Can someone explain RAPIER engines to me?

[Clockwork13]

According to their stats, rapier engines didn't look that good to me. Can anyone explain why everyone seems to be obsessed with them?

[lincourtl]

It's a tradeoff -- a little less efficient in both modes for weight, and ease of use. I suspect they'll be refined (and hopefully remodeled) in future releases though.

[Rusty6899]

They are useful for building Single Stage To Orbit space planes. Basically, they can switch between a very high level of fuel efficiency in an oxygen atmosphere by taking in oxygen as a jet engine, and the functionality of a reasonably efficient rocket engine for when you are in the upper atmosphere/space.

Normally for space planes (although I haven't built many so I am not an expert) you have to have a mix of jet engines and rocket engines to reach orbit, but the RAPIER lets you get a 2-for-1 balance with one engine.

[tipsyMJT]

If you look at them in the designs of space planes they don't always have to be considered inefficient. Plenty of things change when making a SSTO spaceplane with RAPIER engines opposed to the old way of putting both jet and rocket engines in the design. With the old design there had to be around 3 engines in the back at the least (unless you did some weird thing with two fuselages on top of each other), that meant a lot of weight in the rear of the plane which meant you either had to have most of your wings near the rear or put a large counterweight on the front which is impractical so the former is mostly use . This often makes the rear of the plane a very crowded place. With the RAPIERs you can get away with just one engine in the back. The RAPIER's inefficiency may not be compensated for in very quantifiable ways but it's definitely a godsend for people that make SSTO spaceplanes.

Edited January 31, 2014 by tipsyMJT

[Captain Sierra]

The RAPIER allows considerably smaller SSTO construction. Instead of having a minimum of 3 fuselages (one radially attached to a center on either side) typically set up with 2 jets and a rocket, you can now have a single dtack and engine which does both (without doing some sketchy clipping, which some people do). Another advantage is it's the only air breathing engine that can have something attached below it, again without some sketchy cubic strut or I-beam clipping, so you can send an air breathing craft somewhere much easier. Additionally, the RAPIER has a considerably lower flameout point than stock turbojets, making it able to go on longer. Lastly, the auto-switching seems to be pretty well timed as it cuts them over right as they start to flame out and you enver get kicked into a flat spin (if using more than 1).

[ComradeGoat]

If you look at them in the designs of space planes they don't always have to be considered inefficient. Plenty of things change when making a SSTO spaceplane with RAPIER engines opposed to the old way of putting both jet and rocket engines in the design. With the old design there had to be around 3 engines in the back at the least (unless you did some weird thing with two fuselages on top of each other), that meant a lot of weight in the rear of the plane

A turbojet and 2 LV 48-77S engines is lighter than a RAPIER.

The only engine configs I use on my current generation spaceplanes are:

1 turbojet, 2 LV 48-77S

2 turbojets, 1 aerospike

2 turbojets, 1 NERVA

The only one of those where it might make any sense at all to use the RAPIER is the second: possibly a twin RAPIER design to replace the three engines, but I like to build those designs so that at a critical point, usually about 25K when asymmetric flameout starts to look like a risk, I can light all three engines, throttle down and get to orbit with a decentish ISP and a good amount of thrust without intake spamming. The RAPIER doesn't let me do that.

So after a bit of initial enthusiasm and experimentation, I've found it more or less completely useless in my spaceplanes.

[ComradeGoat]

The RAPIER allows considerably smaller SSTO construction. Instead of having a minimum of 3 fuselages (one radially attached to a center on either side) typically set up with 2 jets and a rocket, you can now have a single dtack and engine which does both (without doing some sketchy clipping, which some people do).

Nope. This is my workhorse SSTO shuttle. It's a lovely little spaceplane that flies very fast, makes orbit very quickly, and is basically lovely.

No sketchy clipping. Two cubic struts to mount the engines. If you have some sort of ethical objection to those, you can swap the LV 48-7S engines for the LV 24-77 radial mount engines. Plane still works fine.

I tried swapping this engine array out for a RAPIER, btw. The plane was around 5% heavier and didn't even come close to making orbit.

