Price (or cost) of rocket engines

[99miles]

Ya'll:
I'm trying to figure out what is a fair and reasonable price for first stage rocket engines such as the BE-3, BE-4, Merlin 1D, RD-180, Aestus, F-1, LE-7A, NK-33A, RD-107A. RD-171M, RD-191, RS-68A, Vulcaine 2, and YF-21C. Does anyone have cost or price data that might help me? 

I'm also interested in price or cost data for upper stage rocket engines in the same class as the RL-10.

Another thing I'd like to know is, what are the biggest cost-drivers for building a rocket engine?  Is it fabrication of the fuel tanks? The plumbing?  The nozzle? Assembly? Testing? Quality control processes? From biggest to smallest, what is the correct order of the foregoing, and what did I leave out that is significant to include?

How do the non-recurring costs of design and certification get baked into the commercial price? 

Is making rocket engines a profitable business to be in? 

What is the added cost of making an engine reusable?

[envy887]

Off the top of my head (sources are public with jsut a little searching)

Merlin 1D: some fraction of $1M, SpaceX internal cost. See: https://forum.nasaspaceflight.com/index.php?topic=42923.0

RD-180: about $23 or $24M (used to be $10 million ~2005ish)

RS-25: around $50M

RL-10: around $25M

BE-4: $16M per pair or $8M each (that's the estimated ULA purchase price, Blue can build them internally for less)

RS-68: $10 to $20M each (old info ca. 2006) https://forum.nasaspaceflight.com/index.php?topic=2623.0

[Jim]

Another thing I'd like to know is, what are the biggest cost-drivers for building a rocket engine?  Is it fabrication of the fuel tanks? The plumbing?  The nozzle? Assembly? Testing? Quality control processes? From biggest to smallest, what is the correct order of the foregoing, and what did I leave out that is significant to include?

The turbo pump.  Tanks are part of the stage and not the engines.

[Coastal Ron]

Ya'll:
I'm trying to figure out what is a fair and reasonable price for first stage rocket engines...

Engines are a means to an end, but not the end themselves. In other words, the same question could be asked of rivets or the electronics. They are all necessary components whose worth is defined by the whole - the price of the service they provide.

...such as the BE-3, BE-4, Merlin 1D, RD-180, Aestus, F-1, LE-7A, NK-33A, RD-107A. RD-171M, RD-191, RS-68A, Vulcaine 2, and YF-21C.

Lots of apples and oranges. Different fuels, different sizes, expendable, reusable, etc. I'm not sure how costs for all of these will tell you anything useful.

Another thing I'd like to know is, what are the biggest cost-drivers for building a rocket engine?

Too general of a question in my opinion, because if it were the turbo-pump assembly that was the highest priced item (as an example), and it was 15% of the total cost of an engine, then 85% of the cost of the engine still needs to be considered.

How do the non-recurring costs of design and certification get baked into the commercial price?

There are few purely "commercial" rocket engines being used for orbital rockets today, so you're not going to be able to get a definitive answer. Plus, rocket engine costs are not usually public, so trying to find out that information could turn out to be fruitless.

Is making rocket engines a profitable business to be in?

SpaceX builds their own engines as a means to an end, but not an end itself. In other words, they need low-cost rocket engines that meet their specific needs, and it turns out it's better if they build them themselves - but they don't sell them to anyone else... yet.

Even for Blue Origin, who is willing to sell their BE-4 engine to ULA, they started development of it for their own needs, not specifically for ULA or other potential customers.

What is the added cost of making an engine reusable?

Based on outside observation of SpaceX and Blue Origin, and not being a rocket engineer, my guess would be that it's not too much more expensive to build a reusable rocket engine than an expendable one. YMMV.

What is your goal with this information?

[Lars-J]

What is the added cost of making an engine reusable?

- You need more margins, so in theory it will be slightly heavier than an expendable version.
- You need a simpler start & restart capability, which will cost more to develop but should have the benefit of simpler ground infrastructure. (short term loss, but long term cost gain even if engines are expended)

The last point is not a requirement (the SSME certainly was/is reusable and tricky to start), but if your engine is complex to start the operational costs will be high.

[HMXHMX]

What is the added cost of making an engine reusable?

