RAPIER velocity / thrust curve makes airframe matching difficult

[AeroGav]

I'm not sure if this is a point for General Discussion since I can certainly get to orbit with these things,  however, the inner min-maxer/perfectionist has hit an imponderable.

I've started building my spaceplanes with RAPIERS,  NERVS and liquid fuel only.    I size the wing such that it only needs 2-4 degrees angle of attack to maintain lift in the very thin air above 20km,  which with the stock aerodynamics, this is the optimum AoA for lift/drag ratio.   This means I can fly as high as possible on airbreathing mode, and also climb as well as possible given the meagre thrust of the NERV.

Recently I built a MK2 Trimotor with two NERV and one RAPIER, and no Oxidiser.   The payload was two mk2 crew cabins (4 T), the vehicle mass 30T, 2200 LF.

The problem is, at low altitude, the natural airspeed of this airframe, at 2deg AoA is very low - something like 100 m/s sea level, rising to 240 by 9km.    At these speeds, the single RAPIER provides only 70-80kn thrust, leading to a climb angle of just 5 degrees.  Slick though this airplane is, most of the thrust just gets used keeping airborne, with not much left over to climb.   

After about 15 minutes, I finally reach 10km and the natural, 2deg AoA cruise speed of the airframe is edging up 260m/s owing to the thinner air.  The RAPIER are finally starting to make halfway decent thrust and the climb angle has gone up to a heady 10 deg.  However,  if we go much faster we hit the transonic drag brick wall.    Fortunately, 10km is high enough that the NERV will give about 690 ISP.  It's not the full 800 they do in a vac,  but a lot better than the 160 they insult you with at sea level.    So, I light em up briefly and nose over to a shallow dive until i'm at 1.3 mach.

Of course, this takes you into the RAPIER more speed = more thrust positive feedback loop.   Between 10-20km,  it's tough to hang on to the dang thing and maintain the right climb profile, so that you hit the RAPIER max thrust velocity before the air gets too thin for you to have any power, but not too soon so that you overheat or waste fuel fighting drag.

In summary :

1) Below 10km Subsonic  - I have too much wing, and not enough engine

2) 10-20km  Supersonic - I have slightly too much engine, and it's tough to find the right climb angle before you bust your service ceiling

3) Above 20km Hypersonic - I'd say my airplane performs really well

Out of 2200 starting LF, I use about 45% below 10km while impersonating  a DC3.   
I use the same fuel again lifting myself from 10km subsonic to orbit, leaving about 15% to spare.

Did shoot a video of the launch but it's a big edit job.

 

So What's my point?

Well, first off is it realistic for a SABRE type engine to gain >5x the static thrust at mach 3.7?   I can understand this in a RAMJET, is it also realistic for a turboramjet, or do turboramjets just loose thrust more slowly than a normal engine owing to their ability to bypass the compressor / having a lower pressure ratio optimised for air that is already highly compressed by the inlet system?  I can't find much mention of this in the SABRE design, but does it also have similar features which would explain the incredible thrust growth?

The fact the engine produces such widely differing levels of thrust according to airspeed makes it hard to match to an airframe for all flight regimes, as alluded to above.

 

On a Practical Level...

what steps could you take to mitigate this in the design, and are they actually worth implementing?

1) I could add another RAPIER, which would also help >20km.  Would make the mad supersonic phase even crazier  and it's not really a low TWR SSTO then is it :-)

2) Could put decouplers on the back of the NERVs and attach some engines to use up to mach 1.3 then be jettisoned, but dumping jet engines over the countryside kind of goes against the SSTO spirit

[sal_vager]

I think the issue here is a misconception as to what the rapier is, it's not a ramjet, the whiplash is the ramjet at high speeds.

The rapier switches from an air breathing mode to a rocket mode, you're exchanging air with a percentage of oxygen to a 100% oxidiser.

Also, it isn't intended to be an exact copy of the sabre, which doesn't yet exist as a working engine anyway, the rapier is designed to get to space so has to have the performance to do that, but it's also got to be balanced against existing jets and rockets, it's able to fill both roles, but the cost of that is that it's not a master of either.

[AeroGav]

5 minutes ago, sal_vager said:

I think the issue here is a misconception as to what the rapier is, it's not a ramjet, the whiplash is the ramjet at high speeds.

To my understanding , the SABRE is a rocket engine, that can compress atmospheric air and cool it down with the incoming fuel stream  , in place of using liquid oxygen.     As such it behaves as a rocket engine in both modes, it's not going to care about velocity, though the bell nozzle will have an optimum atmospheric pressure for efficiency?

[sal_vager]

And the rapier isn't the sabre, it's just similar, in open cycle mode it uses the jet module and is a jet, not a rocket, the thrust effects just make it look that way

[numerobis]

You could add a Juno or two. They'll help you at low altitude, and for punching through the transonic regime, without costing you a lot of mass.

Idea the second: less wing. You must need a lot of wing mass to be flying flat and level at 20km.

