Author Topic: Could a new F-1x Engine open the door for an 8.4m kerolox core for SLS?  (Read 43992 times)

Offline Lobo

  • Senior Member
  • *****
  • Posts: 6758
  • Spokane, WA
  • Liked: 556
  • Likes Given: 348
Anyway, I imagine this RP-1 LV with 4 X F-1x would have something similar, if htere was an early abort, the LAS system would save the crew, as it would have for Apollo.

Would probably save the crew, as it probably would have saved an Apollo crew.

Good point!  Nothing's perfect.  But I figure if the Ruskie's LAS sytem would have saved the crews during the N-1 failures, hopefuly 40 years later, ours would too. 

Offline Proponent

  • Senior Member
  • *****
  • Posts: 5596
  • Liked: 1142
  • Likes Given: 684
The question that keeps bugging me is, does a core-plus-boosters design really make sense for an expendable launch vehicle?  If the core goes all the way to orbit, then each kilogram by which it exceeds the design spec is a kilogram of lost payload capability.  The entire stage, which is tossed out after each flight, is therefore mass-critical.  The same goes for the engines, which may need to be expensive, high-pressure beasts like the RS-25, so they can operate at sea level yet still be efficient in vacuo.  (TAN might work well to alleviate this last problem, but that doesn't seem to be on the radar.)

With a two-stages-to-orbit approach, on the other hand, the entire first stage can tolerate relatively inefficient engines optimised for low-altitude operation.  It strikes me that this is likely to be cheaper.

As I said, the above question was bugging me.  So I had a look at some numbers -- see the attached spreadsheet.  Two vehicles capable of imparting an ideal delta-V of 9200 m/s (about enough to reach LEO, after losses are taken into account) to a payload of 120 tonnes, which was about the capability of the two-stage Saturn V (Skylab weighed less; it didn't use the Saturn's full capability).

The first vehicle is a classic two-stager, the first fueled by RP-1 and the second by hydrogen.  The second vehicle consists of a ground-lit hydrogen core flanked by RP-1-powered boosters.

You can get different results with different parameters, but the core-plus-boosters configuration always has a bigger upper stage, easily 50% bigger by dry mass.  It also as a somewhat more powerful lox-hydrogen engine, which must operate efficiently from the ground up.  It's got to be a more expensive vehicle.  I guess that's what RAC-2 told us.

Offline notsorandom

  • Full Member
  • ****
  • Posts: 1721
  • Ohio
  • Liked: 418
  • Likes Given: 91
The question that keeps bugging me is, does a core-plus-boosters design really make sense for an expendable launch vehicle?  If the core goes all the way to orbit, then each kilogram by which it exceeds the design spec is a kilogram of lost payload capability.  The entire stage, which is tossed out after each flight, is therefore mass-critical.  The same goes for the engines, which may need to be expensive, high-pressure beasts like the RS-25, so they can operate at sea level yet still be efficient in vacuo.  (TAN might work well to alleviate this last problem, but that doesn't seem to be on the radar.)

With a two-stages-to-orbit approach, on the other hand, the entire first stage can tolerate relatively inefficient engines optimised for low-altitude operation.  It strikes me that this is likely to be cheaper.

As I said, the above question was bugging me.  So I had a look at some numbers -- see the attached spreadsheet.  Two vehicles capable of imparting an ideal delta-V of 9200 m/s (about enough to reach LEO, after losses are taken into account) to a payload of 120 tonnes, which was about the capability of the two-stage Saturn V (Skylab weighed less; it didn't use the Saturn's full capability).

The first vehicle is a classic two-stager, the first fueled by RP-1 and the second by hydrogen.  The second vehicle consists of a ground-lit hydrogen core flanked by RP-1-powered boosters.

You can get different results with different parameters, but the core-plus-boosters configuration always has a bigger upper stage, easily 50% bigger by dry mass.  It also as a somewhat more powerful lox-hydrogen engine, which must operate efficiently from the ground up.  It's got to be a more expensive vehicle.  I guess that's what RAC-2 told us.
That is an interesting spreadsheet, thanks for providing it. I have been playing with it a bit. I think I see one advantage of the core and booster arrangement. Even though a parallel staged rocket may weigh more the total combined thrust of all the engine is less. From what I can tell that is because the core's engines are doing double duty as both first and second stage engines. To put it another way the engines of a second stage are dead weight until the first stage is expended.