[Levelord]

Nope. This is my workhorse SSTO shuttle. It's a lovely little spaceplane that flies very fast, makes orbit very quickly, and is basically lovely.

http://www.sarahlizzy.com/LabrysIIIHanger.png

No sketchy clipping. Two cubic struts to mount the engines. If you have some sort of ethical objection to those, you can swap the LV 48-7S engines for the LV 24-77 radial mount engines. Plane still works fine.

I tried swapping this engine array out for a RAPIER, btw. The plane was around 5% heavier and didn't even come close to making orbit.

That's a pretty clever design. I hadn't thought about that.

[ComradeGoat]

That's a pretty clever design. I hadn't thought about that.

Thanks! The Mk1 to Mk2 convertor is really handy for that extra space at the back.

[Captain Sierra]

Nope. This is my workhorse SSTO shuttle. It's a lovely little spaceplane that flies very fast, makes orbit very quickly, and is basically lovely.

http://www.sarahlizzy.com/LabrysIIIHanger.png

No sketchy clipping. Two cubic struts to mount the engines. If you have some sort of ethical objection to those, you can swap the LV 48-7S engines for the LV 24-77 radial mount engines. Plane still works fine.

I tried swapping this engine array out for a RAPIER, btw. The plane was around 5% heavier and didn't even come close to making orbit.

I refuse to clip with cubic struts. Instead, I just enable surface attach on those engines and do the same thing with EditorExtensions. Lower part count.

[Wahgineer]

A turbojet and 2 LV 48-77S engines is lighter than a RAPIER.

The only engine configs I use on my current generation spaceplanes are:

1 turbojet, 2 LV 48-77S

2 turbojets, 1 aerospike

2 turbojets, 1 NERVA

The only one of those where it might make any sense at all to use the RAPIER is the second: possibly a twin RAPIER design to replace the three engines, but I like to build those designs so that at a critical point, usually about 25K when asymmetric flameout starts to look like a risk, I can light all three engines, throttle down and get to orbit with a decentish ISP and a good amount of thrust without intake spamming. The RAPIER doesn't let me do that.

So after a bit of initial enthusiasm and experimentation, I've found it more or less completely useless in my spaceplanes.

The thing that you seem to fail to realize is that the RAIPER reduces part count and weight overall. How? Since it doesn't need oxidizer for the first 10-25km (depends on the amount of intakes), you don't need as many rocket fuel tanks. Also, because it's a two in one engine, you have less of a risk spinning out, and don't need to add extra fuel tanks/fuselages to mount the jet engines. So while the Rapier may be heavy and some what inefficient, its unique abilities allows you to reduce the weight/part count over all, especially for the designs that use LVT-30/Turbojets set-up.

[The Error]

Because having o e engine in the place of 2 or more is just Cool.

Rule of Cool, guys.

[ComradeGoat]

The thing that you seem to fail to realize is that the RAIPER reduces part count and weight overall. How? Since it doesn't need oxidizer for the first 10-25km (depends on the amount of intakes), you don't need as many rocket fuel tanks. Also, because it's a two in one engine, you have less of a risk spinning out, and don't need to add extra fuel tanks/fuselages to mount the jet engines. So while the Rapier may be heavy and some what inefficient, its unique abilities allows you to reduce the weight/part count over all, especially for the designs that use LVT-30/Turbojets set-up.

See the design I posted above. No danger of spinning out, 5% lighter, faster, jets also don't use oxidiser, only needs rockets to circularise, no extra mounting points needed (you can mount a 48-7S pretty much anywhere by putting it on a cubic strut).

I can't see replacing two jets with one RAPIER. The RAPIER in jet mode is slower and less powerful, so if you needed two jets to lift a heavy SSTO (e.g. one that uses a NERVA as its main engine), you'll still need two afterwards, only it'll perform worse because the RAPIER is heavier, so you'll need to switch to rocket mode earlier, burn more fuel to get to orbit, AND have less range once there because your plane weighs more.

In summary, for light designs it performs significantly worse than a single turbojet and 2 LV 48-7S engines; for high endurance design it performs dramatically worse than 2 turbojets and a NERVA (good luck getting 5,000 m/s range from a single RAPIER engined design). It's heavy, slow, and it also looks like it's been beaten violently with the ugly stick.