[NathanKell]

Real life rocket engines have 4-8x the TWR of KSP rocket engines

Real life jet engines have 0.6-0.75x the TWR of KSP jet engiines.

As you can imagine, that leads to odd results when you try to combine them, and then compare to reality.

[jwbrase]

8 hours ago, AeroGav said:

 

 

 

 

 

On a Practical Level...

what steps could you take to mitigate this in the design, and are they actually worth implementing?

1) I could add another RAPIER, which would also help >20km.  Would make the mad supersonic phase even crazier  and it's not really a low TWR SSTO then is it :-)

I find that I need about one RAPIER and one NERV for every 15 tons of full-load mass.
 

8 hours ago, sal_vager said:

I think the issue here is a misconception as to what the rapier is, it's not a ramjet, the whiplash is the ramjet at high speeds.

The rapier switches from an air breathing mode to a rocket mode, you're exchanging air with a percentage of oxygen to a 100% oxidiser.

Also, it isn't intended to be an exact copy of the sabre, which doesn't yet exist as a working engine anyway, the rapier is designed to get to space so has to have the performance to do that, but it's also got to be balanced against existing jets and rockets, it's able to fill both roles, but the cost of that is that it's not a master of either.

Technically, neither the Whiplash nor the RAPIER matches the performance of a ramjet, because both can operate at zero airspeed. In any case, its high-speed, high-altitude performance in jet mode is better than that of the Whiplash, and an LF only RAPIER / NERV design that forgoes rocket mode on the RAPIER entirely actually works fairly well.
 

[AeroGav]

6 hours ago, numerobis said:

You could add a Juno or two. They'll help you at low altitude, and for punching through the transonic regime, without costing you a lot of mass.

Idea the second: less wing. You must need a lot of wing mass to be flying flat and level at 20km.

I just did a spot of accountancy -

9,400KG Fuel 

4,000KG Payload 

11,830KG Systems (engines, cockpit, landing gear, service bays, batteries, radiators, reaction wheels)

5,396KG  Airframe (wings, control surfaces, tailcones)

I have really gone to town on the airframe, loadsa wing, tail cones on every open node, been very generous with fins and canards, so the airframe mass is 17.6% of the takeoff weight.   Will be impressed if anyone manages more than that.      The fuel fraction is pretty low for a spacecraft, but Systems is the biggie - no doubt because i'm carrying 6 tonnes of reactor.

It needs about 35kn thrust to maintain level flight, and at times the RAPIER only does twice that, so half the thrust is being lost to do nothing more than just maintain.    If I doubled my thrust, I'd have three times as much excess power therefore climb three times faster, for twice the fuel consumption.  The best solution would be another RAPIER (since i'd probably be able to fly level to 29.5km instead of 25km) but I find the 10-20km part of the profile difficult to get right as it is, let alone making it even more frantic.         It's flown like an "S" shaped curve - as you transition supersonic and the RAPIER goes into Banzai mode,  you have to pitch up to avoid excess speed in thick air,  but then you need to start pitching down again so as to hit your top speed before busting the service ceiling.    I started pitching down at 16.7km and ended up ballooning to 24km, then overshot to 16km, finally hit 1250 on like the third oscillation !

The main thing about adding more engines is extra drag nodes.    I wonder how much two Juno will cost me?   Alternatively I could just mount a pair of  Wheezy turbofans on jettisonable pods,  they get 2.5x the ISP of a RAPIER and by themselves they'd give three times as much thrust (7 times the excess power for climbing!) and i can ditch them as soon as we go supersonic.

[numerobis]

6 hours ago, AeroGav said:

I just did a spot of accountancy -

9,400KG Fuel 

4,000KG Payload 

11,830KG Systems (engines, cockpit, landing gear, service bays, batteries, radiators, reaction wheels)

5,396KG  Airframe (wings, control surfaces, tailcones)

Reduce the dry mass by x% by ditching some of the wings, then reduce the fuel by x% so you maintain the same delta-V. Shedding mass while keeping the same engines means you will climb faster and break through the transsonic regime faster, so you might even save delta-V requirements and be able to reduce fuel even more.

For dry mass, maybe also see if you can do without the reaction wheels.

The Junos are small, so one Juno is a quarter the drag of a Wheezy. It's so small that it's fine to have it be asymmetric, for instance just above your tail cone, and offset the asymmetry with pod torque. It might not be enough extra push for you, even paired up. Incidentally, the Mk0 tank is the most efficient LF tank, which can help you slightly reduce your dry mass.

[AeroGav]

Well the videos of this flight are being uploaded as we speak.

 

 

It turns out, it's not as easy making these videos as Scott Manley makes it look, in fact I go so far as to call it rocket science.      On the first three vids, i had the game sound muted by mistake, so no engine noises,  just me regurgitating all the aero theories i ever heard in a random order into a cheap mike.

The fourth video has game sounds , but a muted mike.    What do you reckon the touchdown velocity of the above vessel to be?

