Author Topic: Did SpaceX (or Any Other Company) Ever Consider Minimum-Cost Design?  (Read 40121 times)

Offline Lars_J

  • Senior Member
  • *****
  • Posts: 6160
  • California
  • Liked: 677
  • Likes Given: 195
They day someone actually flies a MCD launcher and shows that it actually saves $$$ - *Then* I'll believe that there is some actual "minimum cost" with MCD.

Offline jongoff

  • Recovering Rocket Plumber/Space Entrepreneur
  • Senior Member
  • *****
  • Posts: 6807
  • Lafayette/Broomfield, CO
  • Liked: 3987
  • Likes Given: 1684
The only one of those things other companies haven't used is Ethernet.

One interesting aspect of wat Beal did, pressure-fed first stages, was not adopted by SpaceX. They did try cheap composite ablative nozzles, but upgraded to regenerative nozzles later. It would also be interesting to see someone field an operational orbital launcher with a kerosene peroxide first stage. Blue Origin presumably intends to do that eventually, but not with a minimum cost expendable.

I wouldn't be quite so sure that ablative nozzles are really cheaper than regen ones....

~Jon

Offline Lars_J

  • Senior Member
  • *****
  • Posts: 6160
  • California
  • Liked: 677
  • Likes Given: 195
Yeah... SpaceX certainly gave ablative engines a shot. (weren't F1 #1 and #2 launched with an ablative Merlin?)

And now they seem to use regen engines whenever they can. All 12 regen engines that have flown (12?) appear to have run flawlessly.

Offline jongoff

  • Recovering Rocket Plumber/Space Entrepreneur
  • Senior Member
  • *****
  • Posts: 6807
  • Lafayette/Broomfield, CO
  • Liked: 3987
  • Likes Given: 1684
A couple of thoughts in this conversation, since I used to be a huge fan of MCD back in the day (my undergrad was in Manufacturing Engineering):

1-One of Arthur's points was that if there is a price vs. performance continuum, that it was better to let the lower stages be simpler and the upper stages more complex, because a weight of performance improvement on the upper stage is worth a lot more than one on the lower stage.

2-It should be noted that in reality, a) that curve is not continuous--you occasionally get situations where a higher performance solution might actually be cheaper than a lower performance one, b) that curve should be shifting as manufacturing and materials technologies improve over time.  What was once really expensive and a bad trade for a first stage might now be cheap enough that it's the new optimum.

3-MCD like "Lean Manufacturing" is often used more as a buzzword than an actual rigorous engineering discipline.  Talk to almost any manufacturing facility in the world, and they'll claim they're lean...even if the guys who pioneered the field would gag or be rolling over in their graves.

4-Arthur also pointed out that if the first stage was reusable/recoverable, that it would shift the optimal level of complexity higher.  Ie if you can easily reuse a stage several times, you can afford to spend more money building the stage, because you can amortize its cost over more flights.  I think that SpaceX thought that water recovery of their stage was going to be easier than it has turned out, so maybe their design optimized out at higher performance for the stage than would have had they been more conservative about reusability.  (note, I think that reusability is quite feasible, just that the approach they took was one that made it really hard/expensive to incrementally test, ie that there are better ways to do things if you really want to be reusable).

5-Building on the "the curve of cost vs. performance is not continuous" bit, there are new technologies coming out all the time that fit into this category.  Exploiting these discontinuities has been one of my big focuses.  For instance, take Steve Harrington's pistonless pump idea.  It's really not much more complicated than a pressure-fed system but starts getting you closer to the performance of a pump-fed system.  Such a system could easily be more optimal than the technologies to either side of it on the "performance" continuum.

6-There are ways to taking complex systems and find ways of simplifying them substantially while only sacrificing a bit of performance. 

7-The book LEO on the Cheap, which is related to the MCD approach, did mention another way of reaching low-cost than the big dumb booster approach--having the first stage be a "simple reusable"...that's the approach I'm most interested in.