- You need more margins, so in theory it will be slightly heavier than an expendable version.
- You need a simpler start & restart capability, which will cost more to develop but should have the benefit of simpler ground infrastructure. (short term loss, but long term cost gain even if engines are expended)

The last point is not a requirement (the SSME certainly was/is reusable and tricky to start), but if your engine is complex to start the operational costs will be high.

A couple of points of reference:  the RL-10 has performed several hundred starts on a "stock" engine, with primary limits being bearing life and thermal ratcheting on the tubular combustion chamber, and the original spec for engines such as the LR-79 (Thor) was for a minimum of 12 starts, so one could argue that even it was reusable (in the 1950s).

But agree that start (and I'd argue also shutdown) are key factors in the reuse equation.  For TPAs, the rate of onset of thermal shock in the turbine and bearing life will be limiting.

[Lars-J]

A couple of points of reference:  the RL-10 has performed several hundred starts on a "stock" engine, with primary limits being bearing life and thermal ratcheting on the tubular combustion chamber, and the original spec for engines such as the LR-79 (Thor) was for a minimum of 12 starts, so one could argue that even it was reusable (in the 1950s).

Certainly, and as used in the DC-X, the RL-10 certainly proved that it could work in a reusable vehicle.

[Ragmar]

Off the top of my head (sources are public with jsut a little searching)

Merlin 1D: some fraction of $1M, SpaceX internal cost. See: https://forum.nasaspaceflight.com/index.php?topic=42923.0

RD-180: about $23 or $24M (used to be $10 million ~2005ish)

RS-25: around $50M

RL-10: around $25M

BE-4: $16M per pair or $8M each (that's the estimated ULA purchase price, Blue can build them internally for less)

RS-68: $10 to $20M each (old info ca. 2006) https://forum.nasaspaceflight.com/index.php?topic=2623.0

I'm familiar with that BE-4 numbers.  Has there been any speculation on how much AR1 is?  Just flyaway cost.

[envy887]

Off the top of my head (sources are public with jsut a little searching)

Merlin 1D: some fraction of $1M, SpaceX internal cost. See: https://forum.nasaspaceflight.com/index.php?topic=42923.0

RD-180: about $23 or $24M (used to be $10 million ~2005ish)

RS-25: around $50M

RL-10: around $25M

BE-4: $16M per pair or $8M each (that's the estimated ULA purchase price, Blue can build them internally for less)

RS-68: $10 to $20M each (old info ca. 2006) https://forum.nasaspaceflight.com/index.php?topic=2623.0

I'm familiar with that BE-4 numbers.  Has there been any speculation on how much AR1 is?  Just flyaway cost.

I recall that ULA was saying about $25M (per pair, which is what they need to replace RD-180 on Atlas V).

[99miles]

My employer would like me to arrive at some sort of recommendation on whether the prices being asked for  several specific rocket engines are fair and reasonable.  So I would like to plot some data points showing "bang" for the buck for different engines, assuming that they are operated as intended with appropriate fuel systems and casings, etc., but without help from any other rockets. 

I'm comparing like to like, i.e. 1st stage engines to other 1st stage engines.

I am computing "bang" based on the difference between thrust and weight integrated over time from launch/startup to empty, but I'm not sure this is the best choice of parameters to benchmark.  Any suggestions on other meaningful parameters I ought to be looking at?

[Ragmar]

Off the top of my head (sources are public with jsut a little searching)

Merlin 1D: some fraction of $1M, SpaceX internal cost. See: https://forum.nasaspaceflight.com/index.php?topic=42923.0

RD-180: about $23 or $24M (used to be $10 million ~2005ish)

RS-25: around $50M

RL-10: around $25M

BE-4: $16M per pair or $8M each (that's the estimated ULA purchase price, Blue can build them internally for less)

RS-68: $10 to $20M each (old info ca. 2006) https://forum.nasaspaceflight.com/index.php?topic=2623.0

I assume we haven't seen anything on BE-3U or Vinci upperstage prices as well?  Would be interesting to compare to RL-10.

[A_M_Swallow]

If your product is more than 2 years away ask Masten Space how much they charge.

