SpaceX Merlin 1C Engine

[Danderman]

meiza - 15/11/2007  8:34 PM  I'm curious how the F9 has comparable performance to Atlas V 401 even though their first stage has same thrust but less performance and the upper stage is much heavier too... They must have a better mass fraction.

I don't know how they get a better mass fraction than a Centaur. I will say that LEO performance of Atlas V suffers due to the single engine Centaur.

 

[jongoff]

Danderman - 15/11/2007  10:10 PM

meiza - 15/11/2007  8:34 PM  I'm curious how the F9 has comparable performance to Atlas V 401 even though their first stage has same thrust but less performance and the upper stage is much heavier too... They must have a better mass fraction.

I don't know how they get a better mass fraction than a Centaur. I will say that LEO performance of Atlas V suffers due to the single engine Centaur.

Well, by using a denser propellant it actually isn't too hard to beat the Centaur.  Centaur has awesome mass fraction for LOX/LH2, but the low bulk density of the propellants really hurts it.  Also having a high thrust upper stage engine like a Merlin (with vacuum optimized nozzle) probably helps somewhat too.

Another interesting thought is that Musk mentioned that they're looking at a turbopump upgrade next year that would increase thrust and T/W by something like 20-25%.  It could possibly be that the F9 numbers they're quoting are assuming that better, upgraded version of the Merlin-1C, not the current version that just finished development.  I don't have any special insight on this, but that would be my guess.

~Jon

[Frediiiie]

The turbo pump upgrade is evidently in the Merlin 1e engine not due till 2009.
I've no idea what upgrades are in Merlin 1d

[JIS]

meiza - 15/11/2007  9:17 PM

There were two, one by astronautix (old) and one by spacex.
You can read Elon's old updates, I think there he mentions how close they were to reach efficiency, thrust and ISP figures when the original Merlin 1 was developed. I think they were pretty close but somewhat below. I don't remember what Falcon I figures were released when. Falcon 9 wasn't even planned back then.

It's been an ongoing development, both the engine and the rockets. It's not a simple question. Payload, margins and designs have probably fluctuated a lot.

They had problems with ablative nozzle. With regenerative nozzle the ISP and thrust issue go away. They can try to squeeze maximum from the current Merlin design and production line.

[meiza]

Yeah but high mass fraction doesn't help that much since the second stage even with infinite tankage ratio would lose to a lox/lh2 stage in mass (payload needs a certain amount of propellants to accelerate to a certain delta vee even if tankage and engine mass is zero). And thus a bigger first stage is needed.

That's why hydrogen helps with second stages, the overall mass matters there more than in the first stage.

[wannamoonbase]

meiza - 16/11/2007  11:00 AM

Yeah but high mass fraction doesn't help that much since the second stage even with infinite tankage ratio would lose to a lox/lh2 stage in mass (payload needs a certain amount of propellants to accelerate to a certain delta vee even if tankage and engine mass is zero). And thus a bigger first stage is needed.

That's why hydrogen helps with second stages, the overall mass matters there more than in the first stage.

Yep, hard to beat the weight and ISP advantages of LH2.  That's why the best vehicles, Atlas V, Delta 4 and Ariane 5 use cryogenic upper stages.

RP1 maybe more affordable and easier to develop but it is not a superior upper stage fuel.

[kkattula]

I've seen mention of SpaceX plans to build a cryogenic upper stage later on. I guess using the same propellants for bother stages, (and same engine for F9),  makes the initial development process a lot easier. IIRC the Falcon tanks are quite light too, and the engines have high t/w. So the Falcon first stage makes up a bit for the less efficient RP-1 second stage.

Still Falcon 9 compared to an Atlas V 402, LEO 28.5 payload is only 9,900 kg vs 12,500 kg, 79%
But launch mass is 325 mt compared to 546 mt, 59%.

So Falcon 9 seems to be more efficient at payload / gross mass? (3% vs 2.2%)  I didn't expect that.

I believe it's because the RL-10 engine has quite low thrust. Less than 3% of the RD-180 first stage. Whereas the F9 upper stage Merlin is about 11% of the first stage thrust. This makes the F9 a bit better balanced in stage budgets. Atlas V and Delta IV suffer from bad second stage gravity losses.

I guess it depends on your definition of superior fuel.  

[CFE]

It's debatable whether the next step for SpaceX should be an LH2 upper stage, or a "BFR."  I think BFR should be sized to replace the nine-Merlin cluster, rather than trying to compete with the F-1.  There's also the possibility that SpaceX and OSC could work together on an LH2 upper stage, which could see use on both Falcon 9 and Taurus-II/Cygnus.

Of course, it's important for SpaceX to actually get Falcon I working properly before even contemplating its next steps.  I suppose the next Falcon I flight will continue to use the standard Merlin, but I can imagine Merlin IC flying on the fourth Falcon I flight.