8-Getting back to your original question, I think that looking at SpaceX's early years, I think they started out with an MCD approach and kind of morphed from there.  Whether that's an actual repudiation of the approach, differences of opinion (cause lets face it, as much as we'd like to pretend that engineering is an exact science, it isn't--we rarely have enough information to really be as sure about stuff as we engineers tend to pretend we are), or a mistake on SpaceX's part is left as an exercise to the interested reader.

That said, while I may not 100% agree with SpaceX's approach (or I would've been working there for several years now), I think they've got several really good elements, a truly amazing team, and I hope things work out for them.  The more people trying more approaches the better, IMO.

~Jon

Offline jongoff

  • Recovering Rocket Plumber/Space Entrepreneur
  • Senior Member
  • *****
  • Posts: 6807
  • Lafayette/Broomfield, CO
  • Liked: 3987
  • Likes Given: 1684
Yeah... SpaceX certainly gave ablative engines a shot. (weren't F1 #1 and #2 launched with an ablative Merlin?)

And now they seem to use regen engines whenever they can. All 12 regen engines that have flown (12?) appear to have run flawlessly.

Well, part of the problem is that pintle injector engines, particularly high-pressure pintle injector engines aren't always super forgiving on ablatives.  They had a pretty nasty failure several years back while doing intentionally off-nominal runs (which got leaked by an ATK guy who used to frequent here, and which resulted in the only time I ever got a comment by Elon on my blog), and very shortly after that they announced their regen engines.  Quite frankly, having seen them up close, they're beauties!  Regen engines give you a lot more "knobs" to deal with hot streaks and several other injector-related issues.

Having built a few regen cooled pintle injector engines, I can say I definitely think regen's a better match for high-pressure pintles than ablatives.  As I was saying in my previous post, manufacturing technologies move on.  I think it may not be too much longer before there are regen fabrication options that are cheaper and better performance than ablatives...I know of at least one that I'm working on, and I've seen a few others that are promising.

~Jon

Offline Lars_J

  • Senior Member
  • *****
  • Posts: 6160
  • California
  • Liked: 677
  • Likes Given: 195
Jon (or anyone else who knows), this may be going slightly off-topic, but you may be a good person to ask:

Whenever people come up with new ideas for low-cost launch vehicles that drastically will lover the $$$/lb to orbit, most of those proposals seem to have ablative engines in common. But what is the actual ratio of successful production engines that are ablative vs. regen? (RS-68 comes to mind as a recent example, but is it the exception to the rule?

Also, with your work dealing with primarily reusable engines (I assume), isn't regen pretty much required? How would you accomplish an ablative reusable engine?
« Last Edit: 09/05/2010 06:47 am by Lars_J »

Offline madscientist197

  • Regular
  • Full Member
  • ****
  • Posts: 1014
  • Liked: 6
  • Likes Given: 0
RS-68 isn't entirely ablative -- just the nozzle. The throat and chamber are regen.

Ablative might be okay for a minimum cost expendible, but if you're looking at reuse then you probably at least want the chamber and throat to be regen.
John

Offline jongoff

  • Recovering Rocket Plumber/Space Entrepreneur
  • Senior Member
  • *****
  • Posts: 6807
  • Lafayette/Broomfield, CO
  • Liked: 3987
  • Likes Given: 1684
Jon (or anyone else who knows), this may be going slightly off-topic, but you may be a good person to ask:

Whenever people come up with new ideas for low-cost launch vehicles that drastically will lover the $$$/lb to orbit, most of those proposals seem to have ablative engines in common. But what is the actual ratio of successful production engines that are ablative vs. regen? (RS-68 comes to mind as a recent example, but is it the exception to the rule?