[Craftyatom]

My employer would like me to arrive at some sort of recommendation on whether the prices being asked for  several specific rocket engines are fair and reasonable.  So I would like to plot some data points showing "bang" for the buck for different engines, assuming that they are operated as intended with appropriate fuel systems and casings, etc., but without help from any other rockets. 

I'm comparing like to like, i.e. 1st stage engines to other 1st stage engines.

I am computing "bang" based on the difference between thrust and weight integrated over time from launch/startup to empty, but I'm not sure this is the best choice of parameters to benchmark.  Any suggestions on other meaningful parameters I ought to be looking at?

Let me start by mentioning that the below assumes that you are looking at liquid-fueled engines - hybrid engines or solid-fueled engines are completely different beasts, with different metrics to consider.  Most importantly, in the case of solids (and to some extent hybrids), the engine and its propellant are the same thing, and thus are sold as a unit - all of the prices discussed in this thread are for liquid engines, and do not include the prices of fuel tanks, their plumbing, or the fuel they carry.

With that out of the way, I would propose that there are four very important metrics to consider when looking at the price of a liquid-fueled engine:
>Engine Cycle: the engine cycle pretty much defines how "complex" the engine is - simple pressure-fed engines are (theoretically) much cheaper, while complex cycles like various staged combustion types can cost a heck of a lot more.
>Propellant Type: in the world of bipropellant engines, there are a number of different fuel types - they don't differ too much from one another, but it's worth noting that non-cryogenic fuels are usually considered simpler, and hypergolic fuels are quite a bit simpler, so engines that use these fuels are cheaper.  Also worth noting is that monopropellant engines exist, and they're usually simpler/cheaper too.
>Thrust: this is one of the most important first-stage metrics.  Pretty simple, but very important.
>Specific Impulse: this is the main reason anyone uses engines that are more complex.  A higher specific impulse means that a rocket is more efficient, getting more energy out of its fuel.  "Complex" propellants like Hydrogen/Oxygen offer higher specific impulses, as do complex engine cycles like FFSC.  Of course, if the fuel for your rocket engine doesn't make up a significant portion of your vehicle's mass - for example, in a rocket car - then efficiency doesn't matter at all, so specific impulse becomes rather unimportant.

I'd advise against the metric you were using (which is basically delta-V) because it depends not only on the engines, but also the tanks, plumbing, fuel, and extra mass.  When people cite prices for, say, the Merlin 1D, they don't include the cost of the rest of the Falcon 9's first stage, because the tanks and engines are considered separate.  Multiplying the thrust and specific impulse can give you a theoretical max total impulse for the engine itself, not including tankage.

[Rik ISS-fan]

Multiplying the thrust and specific impulse can give you a theoretical max total impulse for the engine itself, not including tankage.

Craftyatom; vary nice informative reply.
But if I'm not mistaken the quoted statement is wrong. Please correct me if I'm wrong!

To my (little) knowledge about propulsion and equations, ISP is used in two equetions.
1) The ISP is derived from (vertical) ground engine firing tests. F = Ve × Mflow = ISP × g0 × Mflow wiki
During a ground firing test the propallent mass flow (Mflow [kg/s]) and generated thrust (F [N | kg.m/s^2]) are measured.
The effective exhaust velocity (Ve [m/s]) is the relationship between the two. The thrust is dependent of gravity (g0). To make the equation vallid in all cases, the notion of specific impuls was introduced (ISP × g0 = Ve)  [s × m/s^2 = m/s]
{G}imperial works with weight, so thrust (F [lbf]) = weight flow (Mflow × g0) [lbs/s] × ISP [s ]. This is always gravity dependant and thus only valid at sea level. {My preference for SI should be clear now.}

2) And ISP is used in the Tsiolkovski rocket equation. dV [m/s] =  Ve [m/s] × ln (M0/Me) [- | kg/kg] =
ISP × g0 x ln (M0/Me)

The total impuls is total fuel mass multiplied by Ve or ISP x g0.
The fuel mass is not given with the engine specs, thus is total impuls a wrong parameter for liquid engine comparison.
I think it will be very hard to find reliable rocket engine cost prices.

For European engines, it was stated that the promethee engine should cost about 1mln, about 1/10th what Vulcain 2 costs. (But the cost of vulcain 2.1; 2.2 and 2.3 ??)