[JIS]

CFE - 17/11/2007  7:59 AM

It's debatable whether the next step for SpaceX should be an LH2 upper stage, or a "BFR."  I think BFR should be sized to replace the nine-Merlin cluster, rather than trying to compete with the F-1.  

What would be commercial benefit in replacing 9 Merlins with one new engine? They should size their engine to get around 20t to LEO or 10t to GEO in two stages one engine each.

[wannamoonbase]

CFE - 17/11/2007  2:59 AM

It's debatable whether the next step for SpaceX should be an LH2 upper stage, or a "BFR."  I think BFR should be sized to replace the nine-Merlin cluster, rather than trying to compete with the F-1.  There's also the possibility that SpaceX and OSC could work together on an LH2 upper stage, which could see use on both Falcon 9 and Taurus-II/Cygnus.

If they are going to be a viable commercial entity then they should see what the market supports.

I possible that the F9 with a high enough flight rate could do all of that and they could just make minor improvements for years to come and lower the individual costs of each launch.  Being in a continous design cycle will become extremely expensive to operations.

However, like you said they need to get the F1 flying and commercially available first, then the F9.  Then worry about BFRs and cryogenic engines and stages.

[Analyst]

Someone less lazy than I am should calculate the Falcon 9 payload capability using the ISPs, stage masses (empty and fueled) and the rocket equation. Maybe some performance funnies disappear sooner than later.

Analyst

[CFE]

JIS - 17/11/2007  2:02 AM

CFE - 17/11/2007  7:59 AM

It's debatable whether the next step for SpaceX should be an LH2 upper stage, or a "BFR."  I think BFR should be sized to replace the nine-Merlin cluster, rather than trying to compete with the F-1.  

What would be commercial benefit in replacing 9 Merlins with one new engine? They should size their engine to get around 20t to LEO or 10t to GEO in two stages one engine each.

The commercial benefit is that a 28-engine monster like Falcon 9 Heavy could become a manageable, four-engine rocket if you dropped a BFR in place of each nine-Merlin cluster.  Perhaps Falcon 9 Heavy can be made to work reliably with 27 engines on the first stage, but that remains to be seen.  Using three BFR's in parallel might solve any reliability problems.

I would also imagine that BFR could be developed fairly inexpensively by scaling the Merlin up.  However, F-1 still went through a lengthy development in spite of Rocketdyne's history with engines like S-3D and H-1.

[josh_simonson]

A big part of SpaceX's plans are to use economies of scale with the Merlin engine, chugging them out at relatively low cost and high reliability.  Sort of the Otrag approach but only with the expensive parts, not the tanks.  Their engine out plans go out the window if they use BFR on the F9, particularly the F9H.  As per their page they believe that the two most reliable designs are single engine/stage or multi-engine with engine out.  The F9H with single engine cores would have three parallel engines that must all work for a successful mission.  

Following their philosophy, I'd expect their first BFR to be a BF5 or BF9 - perhaps with redundant Merlins on the second stage.  Mainly because of the emphasis they've placed on engine-out, and because it would be in a class above F9H rather than competing with it.

It may make more sense in the near term to develop a cryogenic stage, but I think their rumer-mongering of the BFR is to kickstart NASA into contemplating the existence of such a thing in the medium term (~2020).  A cryogenic stage makes particular sense if they're making fuel deliveries for lunar/mars missions because the fuel to be delivered will almost certainly be hydrogen.  If they're going to develop tankage and pad infrastructure to deal with LH2 on payloads, they might as well use it in an upper stage too.

[aero313]

josh_simonson - 19/11/2007  2:02 PM

A big part of SpaceX's plans are to use economies of scale with the Merlin engine, chugging them out at relatively low cost and high reliability.

I still maintain that this "production line" approach is a fallacy.  Automakers don't string together a bunch of single cylinder engines for each car, they build the one engine that is required.  This "mass production" concept doesn't account for the significantly increased amount of touch labor required to integrate the nest of engines.  Every single one of the multitude of fluid connections on each engine (the dreaded "B-nuts") needs to be mated, leak checked, and validated by a second set of eyes (per SpaceX's own flight assurance procedure).  Every one of those connections is a potential source of a mission failure.  The more you have, the lower your reliability - Futron "study" notwithstanding.  Add to this the increased complexity of the thrust structure required to mount all these engines, the added number of actuators, the added number of valves, the added number of control electronics boxes, ...well, you get the idea.  I also maintain that if this vehicle REALLY has engine out capability under all flight regimes (including the first five seconds of flight) then it is not the lowest possible cost.