There have been a number of ablative engines flown...like on the lunar module for both the ascent and descent engines, Kestrel, Merlin 1A, RS-68 (which as ms pointed out is only ablative for the nozzle section), and the AJ-10 which has flown on every Delta-II second stage.  So, I guess you could say they're relatively common on production engines.  Not as common as regen, but not totally unknown.  Typically they seem to be used with pressure-fed stages.  Actually, for pressure fed engines, I can't think of too many regen cooled systems other than Masten's.

Personally, I've never been a fan of ablative cooled systems.  You have no way of building in margins other than making the whole thing thicker, which means it doesn't handle hot-spots well.  Which is a problem, especially for pintle engines and throttleable engines.  Not to mention the hassles associated with testing.  It really wasn't that hard to make regen cooling work, especially for moderate cooling loads.  We did a passable job with only a few hickups ever, even with only one engineer working on it (me) who was busy doing about half a dozen other systems at the same time...Personally, for engines like RS-68 I have a hard time seeing what the big deal is with doing a regen extension, especially since the TP's and injectors were designed for such an assumption originally....but I digress.

Basically, it's an idea that comes up a lot, but I don't think it saves you a ton on cost, but it does save you a bit on pressure drop...but if you have any sort of pump, it seems kind of silly.  Right now for instance, I'd be surprised if an ablative chamber was actually any cheaper for SpaceX to manufacture than their current regen chambers.

Quote
Also, with your work dealing with primarily reusable engines (I assume), isn't regen pretty much required? How would you accomplish an ablative reusable engine?

Well, it depends.  You could make the ablative section removable and swappable after flights.  Make it thick enough for enough relights to cover the mission duration and such...but meh...ablative for RLVs?  We looked into that before we thought we could make regen cooling work, but now in hindsight...it just seems like the wrong tool for the job.

Personal biases come free of charge in this case.

~Jon

Offline Antares

  • ABO^2
  • Senior Member
  • *****
  • Posts: 5181
  • Done arguing with amateurs
  • Liked: 371
  • Likes Given: 228
for engines like RS-68 I have a hard time seeing what the big deal is with doing a regen extension, especially since the TP's and injectors were designed for such an assumption originally.

The RS-68 FTP was designed for the additional flow resistance of pushing fuel through a regen nozzle the size of the existing nozzle?
If I like something on NSF, it's probably because I know it to be accurate.  Every once in a while, it's just something I agree with.  Facts generally receive the former.

Offline jongoff

  • Recovering Rocket Plumber/Space Entrepreneur
  • Senior Member
  • *****
  • Posts: 6807
  • Lafayette/Broomfield, CO
  • Liked: 3987
  • Likes Given: 1684
for engines like RS-68 I have a hard time seeing what the big deal is with doing a regen extension, especially since the TP's and injectors were designed for such an assumption originally.

The RS-68 FTP was designed for the additional flow resistance of pushing fuel through a regen nozzle the size of the existing nozzle?

I probably should say that I'm going off of old, third-hand knowledge on this, so I could be wrong.  I thought it was, but maybe someone closer to that particular project could shed some light on if I'm misremembering things? 

That said, I've been working on some tricks for the nozzle that could get the pressure drop down quite a bit from a traditional tube-wall nozzle (while also being lighter, cheaper, and easier to fabricate)...

~Jon

Offline mmeijeri

  • Senior Member
  • *****
  • Posts: 7772
  • Martijn Meijering
  • NL
  • Liked: 397
  • Likes Given: 822
I wouldn't be quite so sure that ablative nozzles are really cheaper than regen ones....

Not cheaper to build or not cheaper to use? I think Beal's ablative composite nozzles were very cheap. I don't know about their performance, and his launchers never flew. It would be really interesting to know. The example of SpaceX, who started with ablative and then moved to regenerative is an indication (but not proof) that maybe they are not such a good idea after all.