[AeroGav]

1 hour ago, numerobis said:

Reduce the dry mass by x% by ditching some of the wings, then reduce the fuel by x% so you maintain the same delta-V. Shedding mass while keeping the same engines means you will climb faster and break through the transsonic regime faster, so you might even save delta-V requirements and be able to reduce fuel even more.

For dry mass, maybe also see if you can do without the reaction wheels.

The Junos are small, so one Juno is a quarter the drag of a Wheezy. It's so small that it's fine to have it be asymmetric, for instance just above your tail cone, and offset the asymmetry with pod torque. It might not be enough extra push for you, even paired up. Incidentally, the Mk0 tank is the most efficient LF tank, which can help you slightly reduce your dry mass.

I'm going to try two variants of this plane.  

The first will have a pair of Juno pods, which should double my excess power and thus climb rate below 10k.   They get about double the ISP of the RAPIER but the main thing is the time to supersonic will be halved.  It's not much extra mass, the pair of them equals half a type one engine, and i currently carry three.   

EDIT - testing complete.

Original aircraft - 1506/2080 LF remaining at 10,300M and 420M/S velocity.

With two extra Juno - 1733/2180 LF remaining at same altitude and velocity.

The second variant, i'll use C7 hardpoints to make  a pair of jettisonable type 1 engine pods, to which i'll attach Weasels.    They can power the plane up to near 10k and 0.8 mach on their own, with the Rapier shut down, drawing only on the fuel in their own engine pod.   I'll put a parachute in the nose cone of the engine pod, and just before they exceed the operating parameters of the chute, i'll trigger the chutes and jettison mechanism, and start up RAPIER.  

Edited December 10, 2015 by AeroGav

[Val]

Edit: You posted your video while I was typing.

@AeroGav I use a different ascent profile that puts me at 10 km and 1000 m/s in less than 6 minutes with a craft that sounds very similar to yours, if the Oxidizer is removed.Propulsion: 1 RAPIER, 2 LV-Ns   

• 2 Crew Cabins
• Liquid Fuel: 2220
• Dry Weight: 18.4 t.
• Wet Weight: 29.5 t. (without Oxidizer)

The craft has 8.26 wing area with 2° Angle of Incidence + 3.88 wing area from Mk2 parts and control surfaces with no Angle of Incidence. It was designed with 220 Oxidizer standard, but I tested and it can complete an ascent with it removed.

I build up speed to 450 m/s at sea level on RAPIER alone, before starting any kind of ascent. During most of the ascent the craft is facing directly prograde and it's never more than 1°-2° off (AoA), except when speed is below 150 m/s or altitude is above ~22 km.

Time to orbit is ~19 minutes MET, but the last ~7 minutes are using physics warp. The ascent profile is described in detail in the album below.

But granted it does require a very low drag craft. I tried adding RCS ports to it, and then it was no longer able to get past Mach 1 on RAPIER alone.

I hope you can use my observations for something.

Edited December 10, 2015 by Val
Ninja'd

[Val]

1 hour ago, AeroGav said:

It turns out, it's not as easy making these videos as Scott Manley makes it look, in fact I go so far as to call it rocket science.      On the first three vids, i had the game sound muted by mistake, so no engine noises,  just me regurgitating all the aero theories i ever heard in a random order into a cheap mike.

The fourth video has game sounds , but a muted mike.    What do you reckon the touchdown velocity of the above vessel to be?

I can't hear what you're saying in the video. Too low volume 

[numerobis]

The video is pointing out that your outer two vertical stabilizers are not quite head-on to the breeze -- they're getting lift inwards, which means drag.

In the SPH, use the rotate widget and hit 'f' to get into absolute mode, then rotate them imperceptibly; they'll snap to pointing forward.

That, or just delete them.

Edited December 11, 2015 by numerobis

[panzer1b]

On 12/9/2015, 7:09:07, AeroGav said:

1) Below 10km Subsonic  - I have too much wing, and not enough engine

2) 10-20km  Supersonic - I have slightly too much engine, and it's tough to find the right climb angle before you bust your service ceiling

3) Above 20km Hypersonic - I'd say my airplane performs really well

Why are you trying to climb without flying past the cieling?  The entire point (and the only reason you would ever use the rapier) is to get to orbit or do some suborbital stuff.  The whiplash is way way way better at jet mode unless you insist on going so fast that your plane burns alive. 

Overall i have to say that the rapier does perform the best of all the engines as a SSTO engine as that extra thrust at high speeds/high altitudes cuts your dV requirements in rocket mode down by quite a bit.  Its designed solely for high speed though, so dont expect good results out of the rapier itself at lower altitudes/low speed.

If i want a dual mode craft that is designed for both effective atmo travel and SSTO capabilities i usually go with a combo of whiplash and rapier or panther and rapier (just place the 2nd engine on the raper's rear node and clip in), its the most effective system to allow efficient low altitude/lower speed flight, and stuill lets you go orbital if needed.