I've tried to find a use for this engine, but I can't find a single one which isn't solved better with turbojets and careful plane design.

ETA: Even the twin turbojet + aerospike designs: 2 turbojets and an aerospace weighs 3.9 tonnes. 2 RAPIERS (no way you'll get a plane that heavy into orbit with just one) weighs 3.5 tonnes, slightly less, BUT: the RAPIER is less powerful than a turbojet, so you'll need to use rocket mode earlier, or airhog intakes, you still have the problem of asymmetric thrust, and the aerospike has a higher specific impulse.

So for my 3 reference designs:

1 turbojet + 2 LV 48-7S vs RAPIER - turbojet wins (RAPIER design doesn't even reach orbit)

2 turbojets + NERVA vs 2 RAPIERS - not even a contest. NERVA plane will take off from the KSC, land at Laythe, plant a flag and return to Laythe orbit while the poor RAPIER plane is still looking for a filling station somewhere around Duna.

2 turbojets + Aerospike vs 2 RAPIERS (This plane has over 3000 m/s delta V in space) - Aerospike design wins, but not as dramatically as the NERVA design.

Edited January 31, 2014 by ComradeGoat

[Jouni]

In summary, for light designs it performs significantly worse than a single turbojet and 2 LV 48-7S engines; for high endurance design it performs dramatically worse than 2 turbojets and a NERVA (good luck getting 5,000 m/s range from a single RAPIER engined design). It's heavy, slow, and it also looks like it's been beaten violently with the ugly stick.

I've tried to find a use for this engine, but I can't find a single one which isn't solved better with turbojets and careful plane design.

A basic engineering principle is that simplicity is usually better than efficiency. A good enough design that gets things done now easily beats a better design that still needs some fine tuning.

When you are still learning how to build spaceplanes that can reach orbit reliably, RAPIERs make the process faster. If you can just use one kind of engine that doesn't even need manual care during the ascent, you have less problems to solve simultaneously, and you can probably solve them quicker. Later, when you already know how to get things done, you can start experimenting with more complicated designs, if you need that extra efficiency for something.

Similarly, when you need to quickly build a spaceplane for some particular task, you'll often find that RAPIERs are efficient enough to get the job done. And because RAPIERs allow simpler designs, the plane can complete its mission sooner than a more efficient design with separate jet and rocket engines.

[Moar Boosters]

They're not that great on large scale craft.

[ComradeGoat]

Similarly, when you need to quickly build a spaceplane for some particular task, you'll often find that RAPIERs are efficient enough to get the job done. And because RAPIERs allow simpler designs, the plane can complete its mission sooner than a more efficient design with separate jet and rocket engines.

I agree that this engine has a role in helping soften the learning curve of building planes. However, I disagree with the above - the RAPIER is underpowered compared to the turbojet, which means you'll need to fly higher to attain orbital velocity while airbeathing, which means more intakes. The alternative is to start your rocket ascent from a lower speed/lower altitude, increasing your fuel requirement, which means increased complexity and increased weight. The simple fact is that the RAPIER is, basically, gutless and heavy, and to compensate for that you need more fuel, more intakes, more complexity. Turbojet designs can be lighter and faster, and if all you want is an orbital shuttle, just stick a pair of LV 24-77s on your wingtips. Job done.

Regarding more complex designs, I've done a little experiment. I have a plane, the Seakrait Mk1, which is a twin turbojet+aerospike design. If the RAPIER is going to win over a turbojet design, it's here because it's one of the rare cases where a RAPIER design is lighter. I flew to a 100k orbit (ish), and saw how much delta V I had left. I also took a snapshot as the plane crossed 10,000m. I then replaced the turbojets with RAPIERs and removed the aerospike. The plane was still well balanced and flew nicely. They took off with the same level of fuel (this plane does not take off fully fuelled, but has empty tank space to allow it to reach Jool after refuelling in orbit), were both flown in air breathing mode for as long as they were able to make progress to orbit, and then switched to rockets. First the original design, with the aerospike:

Passing 10K, mission time is 2:13, we're doing 220m/s and have 819 units of fuel left.