But starting with something simple is logical enough. It's what von Braun did. It's what the New Space companies are doing. Didn't you guys start with IPA/LOX, ablative, pressure-fed, blow-down because it was a good place to start, only adding regenerative cooling later? Starting with a LOX/LH2 (or tripropellant!) staged combustion engine would have been the other extreme. Similarly, Paul Breed started with a four bipropellant engine vehicle and quickly switched to a single noncryogenic, nontoxic monopropellant engine vehicle because it was an easier place to start, only moving back to bipropellant once he had got the monopropellant vehicle to work. You can tell by his blog posts that Breed really understands incrementalism. I wonder if this is because he is a software guy. Of course Masten and Greason know all about incrementalism too.

So even if MCD doesn't work out it may still be a very good thing to try it, simply from the point of view of incrementalism. And also from the point of view of producing more data points about something that might work. Even if it doesn't work, that's still good to know.
« Last Edit: 09/12/2010 03:40 pm by mmeijeri »
Pro-tip: you don't have to be a jerk if someone doesn't agree with your theories

Offline Antares

  • ABO^2
  • Senior Member
  • *****
  • Posts: 5181
  • Done arguing with amateurs
  • Liked: 371
  • Likes Given: 228
This is a SWAG, but I think part of the reason that SpaceX went regen was that the business case for building the ablative in-house didn't close.  The unit price might've been cheaper, but they would have needed the capital to work with that particular material when nothing else in their product line needs it.  Or, they would have had to contract it out, which they don't seem to like.
If I like something on NSF, it's probably because I know it to be accurate.  Every once in a while, it's just something I agree with.  Facts generally receive the former.

Offline mmeijeri

  • Senior Member
  • *****
  • Posts: 7772
  • Martijn Meijering
  • NL
  • Liked: 397
  • Likes Given: 822
Actually, for pressure fed engines, I can't think of too many regen cooled systems other than Masten's.

The Shuttle OME and Ariane Aestus have regeneratively cooled chambers and radiatively cooled nozzle extensions.
Pro-tip: you don't have to be a jerk if someone doesn't agree with your theories

Offline jongoff

  • Recovering Rocket Plumber/Space Entrepreneur
  • Senior Member
  • *****
  • Posts: 6807
  • Lafayette/Broomfield, CO
  • Liked: 3987
  • Likes Given: 1684
But starting with something simple is logical enough. It's what von Braun did. It's what the New Space companies are doing. Didn't you guys start with IPA/LOX, ablative, pressure-fed, blow-down because it was a good place to start, only adding regenerative cooling later?

Nope, we never did ablative at Masten.  We went straight from heatsink firings (to dial in injector and chamber geometry stuff) to regen cooled.  Mind you, nobody but Ian Moore thought it would work the first time.  But it did.  And also all of our systems have been regulated pressure systems.  They do blow down a bit near the end of the run, but the first 1/2-2/3 of the flight is at a constant tank pressure. 

~Jon

Offline jongoff

  • Recovering Rocket Plumber/Space Entrepreneur
  • Senior Member
  • *****
  • Posts: 6807
  • Lafayette/Broomfield, CO
  • Liked: 3987
  • Likes Given: 1684
This is a SWAG, but I think part of the reason that SpaceX went regen was that the business case for building the ablative in-house didn't close.  The unit price might've been cheaper, but they would have needed the capital to work with that particular material when nothing else in their product line needs it.  Or, they would have had to contract it out, which they don't seem to like.

I think it had more to do with reliability.  High pressure pintle injectors and ablative chambers are a risky mix.  The timing of the switch to regen was very shortly after news came out of a fairly impressive failure during an intentionally off-nominal test.  In fact, I think it was Generic Username who posted it here on NSF that triggered the reaction.

The cost and difficulty of in-sourcing it may have also played into the decision, but my guess is that reliability and performance were just as big of concerns.

~Jon

Offline mmeijeri

  • Senior Member
  • *****
  • Posts: 7772
  • Martijn Meijering
  • NL
  • Liked: 397
  • Likes Given: 822
Nope, we never did ablative at Masten.  We went straight from heatsink firings (to dial in injector and chamber geometry stuff) to regen cooled.  Mind you, nobody but Ian Moore thought it would work the first time.  But it did.