I'm getting a difference of 3,310KN with the default values. Setting the ISP of the second and core to mimic engines likely to be used in real rockets a parallel staged LV still needs less thrust. An upper stage with the J-2x(448s) vs SSME(452s) core gives a difference of 3,100KN. I wondered, what about using the RL-10? An RL-10B-2 at 462s vs SSME at 452 has a difference of 1,663KN. However the second stage would need 33 RL-10s to provide enough thrust!

That difference in needed thrust could lead to a more expensive rocket. From what I can tell the difference needs to be made up by the first stage.

Offline Downix

  • Senior Member
  • *****
  • Posts: 7087
  • Liked: 16
  • Likes Given: 1
The question that keeps bugging me is, does a core-plus-boosters design really make sense for an expendable launch vehicle?  If the core goes all the way to orbit, then each kilogram by which it exceeds the design spec is a kilogram of lost payload capability.  The entire stage, which is tossed out after each flight, is therefore mass-critical.  The same goes for the engines, which may need to be expensive, high-pressure beasts like the RS-25, so they can operate at sea level yet still be efficient in vacuo.  (TAN might work well to alleviate this last problem, but that doesn't seem to be on the radar.)

With a two-stages-to-orbit approach, on the other hand, the entire first stage can tolerate relatively inefficient engines optimised for low-altitude operation.  It strikes me that this is likely to be cheaper.

As I said, the above question was bugging me.  So I had a look at some numbers -- see the attached spreadsheet.  Two vehicles capable of imparting an ideal delta-V of 9200 m/s (about enough to reach LEO, after losses are taken into account) to a payload of 120 tonnes, which was about the capability of the two-stage Saturn V (Skylab weighed less; it didn't use the Saturn's full capability).

The first vehicle is a classic two-stager, the first fueled by RP-1 and the second by hydrogen.  The second vehicle consists of a ground-lit hydrogen core flanked by RP-1-powered boosters.

You can get different results with different parameters, but the core-plus-boosters configuration always has a bigger upper stage, easily 50% bigger by dry mass.  It also as a somewhat more powerful lox-hydrogen engine, which must operate efficiently from the ground up.  It's got to be a more expensive vehicle.  I guess that's what RAC-2 told us.
This layout is putting the thumbs on the scale, comparing a two stage against a 2.5 stage design. This is shoe-horning a layout without any optimization for the advantage of each, with the result being this.

So I went in, and did some real world tweaking.  Increasing the second/core stage ISP to the SSME's, the difference in the fuel needed dropped, significantly.  Taking into account the SSME's throttling, and using that to adjust the fuel burned during initial ascent to 0.12, you suddenly have the second stage neck and neck between the two.  Then adjusting for the first stage specific impulse for sea level of 280 of the F-1A, now the story becomes radically different, with the whole stack for the parallel staged design having a significantly lower lift off mass, 200 metric tons less.
« Last Edit: 05/08/2012 09:21 PM by Downix »
chuck - Toilet paper has no real value? Try living with 5 other adults for 6 months in a can with no toilet paper. Man oh man. Toilet paper would be worth it's weight in gold!

Offline simonbp

My 2 cents on serial vs. parallel:

I set up a representative moderised Saturn V configuration:

First Stage: 4x F-1A, 1400 tonnes, pmf=0.93
Second Stage: 3x SSME, 750 tonnes, pmf=0.89
Third Stage: 4x RL-10B, 200 tonnes, pmf=0.89

With two stages, I get about 80 tonnes to LEO. With three stages, I get 50 tonnes to TLI. The gross mass of the three-stage version is 2400 tonnes.

I then set up a parallelised version of the same rocket, with two boosters and a core, with the core having 75% of its propellant at separation:

Boosters (2x): 2x F-1A, 850 tonnes, pmf=0.93
Core Stage: 5x SSME, 950 tonnes, pmf=0.89
Third Stage: 4x RL-10B, 400 tonnes, pmf=0.89

The core+boosters sends 56 tonnes to LEO, while with the third stage it sends 50 tonnes to TLI. The gross mass of the latter is 3840 tonnes.