[ChrisWilson68]

My employer would like me to arrive at some sort of recommendation on whether the prices being asked for  several specific rocket engines are fair and reasonable.

It's a myth that there's some sort of objective standard for "fair and reasonable".  It's purely subjective.  You might as well ask if a given rocket engine is "good" or "bad".

[spacenut]

For booster engines, what if you figured it cost/lb or kg of thrust.  Upper stage engines, not only per cost per unit, but also factor in the cost/unit delivered to target.  I know kerolox's ISP isn't as great as hydrolox but for payload mass delivered by cost. 

[ChrisWilson68]

If your product is more than 2 years away ask Masten Space how much they charge.

The original poster's employer clearly has no product.  The report is for some political purpose or legal argument.

The term "fair and reasonable" is a term used by lawyers, not by engineers or business people.  If it were a business, they'd be looking to see exactly how particular engines suit their particular needs and if more than one does, how the costs asked for those engines compare.  And they'd be asking those companies for quotes, they wouldn't be asking for information from public sources on engines in general.

[Robotbeat]

For first stage, thrust times Isp versus dollars isn't too bad, though maybe doesn't QUITE value Isp enough (Isp squared is too much for a first stage, about right for a second stage). Probably want T/W as a figure of merit, too. But not sure how to include that... Maybe add $1000 in cost per kilogram of dry mass??

[spacenut]

Ok, so which booster engine is best as a cost/kg delivered to space for booster engines.

And, which 2nd or 3rd stage engine is best as cost/kg delivered to LEO, GTO, moon, Mars? 

Has anyone figured that out yet? 

I would venture to say the Merlin beats out on booster engines.  And maybe if prices were available the BE-3 would win out on deep space.  Then the Merlin vac would probably win out on LEO.  Just guessing. 

All other engines, except maybe the RD-180?  are expensive.  However 5-7 Merlins at $5-$7 million may win out over one RD-180 at $25 million.  Yes, more fuel, but fuel is cheap compared to the engine prices. 

[99miles]

My employer would like me to arrive at some sort of recommendation on whether the prices being asked for  several specific rocket engines are fair and reasonable.

It's a myth that there's some sort of objective standard for "fair and reasonable".  It's purely subjective.  You might as well ask if a given rocket engine is "good" or "bad".

If it was your own money you were using to buy one, you might find fair and reasonable pricing not quite so mythical!

[Kansan52]

I am not an engineer. This sight has such a wealth of knowledge and you have seen a small amount in this thread.

One lesson from this sight is the answer is not the engine, not the rocket, and not even the company. It is the service.

First is the mission you need to accomplish. Then who can do it. Generally plan on signing the launch service contract years in advance.

The list of engines earlier and the estimates given should be the start of a nice chart if engines are the only parameter.

[Nilof]

There is no such thing as a fair and reasonable price for an engine, really. Market economics for rocket engines don't really exist, most engines are custom-designed for a particular customer that goes on to be the only customer for that engine. Furthermore, politics play a major part in who can be allowed to make what for whom. The sample size is just too small.

In addition, the sector is in the process of being disrupted by a number of changes such as improvements in manufacturing (such as 3D printing), and in reuse becoming more of a concern for new launch vehicles due to SpaceX and Blue origin which changes requirements.

[Lar]

Merlin good, all others (currently available) bad.  Done.

OK  no, that's not right. Horses for courses. But one could argue that the total package the F9 booster offers is pretty remarkable on a price/performance basis.

I'm not sure what the OP is after either. He or she turned up with a very open ended question and not a lot of motivation other than "for my employer" ... that's what consultants are hired for, but he or she is getting our thoughts for free. Wonder what we get in return other than intersting discussion (which often is "enough" to make it worthwhile)....

[QuantumG]

There is no such thing as a fair and reasonable price for (anything really)

Honestly, people seem to refer to a "fair price" as if it's obvious to everyone else. In a sense, that's what they mean, the price that you'd expect to pay. It's basically sentimentalism. That price feels low, that price feels high, but that price feels just right. If you need a rocket engine you pay what you have to pay for the application you're buying it for. If you can't afford it you try to build your own, and 99% of the time you go out of business.