Orbit. Ignore the mission time - I fast forwarded until on the light side to get a better photo. Delta V remaining is 1467 m/s, 412 units of fuel, 493 units of oxidiser.

Now for the RAPIER. I wasn't able to climb as steeply to 10K, as it simply did not have the power, necessitating a shallower ascent. It passes 10K at 2:45 - 32 seconds slower, is travelling at 208 m/s, 12 m/s slower (remember, this craft is LIGHTER - most RAPIER craft will be HEAVIER than their equivalent jet powered craft) and has 816 units of fuel left, so it's burned 3 units MORE than the heavier craft.

Now orbit. I really did eke out as much as I could from air breathing mode, until I had to throttle back so much that I couldn't chase my apoapsis any more.

We have 1247 m/s of delta V left, 220 LESS than the HEAVIER craft, and have 396 units of fuel (less than the heavier craft) and 461 units of oxidiser (less than the heavier craft).

Frankly, this is a test the RAPIER should have excelled at. As a RAPIER is heavier than a turbojet + 2 LV 48-7S engines, it's pretty much always going to suck compared to that config. It's got no hope against a NERVA powered spaceplane. If it's going to be competitive, it should be against an aerospike, and it just isn't. Not only that, the aerospike plane was much nicer to fly, given the RAPIER powered plane struggled to gain speed at low altitude.

They don't lead to simpler craft, because you have to overbuild to overcome their deficiencies. As a rocket, they're OK, but heavy. As a jet, they're really quite poor.

[Captain Sierra]

Ima throw another variable in here. Turbojet RAPIER combo, like this:

the advantage here is that a RAPIER weighs the same as an aerospike, but has huge gimballing range (vs aerospike none) and can assist in the jet ascent. I find my old craft which used aerospikes are performing remarkable when the rocket is switched out for a RAPIER. Instead of kicking in the rocket when the turbojets flame out, you can now switch to just the turbojets, and then back to the RAPIER alone, which seems to have a very low flameout point, and milk the crap out of the atmosphere before autoswitch takes you rocket and you head off into space.

ComradeGoat, I challenge you to fly that aircraft in two additional configurations. First, run it with turbojets and a RAPIER. Also, run it with 3 RAPIERs. Then we can see if these engines are truly a no-win scenario.

[BubbaWilkins]

I think the results above are flawed due to an overly aggressive ascent profile and general lack of velocity at lower altitude. You should be pushing 400-600 m/s by the time you reach 10k. With as many intakes as you have, 1600-1800 near 20km. The difference is you have to gradually build up speed in and not just brute force punch into space.

If the 45deg angle of attack at 10km didn't give it away, the fact that you burnt any oxidizer at all before 10km did. With all those intakes and the relatively small mass of that craft, you should be pushing 1800m/s in near level flight at 20km on intake air alone.

All your comparisons above are also flawed in that the thrust performance is nearly 50% that of the Rapier craft in space. granted, your crafts are relatively small and can get away with it. The Rapier engines are clearly superior to any combination of other engines with similar performance.

EDIT:

It just occurred to me that you might have adjusted the oxidizer load. Fair enough, but the ascent angle is still too steep and relative velocities too low.

Edited January 31, 2014 by BubbaWilkins

[SuperBigD60]

I'm finding it hard to argue with Goat's logic, but I do think there's got to be some way to use a Rapier (however out of the box it may be) that no other rocket/jet is as efficient at.

I wonder if maybe some of the SSTO issues get resolved with FAR where your speed on air-breathing mode is effectively limited to about 1700 m/s or so.

Anybody have any other ideas?

[ComradeGoat]

ComradeGoat, I challenge you to fly that aircraft in two additional configurations. First, run it with turbojets and a RAPIER. Also, run it with 3 RAPIERs. Then we can see if these engines are truly a no-win scenario.

Ok, here goes. I'll write this as I do the test.

2 turbojets, 1 RAPIER (RAPIER in manual switching mode to try and get best performance out of it)

At 10K in 1:33, 282 m/s, 818 fuel left (got there quicker than the aerospike variant, consumed about the same amount of fuel). Conclusion - it's fast, which you'd expect it to be with 3 jet engines running simultaneously.