So, it seems to me that the serial version is far superior, having a much lower GLOW and using fewer engines. Neither version is particularly optimised, though, so I could be wrong.

Offline Proponent

  • Senior Member
  • *****
  • Posts: 5596
  • Liked: 1142
  • Likes Given: 684
I've updated the spreadsheet (attached).  It now shows the difference in masses, thrusts and durations between the two configurations.  At the bottom, it also shows some trade ratios, about which more later.  (BtW, I tried to encode a VB macro which calls the solver.  Whenever I run it, however, VB complains that it doesn't know about the solver function.  Anybody know how to fix this?  I'm not much of a Excel guru.)

I think I see one advantage of the core and booster arrangement. Even though a parallel staged rocket may weigh more the total combined thrust of all the engine is less. From what I can tell that is because the core's engines are doing double duty as both first and second stage engines. To put it another way the engines of a second stage are dead weight until the first stage is expended.

I'm getting a difference of 3,310KN with the default values. Setting the ISP of the second and core to mimic engines likely to be used in real rockets a parallel staged LV still needs less thrust. An upper stage with the J-2x(448s) vs SSME(452s) core gives a difference of 3,100KN.

With those Isps, I get that the upper-stage mass increases by 11,000 kg against a lower-stage mass decrease of 10,000 kg.  Upper-stage thrust increases by 300 kN, while lower-stage thrust decreases by 4200 kN.  In other words, each kilogram of first-stage mass reduction costs 1.1 kg of increased second-stage mass (see line 36 of the new spreadsheet), and each newton of thrust saved on the first stage costs us 0.07 N on the second stage (line 37).

The mass trade is definitely bad, because the upper stage is mass-critical (its dry weight is lugged all the way to orbit).  A kilogram of upper-stage structure will be more expensive and entails more risk than a kilogram of lower-stage.

The thrust trade, on the other hand, might look good.  But what we're doing here is replacing relatively cheap thrust from, e.g., a J-2X and an F-1A, with a smaller amount of thrust from an expensive SSME (see the last line of the spreadsheet).  So if we do save money on engines, it certainly won't be in proportion to the change in total thrust.

An Isp of 452 s for the core, however, is not realistic, since the core's engines must operate from sea level to vacuum (whereas the second stage's engines need operate only in a vacuum).  All we can do with a simple model like this is guestimate an appropriate trajectory-averaged Isp.  If I take the average core Isp down to 430 s (the SSME's sea-level Isp is just 363 s), then each kilogram of reduction in booster mass costs 18 kg of core mass, and each kilonewton of reduced booster thrust costs 380 N in core thrust.  I'm skeptical that trade would be economic.

Quote
I wondered, what about using the RL-10? An RL-10B-2 at 462s vs SSME at 452 has a difference of 1,663KN. However the second stage would need 33 RL-10s to provide enough thrust!

Yeah, the RL-10 is just too small for use on an Earth-to-LEO stage of a Saturn V-class vehicle.

This layout is putting the thumbs on the scale, comparing a two stage against a 2.5 stage design.

It's two stages against two stages; where is the extra half stage?

Quote
So I went in, and did some real world tweaking.  Increasing the second/core stage ISP to the SSME's, the difference in the fuel needed dropped, significantly.

As noted above, using the full SSME Isp is not realistic.

Quote
Taking into account the SSME's throttling, and using that to adjust the fuel burned during initial ascent to 0.12, you suddenly have the second stage neck and neck between the two.

In a Shuttle launch, the SSMEs burned about four times longer than the SRBs.  If they were throttled back to 2/3 power for half boost phase, then the boost-burn fraction would be about 0.2.  SLS, with five-segment SRBs, would stage higher and faster than did the Shuttle, which would tend to increase the boost burn fraction.  Hence, I suspect 0.25 is a pretty reasonable number.

Offline Downix

  • Senior Member
  • *****
  • Posts: 7087
  • Liked: 16
  • Likes Given: 1
I've updated the spreadsheet (attached).  It now shows the difference in masses, thrusts and durations between the two configurations.  At the bottom, it also shows some trade ratios, about which more later.  (BtW, I tried to encode a VB macro which calls the solver.  Whenever I run it, however, VB complains that it doesn't know about the solver function.  Anybody know how to fix this?  I'm not much of a Excel guru.)