[Robotbeat]

All I know is the Merlin 1D is a killer engine... If you were pretty good at making lightweight structures (comparable to F9), you could make an expendable SSTO smallsat launcher (500kg?) out of a Merlin 1D first stage. Cost of the engine would be so cheap (if SpaceX charges you, say, half a million) that you could beat the snot out of more other smallsat launchers plus with the added safety of SSTO.

[envy887]

All I know is the Merlin 1D is a killer engine... If you were pretty good at making lightweight structures (comparable to F9), you could make an expendable SSTO smallsat launcher (500kg?) out of a Merlin 1D first stage. Cost of the engine would be so cheap (if SpaceX charges you, say, half a million) that you could beat the snot out of more other smallsat launchers plus with the added safety of SSTO.

I don't think it quite makes it. A Merlin 1D can only lift about 75 tonnes at SL and can only put about 3 tonnes through 9400 m/s. That requires a dry mass fraction under 4% which is much more like ITS than Falcon.

[S.Paulissen]

All I know is the Merlin 1D is a killer engine... If you were pretty good at making lightweight structures (comparable to F9), you could make an expendable SSTO smallsat launcher (500kg?) out of a Merlin 1D first stage. Cost of the engine would be so cheap (if SpaceX charges you, say, half a million) that you could beat the snot out of more other smallsat launchers plus with the added safety of SSTO.

I don't think it quite makes it. A Merlin 1D can only lift about 75 tonnes at SL and can only put about 3 tonnes through 9400 m/s. That requires a dry mass fraction under 4% which is much more like ITS than Falcon.

Elon has already confirmed that the FH side boosters were SSTOs with little to no payload.

[envy887]

All I know is the Merlin 1D is a killer engine... If you were pretty good at making lightweight structures (comparable to F9), you could make an expendable SSTO smallsat launcher (500kg?) out of a Merlin 1D first stage. Cost of the engine would be so cheap (if SpaceX charges you, say, half a million) that you could beat the snot out of more other smallsat launchers plus with the added safety of SSTO.

I don't think it quite makes it. A Merlin 1D can only lift about 75 tonnes at SL and can only put about 3 tonnes through 9400 m/s. That requires a dry mass fraction under 4% which is much more like ITS than Falcon.

Elon has already confirmed that the FH side boosters were SSTOs with little to no payload.

Maybe for the empty booster to a low altitude, low inclination orbit. But a typical reference orbit for small-sat launchers is 500 km sun-synchronous, which is ~600 m/s more demanding. That quickly takes a SSTO payload from zero to negative.

A Falcon-1 SSTO putting up 500 kg would be like a Falcon 9 SSTO putting up 4500 kg. It could be done with 4% dry mass, and to sun-sync with 3% dry mass. But not with Falcon's 4.5 to 5.5% dry mass.

[Robotbeat]

Falcon booster was supposed to get 25:1 mass fraction when optimized, i.e. 4%.

[99miles]

RS-25: around $50M

RS-68: $10 to $20M each (old info ca. 2006) https://forum.nasaspaceflight.com/index.php?topic=2623.0

Suppose I want to try to develop a "fair and reasonable" price for the RS-68 by comparing it to the RS-25 and doing a parametric price analysis based on thrust, or possibly thrust x specific impulse.

Apart from the RS-25's man-rating, what other properties of these two engines would make that analysis misleading?  How could I correct for them?

[Kabloona]

RS-25: around $50M

RS-68: $10 to $20M each (old info ca. 2006) https://forum.nasaspaceflight.com/index.php?topic=2623.0

Suppose I want to try to develop a "fair and reasonable" price for the RS-68 by comparing it to the RS-25 and doing a parametric price analysis based on thrust, or possibly thrust x specific impulse.

Apart from the RS-25's man-rating, what other properties of these two engines would make that analysis misleading?  How could I correct for them?

I'm dubious that you can get any meaningful result from such a parametric analysis. The RS-68 is significantly different from the RS-25 in that it has substantially fewer parts (Wikipedia says something like 80% fewer) and has a fundamentally different nozzle design (channel wall cooling with an ablative nozzle extension, vs brazed tube cooling in the RS-25).

