your upper stage is a stainless steel monster. You absolutely don't want to have one of these stranded in a LEO orbit randomly decaying over days or weeks and then coming down in a totally random and uncontrollable place.

Like, I think every crewed LEO spacecraft has redundant (often combined) attitude control and deorbiting systems. The reason here is that you absolutely NEED to deorbit your crew in a timely and controlled manner, but while with a (uncrewed) Starship this is not about the safety of a crew but about the safety of people on the ground, at least potentially. Even one 100 tonnes steel Starship coming down only after weeks in a totally random place would feed a media frenzy and would be very, very bad PR. And for good reasons.

So, yes: I think SpaceX will not only have to demonstrate they can reliably do a targeted deorbit burn, they also will need some redundancy here for the case that during the 37th flight or so something will fail to work. "Fail early and often" is not really an option here when others may have to cope with what follows from your failure and not just your very own bottom line.

Just as with Crew Dragon by the way, SpaceX didn't do any "try, fail, rinse and repeat" approach with crews.

[zubenelgenubi]

Not so long ago a large Chinese upper stage was left to decay and land wherever it wanted to. Not for the first time I think.

The CZ-5B leaves an upper stage in orbit because it has no third stage and injects he payload directly (payload does perform insertion burn itself), the regular CZ-5 does not, as like the majority of large upper stages it has further stages to perform final orbit insertion.

Upper stages: one kg saved on the upper stage roughly equals one more kg to orbit.

Re: the PRC: Seems like a fundamental design flaw dictated by political masters who don't have to answer to their subjects--witness PRC launches dropping deadly poisonous hypergolically-propelled first stages on their own country, over and over and over again. 🤔  Ditto USSR/Russia.

(Further discussion of the previous paragraph belongs in "Space Policy.")

Also, it's the 2020's, not the 1960's.  Rocket engineering for the major players has moved far beyond just getting the payload to orbit being a tough-to-achieve goal.

This seems like a solution looking for a problem.

Agreed. 👍

[nicp]

Finally I'm typing on a confuser and not on a phone.
Right, I _really_ don't want this to be another Starship thread.
My point is that big upper stages, whether New Glenn, Starship etc need redundancy and reliability in deorbit -  which will require attitude control - to drop lumps of spacecraft into - say - the Pacific spacecraft graveyard where they have been going for decades. See Mir.
But the Chinese, and my imagined but not improbable variation of the recent Starship test flight (yes, a test I know) could have dropped not quite anywhere, but a lot of places.
My concern with Starship is I believe it showed no redundancy, to a point where I think it was careless - but, to repeat I don't want this to be a Starship thread. We have too many of those and I certainly can't keep up with 10% of them. (BTW I like Spacex)

Perhaps the subject should be "regulation of spacecraft of beyond X tons and controllability thereof given any possiblility of atmospheric re-entry and perhaps hitting Nic's house?" :-)

I'm actually serious. What happens after several tons of unguided space derelict drops onto New York? Or Rome?

[edkyle99]

Starship will ultimately have to have a robust, redundant attitude control system if it is to achieve its goals.  Propellant transfer requires long-term station keeping and multiple rendezvous/docking events.  Recovery requires retro burns and reentry, all under precise control. 

STS dedicated something like 19 of its 120 tonnes to RCS and OMS.  Along with two OMS engines there were 44 RCS thrusters.  They fired down to 70,000 feet during reentry.  S-IVB APU was  much lighter, at probably less than 1 tonne of the in-orbit vehicle's 62 to 89 tonnes (including payload) but only had to work for a few hours and did not have to do reentry burns.  Both STS and S-IVB APU used N2O4/MMH pressure fed.

Starship intends to go to do dozens upon dozens of rendevous-docking-refuelings in LEO and then go to the Moon and back, and reenter, and land.  It has to be under absolute control, even when the occasional thruster fails.  ITF-3 showed no signs of control, in my opinion.  Whatever was being tested was only a skeleton system, I think, and it worked not at all.  My guess is that SpaceX is working on an operational ACS that we will see in the future, hopefully - and it isn't ullage gas!

As for China's CZ-5B core stage, I see it as illegal in the modern era.  No large piece of hardware should be allowed to reenter uncontrolled in that manner.  Of course the US was guilty of the same offense with the SA-513 second stage and with Skylab and with the Skylab shroud.

 - Ed Kyle

[Solarsail]

So got to thinking about ways to do this with large rocket stages (ignoring space stations for now)...  I suspected that you could use continuous thrust with very small motors to reach a sufficiently small patch of the Earth for reentry.  (Counting the Pacific ocean as 'small' ...)

Constantly rotate against orbit, retrograde at apogee, straight down mid way between apogee and perigee, prograde at perigee, and straight up midway through ascent back to apogee.  If you come down steep enough to perigee, then you don't loose spend multiple orbits loosing apogee to drag...  So one big maneuver spanning multiple orbits.  Tried that out in Kerbal (with RSS/RO) and I'm surprised by how little delta-v that wastes.  Starting in a 305km circular orbit I flew to a 360 by 180km orbit twice...  Spent 56 m/s doing short impulsive burns, and 77 m/s on one 2h long burn.

Now for some serious armchair engineering;  Can a maneuver and corresponding thrusters for something like that be matched to the boiling rate for LH2 in the main stage tank?  Because something like the CZ-5B should have noticeable residuals left, if thrusters can be run off of the boil-off.

I suppose a related conversation for things built more like satellites or space stations is whether you could do this usefully with electric propulsion instead.

[Barley]

I suppose a related conversation for things built more like satellites or space stations is whether you could do this usefully with electric propulsion instead.

I suspect not.  Below some altitude a satellite are will always reenter "near" the next perigee so you can target point Nemo.  Above some higher altitude the satellite will not reenter on this orbit.  In between the satellite may or may not reenter.  You want to avoid this range of perigees for a targeted reentry.  The question is how wide the region of uncertainty is (given whatever knowledge you have of sunspots, satellite orientation etc. can be mustered.)  I suspect the difference in perigee to avoid the uncertain region is at least several tens of km.  Avoiding this from LEO would require a delta-V of several (~10) m/s in less than an orbit (~90 minutes).   mN thrusters are not going to achieve that for reasonable sized satellites.

To put it another way the determining factor is how well you can predict the reentry point when close to the boundary. This is a horribly ill-conditioned problem so you need to avoid the boundary by going from "clearly will not reenter soon" to "clearly will reenter at the next perigee" in less than one orbit.

[Solarsail]

Yes, that was what I was trying to accomplish by raising apogee at the same time.  Though I'm in over my head again on the math for that...  I figured that having a steeper approach to the atmosphere would help to dive into the atmosphere more abruptly...

Chemically fueled thrusters running for 2-8 hours might then amount to an in between of electric propulsion at the slow end, and the turbopump powered main stage engines at the other.

If it is only the perigee that matters...  then I think multiple impulsive burns at apogee are about 1/3 the dV, but places a noticeable minimum acceleration on the maneuver.

Hm.  If electric propulsion in LEO (as drag compensation) is too weak for a controlled deorbit, that would complicate some concepts of giant space stations or mega-structures built there...