Started to get asymmetric flameout and had to reduce thrust lower than I normally would, as expected (25K - normally this plane can run at full power to 27K). Switched turbojets off at 30K (normally would be a bit higher) and going full thrust on the RAPIER in air breathing mode.

I'm pushing apoapsis, but only just. Let's see how close this can get me to orbital speed. I'm slower than I would have hoped for at this point because I had to throttle back earlier. Presumably adding intakes would have mitigated this.

I'm having to pitch up at 45º to push apoapsis - damn this thing is gutless in air breathing mode!

Flameout at 36K, reduced power, it's reignited.

I'm about to hit apoapsis, got to switch to rocket mode now.

Reached orbit without complications. Apoapsis 108km, Periapsis 100K, 1249 m/s delta V remaining, 381 fuel, 449 oxidiser.

OK, now for the three RAPIER config.

Initial thoughts on pulling off the runway - this feels about as powerful as the twin jet + aerospike config.

1:50 to 10K, 816 fuel left. Second fastest, but it's using more fuel than either of the turbojet configs.

Again, starting to have to throttle back at 25-26K

At 32.5 K I couldn't push apoapsis any more. Throttled up, engines switched to rocket mode.

Orbit, 107 * 102, 1046 m/s Delta V remaining. 371 fuel, 431 oxidiser.

So in order:

To 10K:

2 jets, 1 RAPIER: 1:33

3 RAPIERS: 1:50

2 jets, 1 aerospike:2:13

2 RAPIERs: 2:45

Take home message - 3 jets are faster than 2, turbojets are faster that RAPIERs (entirely expected result).

Delta V remaining on 100K orbit:

2 jets, 1 aerospike: 1467 m/s

2 jets, 1 RAPIER: 1249 m/s

2 RAPIERS: 1247 m/s

3 RAPIERs: 1046 m/s

Note that the more RAPIERS, the less Delta V the plane has left in orbit. 2 RAPIERs is gutless and you don't lose anything by dropping one of the RAPIERs and adding 2 jets. HOWEVER, you do need to faff about manually managing the switching mode.

The two planes powered entirely by RAPIERS suck, basically, a lot. 2 jets and 1 RAPIER is fast at low altitude, but in space and high altitude jets and an aerospike beats it every time.

Don't even get me started on 2 jets + 1 NERVA.

[ComradeGoat]

I think the results above are flawed due to an overly aggressive ascent profile and general lack of velocity at lower altitude. You should be pushing 400-600 m/s by the time you reach 10k. With as many intakes as you have, 1600-1800 near 20km. The difference is you have to gradually build up speed in and not just brute force punch into space.

It just occurred to me that you might have adjusted the oxidizer load. Fair enough, but the ascent angle is still too steep and relative velocities too low.

I climb steeply until about 15K, when I reduce my climb to pick up speed. I'm basically horizontal at 25-27K. This is not a rocket ascent - I just don't see the point in wasting fuel to make pretty flames at 15K and 800 m/s when I can punch through 800 m/s at 20K like it's not there.

I reiterate, I am not flying a rocket ascent profile, but going horizontal too early in space planes loses efficiency to atmospheric friction. You could spend your entire fuel load at 10K and never reach 500m/s if you fly like that.

What is the point in going supersonic at 8K? It doesn't make the engines work any better, you don't need it to collect intake air, and the amount of drag is horrible

I didn't adjust the oxidiser loads at all. All planes took off with the same amount of fuel and oxisider.

ETA: Here, to prove I'm not a muppet and do actuality know how to fly a spaceplane. This is the original variant (aerospike + jets) picking up speed at 27K:

Edited January 31, 2014 by ComradeGoat

[Jouni]

I agree that this engine has a role in helping soften the learning curve of building planes. However, I disagree with the above - the RAPIER is underpowered compared to the turbojet, which means you'll need to fly higher to attain orbital velocity while airbeathing, which means more intakes. The alternative is to start your rocket ascent from a lower speed/lower altitude, increasing your fuel requirement, which means increased complexity and increased weight. The simple fact is that the RAPIER is, basically, gutless and heavy, and to compensate for that you need more fuel, more intakes, more complexity. Turbojet designs can be lighter and faster, and if all you want is an orbital shuttle, just stick a pair of LV 24-77s on your wingtips. Job done.