By reducing the parts count and getting away from the labor-intensive process of brazing the nozzle tubes, the RS-68 can be made significantly cheaper than the RS-25. These are major factors in the cost difference that have nothing to do with parametric thrust or ISP comparisons.

Maybe I'm missing the point of what you're trying to do, but the actual reason an RS-25 costs so much more than an RS-68 has to do with the design and manufacturing issues, and just doing a parametric comparison of thrust or ISP ignores those factors.

The way NASA would to do a true "should cost" analysis is to estimate the cost of manhours and materials that go into a typical component (say, a channel wall nozzle vs a brazed tube nozzle) based on their extensive experience with engine design, construction, and procurement. And they would estimate that an RS-25 nozzle "should cost" much more than an RS-68 nozzle with its less labor intensive channel wall cooling and ablative extension. And so on, for each component or subsystem of the engine.

They would then end up with a bottom line "should cost" number for each engine based on its unique design and specifications, and the RS-68 number would likely be significantly less than the RS-25 number. Both numbers would probably be around what envy887 posted above.

So your comparison problem is that you can't "correct" for the differences in design unless you know (to pick just one example) what it "should cost" to build a channel wall nozzle with an ablative extension, as compared to a brazed tube nozzle,and that's a level of knowledge most people on this site, including me, don't have.

[spacenut]

Does anyone know the cost of the BE3U? 

[Davidthefat]

What it costs to make something doesn't mean it's what they sell it for and the R&D costs rolled up into it.

[john smith 19]

Suppose I want to try to develop a "fair and reasonable" price for the RS-68 by comparing it to the RS-25 and doing a parametric price analysis based on thrust, or possibly thrust x specific impulse.

Apart from the RS-25's man-rating, what other properties of these two engines would make that analysis misleading?  How could I correct for them?

Wow, you sound like some new hire whose been set a nearly impossible challenge by their boss.

There are 2 ways to do this. Top down and bottom up. What you're doing is top down. You seem to be trying to draw a line with 2 points only.

You could start by finding a standard Work Breakdown Structure for a rocket engine (AIAA should have something, or NASA STI or DTIC for known cycles) then getting cost estimating data from production engineering handbooks and working through the numbers. Parts counts is a factor, as is touch labor (a USAF general commented an RL10 took more touch labor than a Ferrari). Cost models developed for photocopiers (which deal with part count rather than part mass, which is what most aerospace CER's basically deal with) seemed to be more accurate according to Len Cornier.

What do you want?
What constraints do you have on on propellants?
How many do you need? 1 big engine on the basis that fewer parts --> less to go wrong, or multi engines (like Saturn V first stage) so vehicle does not suffer loss of mission if one engine fails in flight?
What's the chamber pressure?

IIRC there was a cost estimating equation worked out on one of the old sci.space. newsgroups which had been fitted to most of the large engines up to that time. Chamber pressure was a big differentiator (as it tends to be for combustion instability).

Otherwise the question is so open as to be meaningless.

A "fair price" as a mfg is one that lets me sell enough engines often enough to make a decent profit. A fair price to you (as a customer) is one that lets you build your vehicle within its target cost.

[ZChris13]

-snip-

-snip-

John... that post was over two years ago.

[TinkerLaspeyrs]

99Miles
This is a human economics question about uncertain "known" data biased by prior knowledge.You may as well be attempting to price seasonal fashions where the cost of production may not change much but the market accepts the current price out of vanity. Price is influenced by the prime human motives of greed and fear both for suppliers and buyers. This is reduced in a more vertically integrated operation. Where external suppliers are used they have their own amount of greed. So minimum price and maximum profit favors the vertically integrated. Maybe you could try the Delphi method utilizing some of the expansive knowledge of this forums members. Maybe sort of asking for trouble but you could limit responses as the Delphi method suggests, under pain of exclusion for over enthusiastic contributors. Treat them as outliers as the enthusiasm could be a sign of bias.
A good economist is always vague enough to always be right afterwards. Because it all depends.
You really have to define the utility to the customer. If the customer is the rocket manufacturer then the price Kg/orbit is not the same as the price Kg/orbit to the same orbit for a satellite operator if the satellites mass is half the rockets capability.
Y'all have fun now.