I think I see one advantage of the core and booster arrangement. Even though a parallel staged rocket may weigh more the total combined thrust of all the engine is less. From what I can tell that is because the core's engines are doing double duty as both first and second stage engines. To put it another way the engines of a second stage are dead weight until the first stage is expended.

I'm getting a difference of 3,310KN with the default values. Setting the ISP of the second and core to mimic engines likely to be used in real rockets a parallel staged LV still needs less thrust. An upper stage with the J-2x(448s) vs SSME(452s) core gives a difference of 3,100KN.

With those Isps, I get that the upper-stage mass increases by 11,000 kg against a lower-stage mass decrease of 10,000 kg.  Upper-stage thrust increases by 300 kN, while lower-stage thrust decreases by 4200 kN.  In other words, each kilogram of first-stage mass reduction costs 1.1 kg of increased second-stage mass (see line 36 of the new spreadsheet), and each newton of thrust saved on the first stage costs us 0.07 N on the second stage (line 37).

The mass trade is definitely bad, because the upper stage is mass-critical (its dry weight is lugged all the way to orbit).  A kilogram of upper-stage structure will be more expensive and entails more risk than a kilogram of lower-stage.

The thrust trade, on the other hand, might look good.  But what we're doing here is replacing relatively cheap thrust from, e.g., a J-2X and an F-1A, with a smaller amount of thrust from an expensive SSME (see the last line of the spreadsheet).  So if we do save money on engines, it certainly won't be in proportion to the change in total thrust.

An Isp of 452 s for the core, however, is not realistic, since the core's engines must operate from sea level to vacuum (whereas the second stage's engines need operate only in a vacuum).  All we can do with a simple model like this is guestimate an appropriate trajectory-averaged Isp.  If I take the average core Isp down to 430 s (the SSME's sea-level Isp is just 363 s), then each kilogram of reduction in booster mass costs 18 kg of core mass, and each kilonewton of reduced booster thrust costs 380 N in core thrust.  I'm skeptical that trade would be economic.
Compare time at SL vs Vac for engine burn.  It will be operating at Vac for far longer than at SL, and if the engine is throttled down, leaving the major lifting to the boosters, that SL vs Vac ratio is tilted even further to Vac.
Quote
Quote
I wondered, what about using the RL-10? An RL-10B-2 at 462s vs SSME at 452 has a difference of 1,663KN. However the second stage would need 33 RL-10s to provide enough thrust!

Yeah, the RL-10 is just too small for use on an Earth-to-LEO stage of a Saturn V-class vehicle.

This layout is putting the thumbs on the scale, comparing a two stage against a 2.5 stage design.

It's two stages against two stages; where is the extra half stage?
the way it is handled is comparable is all I meant
Quote
Quote
So I went in, and did some real world tweaking.  Increasing the second/core stage ISP to the SSME's, the difference in the fuel needed dropped, significantly.

As noted above, using the full SSME Isp is not realistic.
Adjusting for throttling and SL vs Vac time spent and a few other optimizations I know of, the average isp for an SSME for full burn in this comes to 443.
Quote
Quote
Taking into account the SSME's throttling, and using that to adjust the fuel burned during initial ascent to 0.12, you suddenly have the second stage neck and neck between the two.

In a Shuttle launch, the SSMEs burned about four times longer than the SRBs.  If they were throttled back to 2/3 power for half boost phase, then the boost-burn fraction would be about 0.2.  SLS, with five-segment SRBs, would stage higher and faster than did the Shuttle, which would tend to increase the boost burn fraction.  Hence, I suspect 0.25 is a pretty reasonable number.
Not by my math, quite off in fact.  Further, your argument above for SL vs Vac on engines, apply to the theoretical F-1x for a bit.  Based on time at SL vs Vac, the average F-1A isp is only 281-284, as it spends next to no time at vac, and most of its time in the thicker atmosphere.  This throws your whole curve off.
chuck - Toilet paper has no real value? Try living with 5 other adults for 6 months in a can with no toilet paper. Man oh man. Toilet paper would be worth it's weight in gold!