Interesting. Did you use a double wall or something a bit more advanced? Or is that an impertinent question? ;)

Quote
And also all of our systems have been regulated pressure systems.  They do blow down a bit near the end of the run, but the first 1/2-2/3 of the flight is at a constant tank pressure. 

Am I confusing this with what Paul Breed did or am I just misremembering what I read?

Anyway, thanks for all that information! Straight from the horse's mouth, it doesn't get much better than that. One of the great things about continuous integration.
Pro-tip: you don't have to be a jerk if someone doesn't agree with your theories

Offline mmeijeri

  • Senior Member
  • *****
  • Posts: 7772
  • Martijn Meijering
  • NL
  • Liked: 397
  • Likes Given: 822
High pressure pintle injectors and ablative chambers are a risky mix.

That's a very interesting piece of information. Wasn't the Apollo LM ascent engine designed for maximum simplicity in order to maximise safety? Would it have been better to use a regeneratively cooled engine? And was this known in the sixties?
Pro-tip: you don't have to be a jerk if someone doesn't agree with your theories

Offline jongoff

  • Recovering Rocket Plumber/Space Entrepreneur
  • Senior Member
  • *****
  • Posts: 6807
  • Lafayette/Broomfield, CO
  • Liked: 3987
  • Likes Given: 1684
Nope, we never did ablative at Masten.  We went straight from heatsink firings (to dial in injector and chamber geometry stuff) to regen cooled.  Mind you, nobody but Ian Moore thought it would work the first time.  But it did.

Interesting. Did you use a double wall or something a bit more advanced? Or is that an impertinent question? ;)

It's on the blog, so it's not impertinent.  We used a Chamber-Saddle-Jacket config similar to what XCOR and Swiss Propulsion Lab have used.  Not very relevant to Merlin sized engines, but for small engines it's great.  Allows you to take the thing apart for post-firing inspections.  That said, we're drifting off-topic again...

Quote
Quote
And also all of our systems have been regulated pressure systems.  They do blow down a bit near the end of the run, but the first 1/2-2/3 of the flight is at a constant tank pressure. 

Am I confusing this with what Paul Breed did or am I just misremembering what I read?

I think you're thinking of what Carmack did.  He never was able to get regen cooling to work (though after I did all the work at Masten making that work, I wouldn't be surprised if Ben was able to make something like that work for them in the future), and only did regulated feed pressure when they needed more performance for the NGLLC.

Quote
Anyway, thanks for all that information! Straight from the horse's mouth, it doesn't get much better than that. One of the great things about continuous integration.

Yup...though any further questions probably ought to be taken over to the Masten thread...though I guess the title says "Did SpaceX (*or Any Other Company*)...

~Jon

Offline jongoff

  • Recovering Rocket Plumber/Space Entrepreneur
  • Senior Member
  • *****
  • Posts: 6807
  • Lafayette/Broomfield, CO
  • Liked: 3987
  • Likes Given: 1684
High pressure pintle injectors and ablative chambers are a risky mix.

That's a very interesting piece of information. Wasn't the Apollo LM ascent engine designed for maximum simplicity in order to maximise safety? Would it have been better to use a regeneratively cooled engine? And was this known in the sixties?

For lower pressure engines it's not as big of a deal...unfortunately this is getting into an area of info that might be ITAR sensitive...sorry.

~Jon

Offline mmeijeri

  • Senior Member
  • *****
  • Posts: 7772
  • Martijn Meijering
  • NL
  • Liked: 397
  • Likes Given: 822
Thanks!
Pro-tip: you don't have to be a jerk if someone doesn't agree with your theories

 

Advertisement NovaTech
Advertisement Northrop Grumman
Advertisement
Advertisement Margaritaville Beach Resort South Padre Island
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
1