The lowest cost and most reliable launch vehicle will consistently be the one with the fewest parts.

aero313 - 19/11/2007  1:46 PM  

josh_simonson - 19/11/2007  2:02 PM  A big part of SpaceX's plans are to use economies of scale with the Merlin engine, chugging them out at relatively low cost and high reliability.

 I still maintain that this "production line" approach is a fallacy.  ... The lowest cost and most reliable launch vehicle will consistently be the one with the fewest parts.

Soyuz R-7 family used this strategy. It is the most successful, most launched vehicle of all time.

Still, the evolutionary path for Soyuz is to reduce parts. So both are true.

Perhaps the best way to look at it is through the eyes of a new launch vehicle source. You know you're sources of error and economics are tied to minimizing the number of new things. So it's far more risky the number of different engines and systems than multiple systems. And as a business, you sell what you have.

After you've proven the systems/engines, then the argument is over proven systems verses unproven systems. But at some point, the economics of loss/risk begin to favor enhanced redesign/replacement, so then (like Soyuz U/2/3 or Atlas II/V) you factor in incrementally changes.

Space-X is doing the correct strategy here. Don't get dogmatic - they will doubtlessly reduce engine number with bigger/fewer engines when appropriate. For now, its make a reliable Falcon 1 with Merlin 1c's , then increase lift capacity with a Falcon 9 using Merlin 1c's, then make a heavy out of 3 CBC's ala Delta IV. Pull that off, and you have still a potentially excellent LV business irrespective of number of engines.

Or, fiddle fart with vehicles, optimizing engines for each, and have little history for each of your engines/systems, so you don't know how much exposure you have with each, and you die the death of 1,000 cuts, and everyone laughs at you. Great. They are hard enough on Space-X as it is with F1 and F9. Rather do it with multiple Merlin 1c's.

[William Barton]

It sometimes seems to me that princple "The lowest cost and most reliable launch vehicle will consistently be the one with the fewest parts" can only be stretched so far. As for the automotive analogy, while it's true manufacturers don't build cars with a large number of small one-cyclinder engines, they also don't build cars with large one-cylinder engines with "the fewest parts." Automotive engines are lineally descended from late 19th century reciprocating engine technology and are designed the way they are for historical reasons more than anything else (remember the Wankel-engine Mazda?). By the same token, you don't see too many single-engine commercial airliners out there, either.

[Jim]

But on the converse, there aren't aircraft with more 4 engines.  27 engines is not good

[William Barton]

Jim - 19/11/2007  4:16 PM

But on the converse, there aren't aircraft with more 4 engines.  27 engines is not good

So it seems as if we are approaching a principle that there's is a point of diminishing return in each direction. B-52 was designed with 8 engines 60 years ago, and my guess is because that's how many of the size engines they could make back then that it took for that size aircraft. I hazily remember there was a proposal decades back to re-engine the B-52 with 4 big turbofans. So what's the optimum number of engines to get realistic engine out capability on an LV (assuming engine out a T-0 means not launching)? Five?

[sitharus]

It's a matter of diminishing returns. Using many units of a small cheap engine lowers manufacturing costs, but it will generally raise integration and running costs. Developing a large efficient engine takes a lot of capital, but running it will be cheaper, as there are fewer duplicate components adding to the dead weight.

However, duplication can prove useful, so long as you make sure you're not causing excessive vibration...

[edkyle99]

William Barton - 19/11/2007  3:32 PM

Jim - 19/11/2007  4:16 PM

But on the converse, there aren't aircraft with more 4 engines.  27 engines is not good

So it seems as if we are approaching a principle that there's is a point of diminishing return in each direction. B-52 was designed with 8 engines 60 years ago, and my guess is because that's how many of the size engines they could make back then that it took for that size aircraft. I hazily remember there was a proposal decades back to re-engine the B-52 with 4 big turbofans. So what's the optimum number of engines to get realistic engine out capability on an LV (assuming engine out a T-0 means not launching)? Five?

When was the last time an engine on a space launch vehicle went "out" in a recoverable way anyway?  

Sea Launch Zenit early this year suffered an engine failure at T-0 essentially - the kind of engine-out failure that I doubt any launch vehicle could recover from.  The same could probably be said for last year's GSLV Vikas booster engine failure right off the pad and the Soyuz Foton failure a few years ago.  Before that there were two Proton second stage engine failures in 1999 due to "foreign particles".  Falcon 9 wouldn't have survived those failures since it only has one second stage engine.  

Before the Proton failures I think you have to go back to the AC-74 Atlas booster engine failure back in 1993, but that was about 75-ish Atlases ago.  That rocket suffered a failure that caused reduced thrust beginning about 25 seconds after liftoff.  Perhaps that would have been a recoverable failure for a Falcon 9, but that was one instance in 14 years.  There have probably been roughly 1,000 space launches worldwide, give or take a few dozen, since then.

 - Ed Kyle