Offline modemeagle

  • Full Member
  • ***
  • Posts: 390
  • Waleska, GA
  • Liked: 27
  • Likes Given: 3
Data from my simulator for SLS with single F-1 engine boosters.

F-1A ISP
Sea Level: 270
Vacuum: 310
Average: 295 (95% of vacuum isp)
Burn Time: 163 seconds

SSME ISP
Sea Level: 364
Vacuum: 453
Average: 444 (98% of vacuum)
Burn Time:570.5 seconds

Using the vacuum ISP for calculation results in only a small error in calculations.

Offline Proponent

  • Senior Member
  • *****
  • Posts: 5596
  • Liked: 1142
  • Likes Given: 684
As noted above, using the full SSME Isp is not realistic.
Adjusting for throttling and SL vs Vac time spent and a few other optimizations I know of, the average isp for an SSME for full burn in this comes to 443.
Quote
Quote
Taking into account the SSME's throttling, and using that to adjust the fuel burned during initial ascent to 0.12, you suddenly have the second stage neck and neck between the two.

I'm still wondering why a boost burn fraction as low as 0.12 would be applicable the Shuttle, much less to an SLS-like booster, but let's go ahead use that figure while also boosting the core's Isp all the way to 450 s and dropping the stage 1/booster Isp to 280 s (not that the last two make a lot of difference).  Then, as you say, the masses of the structures of the 2-stage and and core-plus-boosters vehicles are similar.  However, every kilogram of reduced booster mass comes at the cost of nearly a kilogram of core mass.  That's not good.  And every newton of reduction in booster thrust costs about half a newton of core thrust.  That's probably not good either if the core engines are SSME-like while the booster engines are likely cheaper per unit thrust.

In addition, the model is probably biased in favor of the core-plus-boosters configuration, in that it assumes the same delta-V losses for both configurations despite the core-plus-booster configuration's 1) greater size and more complex shape, both of which will tend to increase drag, and 2) longer burn time.
« Last Edit: 05/11/2012 03:50 AM by Proponent »

Offline dwightlooi

  • Member
  • Posts: 83
  • Liked: 3
  • Likes Given: 0
The more I ponder the issue, the more convinced I am that the whole SLS architecture is plain stupid. It does not appear to be designed to accomplish the goal of space exploration, but rather to keep Shuttle suppliers and workers employed doing the same approximate things that they were doing before.

If you are dusting of the F1A -- that's a 2,065K lbs thrust engine. The logical SLS alternative based on that will be three parallel 5m booster stages with 1 x F1A a piece. You cross tank to the central engine the same way the Falcon Heavy does and put the J2X on the upper stage. That's an 75~80 ton to LEO vehicle right there. By designing your mission the right way you can scrap th need for a 130 ton vehicle and make do with 80 tons. Plus, because the vehicle does not rely on low thrust LH2/LOX engines, you can operate it without the side boosters and still get the 22 ton MPCV into orbit.

Offline vulture4

  • Full Member
  • ****
  • Posts: 1006
  • Liked: 312
  • Likes Given: 90
Just a couple thoughts -

Both the SSME and the F1 would be very expensive today due to their fabricated tube wall construction. On the SSME the tubes are as fine as spaghetti above the throat, and all the thousands of welds are critical. Moreover the preburner cycle used on the SSME requries a very complex injector plate to combine LH2, LO2, and the turbopump exhaust. These engines were designed when labor was (relatively) cheap and CNC machining had not been invented.

The SpaceX Merlin-2 is a new design loosely and indirectly based on the F1, incorporating modern channel-wall design, but keeping the F1s innovative film-cooling for the nozzle extension. It would have major advantages in manufacturing cost and also, potentially, for maintenance and reusability.
« Last Edit: 06/24/2012 08:07 PM by vulture4 »

Offline 93143

  • Senior Member
  • *****
  • Posts: 3038
  • Liked: 292
  • Likes Given: 1
SSME Block III was intended to use channel-wall construction, along with (I believe) other cost-reduction measures and improvements in robustness.  RS-25E should keep these features, though it doesn't need the same level of in-depth advanced health monitoring planned for the Block III because it's only used once.
« Last Edit: 06/24/2012 08:00 PM by 93143 »

Offline vulture4

  • Full Member
  • ****
  • Posts: 1006
  • Liked: 312
  • Likes Given: 90
SSME Block III was intended to use channel-wall construction, along with (I believe) other cost-reduction measures and improvements in robustness.  RS-25E should keep these features, though it doesn't need the same level of in-depth advanced health monitoring planned for the Block III because it's only used once.
As an expendable, how would if compare to the RS-68, which has considerably greater thrust?