Adding more fuel only makes the spaceplane heavier, not more complex. Adding mass means that the plane relies more on thrust and less on lift, making it simpler, as you don't have to spend that much effort on balancing CoL vs. CoM for different stages of flight. And as balance is no longer that important, you can modify the plane for different missions by adding and removing components, and still be quite sure that it will behave nicely.

Modularization is another good engineering practice. Things become simpler, if you don't set too ambitious goals for individual designs, but break the goals down into smaller subgoals. If you want a spaceplane that can carry 5 kerbals to Laythe, don't build a spaceplane that can carry 5 kerbals to Laythe. Build a simple spaceplane that can carry 5 kerbals to LKO instead. That plane can now use RAPIERs, because the design goal is quite modest. Then launch a transfer module that can get the plane (or a fuel tank or a spacestation module) to the orbit of Laythe. As the spaceplane and the transfer module can be pretty much independent from each other, the overall design is simpler than in a single capable spaceplane. Especially when you realize that you actually wanted a spaceplane that can carry 10 kerbals to Laythe.

[Claw]

I climb steeply until about 15K, when I reduce my climb to pick up speed. I'm basically horizontal at 25-27K. This is not a rocket ascent - I just don't see the point in wasting fuel to make pretty flames at 15K and 800 m/s when I can punch through 800 m/s at 20K like it's not there.

I reiterate, I am not flying a rocket ascent profile, but going horizontal too early in space planes loses efficiency to atmospheric friction. You could spend your entire fuel load at 10K and never reach 500m/s if you fly like that.

What is the point in going supersonic at 8K? It doesn't make the engines work any better, you don't need it to collect intake air, and the amount of drag is horrible

I didn't adjust the oxidiser loads at all. All planes took off with the same amount of fuel and oxisider.

ETA: Here, to prove I'm not a muppet and do actuality know how to fly a spaceplane. This is the original variant (aerospike + jets) picking up speed at 27K:

I agree, I try to climb up above 12km quickly. I'm usually only doing 200m/s at 10km.

Although I've been trying desperately to get a good use out of some rapiers. I feel like I'm onto something, but they still don't seem to have enough oomph. I haven't had much success on the smaller side, so I'm trying on the heavier side.

[ComradeGoat]

Modularization is another good engineering practice. Things become simpler, if you don't set too ambitious goals for individual designs, but break the goals down into smaller subgoals. If you want a spaceplane that can carry 5 kerbals to Laythe, don't build a spaceplane that can carry 5 kerbals to Laythe. Build a simple spaceplane that can carry 5 kerbals to LKO instead. That plane can now use RAPIERs, because the design goal is quite modest. Then launch a transfer module that can get the plane (or a fuel tank or a spacestation module) to the orbit of Laythe. As the spaceplane and the transfer module can be pretty much independent from each other, the overall design is simpler than in a single capable spaceplane. Especially when you realize that you actually wanted a spaceplane that can carry 10 kerbals to Laythe.

That's actually how I do my Laythe missions - I ferry planes there by nuclear tug, or if they're designed to be able to visit outer moons, under their own power.

If you use the RAPIER, your plane will be bigger, heavier, take more fuel from your nuclear tug to drag to Laythe, and waste more precious fuel every time you go visit the surface in your heavy, slow, RAPIER plane.

Small and nimble still works if you're building modular. The advantages don't change. A plane that can get to LKO is lighter and smaller if it uses a jet and 2 48-7S engines than if it uses a RAPIER. It's easier to haul to Jool and gets to and from the Laythe surface much more cheaply.

But, you know, enjoy hauling that fuel to Jool for your airborne SUV. ;-)

[Rhomphaia]

Purely mathematically there is no reason to use the rapier, just like there is no mathematical reason why you would ever use the Lv-909, Poodle, Skipper or the Mainsail...

...Or basically any engine other than the 48-7S or LV-N.

47-7S is seriously overpowered.