Offline Downix

  • Senior Member
  • *****
  • Posts: 7087
  • Liked: 16
  • Likes Given: 1
SSME Block III was intended to use channel-wall construction, along with (I believe) other cost-reduction measures and improvements in robustness.  RS-25E should keep these features, though it doesn't need the same level of in-depth advanced health monitoring planned for the Block III because it's only used once.
As an expendable, how would if compare to the RS-68, which has considerably greater thrust?
Depends on volume.  At the same volume of let's say 10/year, the RS-68 would be about 15% less expensive.

Of course the SSME would have higher impulse, allowing one to use a parallel staged setup rather than a serial, saving the cost of the upper stage.
chuck - Toilet paper has no real value? Try living with 5 other adults for 6 months in a can with no toilet paper. Man oh man. Toilet paper would be worth it's weight in gold!

Offline FinalFrontier

  • Senior Member
  • *****
  • Posts: 4012
  • Space Watcher
  • Liked: 437
  • Likes Given: 159
SSME Block III was intended to use channel-wall construction, along with (I believe) other cost-reduction measures and improvements in robustness.  RS-25E should keep these features, though it doesn't need the same level of in-depth advanced health monitoring planned for the Block III because it's only used once.
As an expendable, how would if compare to the RS-68, which has considerably greater thrust?

Non starter RS 68 will not be used, ever. Far too much money and time required+stage changes to accommodation heavier less fuel efficient engines ect.

List goes on but the reasons have been discussed endlessly on many threads throughout the past 6 years.
« Last Edit: 06/24/2012 08:22 PM by FinalFrontier »
3-30-2017: The start of a great future
"Live Long and Prosper"

Offline 93143

  • Senior Member
  • *****
  • Posts: 3038
  • Liked: 292
  • Likes Given: 1
There's also the base heating issue.  RS-68s as they are now don't do well clustered under a wide core stage with large SRBs nearby.

Offline RyanC

  • Full Member
  • ****
  • Posts: 402
  • SA-506 Launch
  • Liked: 51
  • Likes Given: 9
Neither would have the J-2 in the Saturn II ground-launch concepts; which was why they developed heat shield concepts to protect the J-2s from the base heating environment imposed by the SRBs.

Offline Downix

  • Senior Member
  • *****
  • Posts: 7087
  • Liked: 16
  • Likes Given: 1
Neither would have the J-2 in the Saturn II ground-launch concepts; which was why they developed heat shield concepts to protect the J-2s from the base heating environment imposed by the SRBs.
Saturn II used HG-3's for the ground stage. Also the J-2 was regen unlike the RS-68.
chuck - Toilet paper has no real value? Try living with 5 other adults for 6 months in a can with no toilet paper. Man oh man. Toilet paper would be worth it's weight in gold!

Offline geoallegrezza

  • Member
  • Posts: 22
  • Liked: 0
  • Likes Given: 0
Saturn II used HG-3's for the ground stage. Also the J-2 was regen unlike the RS-68.

Some were HG-3.  MLV-SAT-INT-18 and -19 used J-2s with 120-inch solids and Minuteman solids, respectively.

Offline Prober

  • Senior Member
  • *****
  • Posts: 10313
  • Save the spin....I'm keeping you honest!
  • Nevada
  • Liked: 702
  • Likes Given: 728
  Taking into account the SSME's throttling, and using that to adjust the fuel burned during initial ascent to 0.12, you suddenly have the second stage neck and neck between the two. 

Starting is the problem with the SSME.  So when it comes to throttling what is the lowest level the SSME can operate?  Idea being startup at launch, throttle down until needed.

Can we deep throttle down the SSME? 


« Last Edit: 06/26/2012 04:15 PM by Prober »
2017 - Everything Old is New Again.
I have not failed. I've just found 10,000 ways that won't work. ~ by Thomas Alva Edison

Tags: