TWR in a winged vessel..

[AeroGav]

There's been a lot written on this board about optimal TWR for rockets.  Math got involved,  some of which went over my head,  but the consensus seemed to be that 2 was a good number.

What I'd like to know is if the same Math can be used to prove the "best" TWR for a spaceplane in closed cycle mode.      

Now before everyone heads for the hills,  try not get too freaked out by the fact it has wings.

Wings generate lift, in exchange for creating a bit of extra drag and extra mass.    If you angle the wings up at 5 degrees relative to the fuselage, tweak the tailplane/canard so the plane holds a nose angle between half a degree and one degree nose up when  SAS is set to prograde,  your lift/drag ratio is going to be more or less constant throughout the closed cycle part of your flight.

Another way to look at it, is that since we have wings, and wings are counteracting gravity, we only have to be concerned with drag.  

So, whilst in a rocket, we are concerned about our margin of Thrust over Gravity,  in the airplane,  we are concerned about the margin of Thrust to Drag.    The greater the margin in favour of Thrust,  the lower our losses, but adding extra engines to increase TWR adds dry mass, and means we have less delta V to start with.        

Of course, this comparison is only valid if the airplane is generating enough lift to support itself  while building velocity.   You could set the wings at a lower angle, or zoom climb to a high altitude on jet power before going closed cycle,  and temporarily get lower drag, but it won't be sustainable as you'll soon fall back into thicker atmosphere and be worse off.      That is why it is best to chase optimal "lift:drag" ratio rather than absolute lowest possible drag.

 

 

Let's say it weighs 30 tons.  In Kerbin gravity, that works out to a gravity force of 300kn if you round things up  a bit.

Do we need 300kn lift ?  Well, not quite.     Let's say we're going at 1400 m/s -  that's actually two thirds of orbital velocity, meaning that orbital freefall is already going to be cancelling most of gravity.    We only need to get enough lift to make up the difference, and stop the airplane descending.    As our velocity continues to increase, our apparent weight decreases.   Since we're holding a constant AoA at constant lift/drag,   our lift now exceeds weight,  and we drift upwards, until the thinner air causes lift to no longer exceed weight.   At this new higher altitude however, drag is less, and so on.   As you can see,   this means it gets easier and easier to accelerate as time goes on,   and fuel burnoff is not the only factor at play.

At this point , some of you will be saying "so what" because,  with RAPIERs in closed cycle mode,  your TWR on this phase of flight is going to be so high as to make  such optimisations pretty irrelevant.

However,  this past week or so,  I've been building craft for rescaled Kerbin.      I'm currently using a rescale factor of 3.2 which raises orbital velocity from 2200 m/s (mach 7) to 4200 m/s (mach 14).      This makes things more than twice as difficult for a spaceplane.

Stock, you can get 1600 m/s air breathing,  and so theoretically only need 600 m/s delta V in closed cycle mode to make orbit.   

With the rescaled system,  even if you manage 1600 m/s air breathing, you need another 2600 delta V to make orbit - more than four times as much.        I don't think there's any chance of packing enough fuel in to do that with a 305m/s specific impulse RAPIER -  800 m/s NERVs are probably your only option.      However, they have much worse TWR -  3 tons for 60kn, instead of 180kn for 2 tons.     Pure chemical rocket engines are better still - the Dart manages RAPIER thrust levels for just one ton,  and is far from the highest TWR engine available.

Compounding the issue,   at the start of your closed cycle burn,  you are nowhere near orbital velocity on rescaled Kerbin, thus are still feeling the full effects of gravity.  This holds you deeper in the atmosphere,  with greater drag losses, for longer.

Thus,  this "optimal" spaceplane TWR becomes such a vexed question.

Starion 2 - A working prototype

I can SSTO with the rescaled Kerbin but such a craft ends up with little fuel left over in a very low orbit.      So the Starion 2 was a mod of the original SSTO that dumps its jet engines at flameout to get extra delta V.        Two whiplashes, three nervs.

Takeoff weight 47 Tons.   Upper stage mass (after Whiplash stage  separation) 38 tons

Spoiler

Why Whiplashes instead of RAPIERs?    Cost.        Whiplashes fade pretty badly after 1100 m/s , and you can expect another 300 m/s top speed with RAPIERs in air breathing mode, but Whiplash are only 2200 Kredits vs 6000 for the Rapier,  and can lift more payload due to better low speed performance.    I did contemplate carrying a single RAPIER to orbit,  and use two droppable Panthers to boost us to mach 2.8,  where the single RAPIER gets ramjet-happy and will boost us to a top speed 300 m/s or so higher than the pair of whiplashes would have managed.  But that only takes 10% or so off our closed cycle burn requirement to get to orbit , hardly worth the mass penalty this RAPIER would bring.   And dumping it is uneconomic !

Mass breakdown  of the upper stage

Stage 2 Mass 38.6t

Liquid fuel        19 t
Nerv engines       9t
Payload            6.1t       (cockpit, crew cabins, docking gear, docking fuel)
Aerodynamic bits   4.5t       (wings, control surfaces, intakes and cones)

Flight Logs

I took some screenies on the way up.   Due to the AeroData GUI being open,  they give some insight into how drag losses changed as the flight progressed :

Time	Velocity Weight	Alt	L/D     Drag
(sec)	(m/s)	 (kN)	(M)	(ratio)	(kN)
					
752     1491	 345	32340   3.7     73
827     1674     328    33435   3.7     80
972     2186	 296	38824   3.6     64
1102	2723	 264	47842   3.6     28
1174	3114	 248	50712   3.5     26
1304	3904	 218	61815   3.4     12

Imgur album of the ascent shots from which this data was transcribed - 

https://imgur.com/a/xNSfY

[Geschosskopf]

Once a spaceplane is done breathing air, the air is so thin that its wings are doing essentially nothing:  they're not producing any noticeable lift, nor are they producing noticeable drag.  The thing is only maintaining altitude or even climbing by virtue of its velocity at its current altitude and/or its raw thrust.  IOW, at that point, the vehicle is essentially the same as the sustainer stage on a conventional rocket lifter.  So rocket rules apply.

[Spricigo]

14 hours ago, AeroGav said:

...   I don't think there's any chance of packing enough fuel in to do that with a 305m/s specific impulse RAPIER -  800 m/s NERVs are probably your only option.      However, they have much worse TWR -  3 tons for 60kn, instead of 180kn for 2 tons.     

 

 

The image (with all that info tabs) is quite telling.

Weight 350kN, TWR 0.27

Quick math tell us 3 nervs are providing 95kN of thrust (so Isp it's nowhere near 800s). And we need to cope with 65kN of drag. So there's only 30kN left to accelerate the vessel.

Conclusion: A single rapier would provide better TWR and consume less fuel.

edit: I wondering where that 0,27 come from. Certainly an optical illusion cause by my phone tiny screen.

Thanks for the correction, @Aegolius13

Edited November 9, 2017 by Spricigo

[Aegolius13]

14 hours ago, Spricigo said:

The image (with all that info tabs) is quite telling.

Weight 350kN, TWR 0.27

Quick math tell us 3 nervs are providing 95kN of thrust (so Isp it's nowhere near 800s). And we need to cope with 65kN of drag. So there's only 30kN left to accelerate the vessel.

Conclusion: A single rapier would provide better TWR and consume less fuel.

Can't say I follow.  According to AeroGUI (by Valentina's mouth), this plane is producing 178 kN of thrust, which means the NERVs are operating at close to peak efficiency (which would be 180).  They still have issues due to inherently bad TWR, but I don't think it's related to altitude here

 

17 hours ago, AeroGav said:

With the rescaled system,  even if you manage 1600 m/s air breathing, you need another 2600 delta V to make orbit - more than four times as much.        I don't think there's any chance of packing enough fuel in to do that with a 305m/s specific impulse RAPIER -  800 m/s NERVs are probably your only option.      However, they have much worse TWR -  3 tons for 60kn, instead of 180kn for 2 tons.     Pure chemical rocket engines are better still - the Dart manages RAPIER thrust levels for just one ton,  and is far from the highest TWR engine available.

It looks like you're doing a TSTO with Whiplashes detaching?  If you do it that way, your plane should be pretty much full of fuel when you jettison the tanks, right?  If so, you can just run the Rocket Equation to figure out the mass ratio you need.  With a 340s ISP engine, your stage needs to be about 65% fuel.  Spaceplanes of course add extra weight for wings, landing gear, etc., but this still sounds within the realm of what's doable.  

 

Going back to the original topic - personally I try to have a starting TWR at least in the vicinity of 1 for spaceplane rocket stages, but that's not based on anything scientific, and there are a few variations.  For example, if I'm planning on using nukes, they will probably have a significantly lower TWR that that.  My solution in those cases is to bring some oxidizer along, and run the RAPIERS in jet ROCKET mode to get a burst of higher thrust after my jets cut out.  This pushes apoapsis out and buys more time for the weaker nukes to get the rest of the way to orbital velocity.  

 

 

 

 

Edited November 9, 2017 by Aegolius13

[Spricigo]

7 hours ago, Aegolius13 said:

My solution in those cases is to bring some oxidizer along, and run the RAPIERS in jet  rocket mode to get a burst of higher thrust after my jets cut out.  This pushes apoapsis out and buys more time for the weaker nukes to get the rest of the way to orbital velocity.  

I suppose you don't need oxidizer for jet mode.

In any case this is a matter of Power. The vessel engines need to be able to 'pump in' more energy than what's lost to gravity and aero drag. 

[AeroGav]

8 minutes ago, Spricigo said:

I suppose you don't need oxidizer for jet mode.

In any case this is a matter of Power. The vessel engines need to be able to 'pump in' more energy than what's lost to gravity and aero drag. 

Just doing some quick maths,     I staged at 1400 m/s with 5k dV,   reached orbit at 4200 m/s with 1400 dV remaining.     

So we used 3600dV to add  2800 m/s velocity,  IOW losses of 28%.     I don;t know how trustworthy Kerbal Engineer is in the VAB,  but I find I have to overbuild rockets by at least 50% to be sure of making orbit.

It looks like the design could potentially carry a bit more fuel, without losses absorbing all the gains.      I suppose there's a cost vs capability metric here.     KSP attaches an unrealistically high price to liquid fuel ,  so if you're adding more fuel and most of that fuel is just going to lift itself, then it might be best not to go there.   Obviously, if you make the upper stage heavy enough to require more Whiplashes, that's also cost.        5k is a lot of delta V to put in a single stage that has to have reasonable TWR anyway.

 

9 hours ago, Aegolius13 said:

It looks like you're doing a TSTO with Whiplashes detaching?  If you do it that way, your plane should be pretty much full of fuel when you jettison the tanks, right?  If so, you can just run the Rocket Equation to figure out the mass ratio you need.  With a 340s ISP engine, your stage needs to be about 65% fuel.  Spaceplanes of course add extra weight for wings, landing gear, etc., but this still sounds within the realm of what's doable

Would be something to see for sure.   @eloquentJane might be interested in the results.          Remember,  I used a rescale factor of 3.2 for Kerbin and 1.4 for the atmo

[Gordon Fecyk]

9 hours ago, Aegolius13 said:

My solution in those cases is to bring some oxidizer along, and run the RAPIERS in jet mode to get a burst of higher thrust after my jets cut out.

I've been playing with a part mod to stretch the 'air breathing' aspect of the Rapier a little further. Even though the jet mode flames out above 20 km, the intakes are still gathering some air. What if we could make oxidizer out of it?

The compressor parts are going through part balancing, but right now I have a part that uses about half of the power of an industrial oxygen concentrator and some mineral acting as a sieve of sorts, to make oxidizer out of intake air.

While the video there demonstrates unrealistic power consumption, the parts can still replicate this behaviour while consuming about fifty times the electric charge. I can squeeze a few hundred more m/s and a few more km of altitude out of the Rapier before going full closed cycle. The compressors can also turn a stock rocket engine into an air-augmented rocket of sorts.

[Aegolius13]

1 hour ago, Spricigo said:

I suppose you don't need oxidizer for jet mod

Yeah... that.

[Spricigo]

47 minutes ago, AeroGav said:

Just doing some quick maths,     I staged at 1400 m/s with 5k dV,   reached orbit at 4200 m/s with 1400 dV remaining.     

So we used 3600dV to add  2800 m/s velocity,  IOW losses of 28%.     I don;t know how trustworthy Kerbal Engineer is in the VAB,  but I find I have to overbuild rockets by at least 50% to be sure of making orbit.

You also raised the altitude what also have its deltaV cost, meaning that your losses are lower than you expected (unless you are counting that as 'gravity losses')

Off course the that losses, or either the ratio at which you accumulate it, are not constant, aero drag decrease  with altitude, effective gravity decrease with horizontal velocity. So you don't need a higher TWR all the way, as Aegolius suggested with rapiers, just enough to get through the demanding phase and get enough time to your weaker engines do their  thing. 

[AeroGav]

18 minutes ago, Spricigo said:

You also raised the altitude what also have its deltaV cost, meaning that your losses are lower than you expected (unless you are counting that as 'gravity losses')

You don't have the formula for that do you?   We gained 92km altitude..

[Laie]

17 hours ago, Geschosskopf said:

Once a spaceplane is done breathing air, the air is so thin that its wings are doing essentially nothing:  they're not producing any noticeable lift, nor are they producing noticeable drag.

Mechjeb's drag meter disagrees; so does the debug readout, or (probably) the seismic accelerometer if you turn off the engines and try to coast. On a plane pointing 20deg above prograde, drag after shutdown easily amounts to 0.1g, and often more. I find that it decreases nowhere as fast as I'd like it to, sometimes even the contrary. 1200m/s @ 30km or 2400m/s @ 45km produce comparable drag.

When I was dabbling in LF-only designs, I found that 0.3g was about the least acceleration that I could still get to orbit, eventually(1). Most of that was spent fighting drag and gravity, only a small fraction went into acceleration. Trouble started even before, as I couldn't milk the jets for all their potential airspeed: I needed upwards momentum to keep me aloft long enough for the nukes to do their thing, forcing me into a faster and steeper climb and only getting a mere 1100m/s or less out of the jets.

Increasing rocket TWR just a little makes it a lot easier, you can afford to point more prograde-ish, thus reducing drag. So you not only have a higher TWR to begin with, you also incur fewer drag and gravity losses, can eject from the usable atmosphere at a shallower angle and higher speed.... Win-win-win across the board. I'd say that anything beyond TWR=1 is overkill.

(1) I was carrying some payload, too, so TWR didn't increase all that much due to spent propellant. Perhaps 0.45g by the time I made orbit. With less deadweight, you may be able to afford a lower TWR at first, as it increases faster and to higher values before you are done.

[Spricigo]

2 hours ago, Laie said:

Increasing rocket TWR just a little makes it a lot easier, you can afford to point more prograde-ish, thus reducing drag. So you not only have a higher TWR to begin with, you also incur fewer drag and gravity losses, can eject from the usable atmosphere at a shallower angle and higher speed.... Win-win-win across the board. I'd say that anything beyond TWR=1 is overkill.

A good sum up of the advantages of a bit of extra TWR at the right moment. The hard part is, with all those variable, figure out when the right moment is happening.

[AeroGav]

6 hours ago, Laie said:

Mechjeb's drag meter disagrees; so does the debug readout, or (probably) the seismic accelerometer if you turn off the engines and try to coast. On a plane pointing 20deg above prograde, drag after shutdown easily amounts to 0.1g, and often more. I find that it decreases nowhere as fast as I'd like it to, sometimes even the contrary. 1200m/s @ 30km or 2400m/s @ 45km produce comparable drag.

When I was dabbling in LF-only designs, I found that 0.3g was about the least acceleration that I could still get to orbit, eventually(1). Most of that was spent fighting drag and gravity, only a small fraction went into acceleration. Trouble started even before, as I couldn't milk the jets for all their potential airspeed: I needed upwards momentum to keep me aloft long enough for the nukes to do their thing, forcing me into a faster and steeper climb and only getting a mere 1100m/s or less out of the jets.

The mistake is pitching to 20 degrees above prograde.    The lift:drag ratio is extremely sensitive to single digit changes in angle of attack.     My airplane flies at about 0.4 degrees above prograde when you set prograde hold due to built in nose up trim in the canards.  The wings are angled up relative to the fuselage at 5 degrees.   In this condition you get about 3.7 to one lift:drag ratio at mach 2 through to mach 7.    

It's got some trim flaps that raise the nose to 1.7 degrees over prograde - this reduces the lift : drag ratio to  a bit under 3.2 to one.      Above 5 degrees AoA the numbers will be pretty awful.

My pitch angle (relative to the horizon) varied from -0.4 nose down to a max of 3.6 nose up.  So the engines were thrusting us forwards.   I have tried getting into zoom climbs when on jet power on other space planes and found it uses more fuel than trying to keep it flat.    Your engines are  thrusting down against gravity instead of straight up,   but your wings make less lift because you prematurely climb to a high altitude (before having enough velocity to get sufficient lift up there) , then either fall down into the thicker atmosphere or pitch up to a really inefficient angle of attack trying to make lift, where the lift drag ratio is probably less than 1.

1. Maintain the angle of attack that gives best lift drag ratio at all times when on closed cycle power.  You're on the expensive juice, don't waste it. 

2. Try to design the aircraft in such a way as to make this easy to do.

3. When still on jet power,  the rules are different.    Try to get your best top speed in level flight,   even if this means pushing the nose down to an angle of attack that gives less than optimum lift:drag ratio,  because you need to be low enough in the atmosphere to make power.     Stay just under 17km on Whiplashes and 21km on Rapier.

Edited November 10, 2017 by AeroGav

[AeroGav]

9 hours ago, Spricigo said:

Off course the that losses, or either the ratio at which you accumulate it, are not constant, aero drag decrease  with altitude, effective gravity decrease with horizontal velocity. So you don't need a higher TWR all the way, as Aegolius suggested with rapiers, just enough to get through the demanding phase and get enough time to your weaker engines do their  thing. 

Well, at 972 seconds, 2186 m/s (about half orbital velocity)  and 38km (about 40% of the way to edge of the atmosphere) I had 64kn drag .   But my next screenie, at 1102 seconds, had me at 2723 m/s and 47km (about 50% of the way to the edge of the atmosphere)  , drag had fallen to 28kn.

So,   we could replace that third nuke with a Terrier, then dump it and its associated tank.   That means a Terrier firing for 368 seconds, over which time it would consume 2811 kg fuel. After that point it would depart and not burden us with useless dry mass. Hmm.

[Spricigo]

13 hours ago, AeroGav said:

You don't have the formula for that do you?   We gained 92km altitude..

well, I  just use the vis-viva equation a few times to get into the ballpark.

v=orbital velocity

GM = standard gravitational parameter of the central body

r=current altitude (measured from body's center)

a=semi-major axis.

 

First lets use it to find GM, you ended in 125km orbit (r=a=(600km*3.2)+125km=2.045Mm) at 4.2km/s. Insert everything there (or whatever calculator you prefer) and we have GM=3.62x10¹³ m³/s²

Now, in most screenshots you have the navball in surface mode (obvious reasons), so you are not accounting for the little boost from the rotation of the planet (something I didn't noticed earlier). If that was stock  I'd just add 175m/s and carry on, you could just check the value  for the scale you are using or calculate it. From the screenshot you had Pe=-1750km (170km from Kerbin's center) and Ap32km(1952km form Kerbin's center), so r=1.95Mm, a=1.06Mm, GM=3.62x10¹³ m³/s²  using the solver again v=1.7km/s. (rotation speed is about 290m/s?)

Ok, at that point I will assume that we can have an instantaneous boost of velocity that raise the orbit to 32km x 125km. (r=1.95Mm, a=2Mm, GM=3.62x10¹³ m³/s²)  solving fo v=4.35km/s. so we would need that instantaneous boost to be 4.35-1.7=2.65km/s.

At the apoapsis of that 32x125km orbit (r=2.05Mm, a=2Mm, GM=3.62x10¹³ m³/s²)  the velocity would have dropped to 4.15km/s, which is 50m/s short of a 125x125km orbit.

So you have it, 2.7km/s to raise your velocity (2.5km/s) and orbital height (200m/s). Obviously, there is no instantaneous boost of velocity by 2,7km/s, you need to break it down into smaller steps to makes the calculation more accurate. (How much would you like to make a spreadsheet?) Also notice the roundings along the way.

 

4 hours ago, AeroGav said:

Well, at 972 seconds, 2186 m/s (about half orbital velocity)  and 38km (about 40% of the way to edge of the atmosphere) I had 64kn drag .   But my next screenie, at 1102 seconds, had me at 2723 m/s and 47km (about 50% of the way to the edge of the atmosphere)  , drag had fallen to 28kn.

So,   we could replace that third nuke with a Terrier, then dump it and its associated tank.   That means a Terrier firing for 368 seconds, over which time it would consume 2811 kg fuel. After that point it would depart and not burden us with useless dry mass. Hmm.

You are already taking in account lift, weight, drag and thrust. Add to the mix the centrifugal force*  =mv²/r  where m=mass of the craft, v=orbital speed and r=distance from Kerbin's center.

About the idea of replace one of the nukes with a chemical droppable stage: why not try a Swinvel, Relliant or some Thuds? There are two advantages, 1.more excess thrust to deal with losses 2.faster consumption of fuel, thus weight reduction, thus increase in TWR. . Off course there is issue of how the  higher wet mass  and drag of all those parts will affect the jet performance and the cost in comparison with the cheaper terrier or recoverable nerv (beside the point: to some extent solvable by making the chemical engine recoverable also)

*for the sake of calculations, not entering the whole force vs pseudo-force vs just inertia debate.

Edited November 10, 2017 by Spricigo

[FleshJeb]

You want to have a truly ridiculous notion--Try a 3-STO. If you're going to toss away Whiplashes, tossing cheap rocket parts isn't that much worse.

I tested about 10 different major configurations on normal size Kerbin, and extrapolated remaining dV for 3.2x.

https://www.dropbox.com/s/c6zs66lgjs56zvj/AeroGav 3STO.craft?dl=

Flight profile:

Get to crusing altitude
Hold 0 pitch for the jet speed run
Fire NERVA and slowly pitch to +5
At jet burnout, ditch jets and fire Terriers.
Pitch to +8 until the lift begins to run out (AOA gets past 5 n the Aero GUI)
Discard side tanks and Terriers when all fuel runs out. (I had a little Liquid Fuel left when the Oxidizer ran out.)
Pitch to prograde.
(I had over 4k dV remaining after circularizing on normal Kerbin)

AG1 - Whiplashes
AG2 - NERVA
AG3 - Terriers

AG 5 - Service bays
AG 6 - Solar Panels

FYI, I found out that the 1.25m reaction wheels are remarkably draggy. I tucked mine in service bays.

EDIT: This one is better:

https://www.dropbox.com/s/f29u0sl2k4r3130/AeroGav 2STO.craft?dl=0

 

 

Edited November 10, 2017 by FleshJeb

[Spricigo]

2 hours ago, FleshJeb said:

You want to have a truly ridiculous notion--Try a 3-STO. If you're going to toss away Whiplashes, tossing cheap rocket parts isn't that much worse.

I tested about 10 different major configurations on normal size Kerbin, and extrapolated remaining dV for 3.2x.

Next steps: 

1.Replace jets with SRBs

2.Remove wings.

[AeroGav]

9 hours ago, Spricigo said:

About the idea of replace one of the nukes with a chemical droppable stage: why not try a Swinvel, Relliant or some Thuds? There are two advantages, 1.more excess thrust to deal with losses 2.faster consumption of fuel, thus weight reduction, thus increase in TWR. . Off course there is issue of how the  higher wet mass  and drag of all those parts will affect the jet performance and the cost in comparison with the cheaper terrier or recoverable nerv (beside the point: to some extent solvable by making the chemical engine recoverable also)

The 3 nerv version probably has a near-optimum thrust-drag ratio already at the start of the closed cycle burn.    Later it becomes excessive, which is why the 2 nerv 1 terrier version might improve things , by allowing us to shed some dry mass after 2700km.  

A swivel, thud or reliant is heavier and puts more load on the jet stage, though atm that could easily carry more.   But atm the mass would be better devoted to more fuel or payload than engine.   As for carrying the LFO engine to orbit if choosing an expensive one like a Dart, that kind of defeats the purpose of mass shedding.   Also, its worth bearing in mind these LFO tanks actually have quite high recovery value even when empty.  It can actually make sense to chuck the LF tanks and low tech LFO engine, but keep the empty LFO tank, if you look at recovery value vs mass.

3 hours ago, FleshJeb said:

DIT: This one is better:

Just remember to actually fly the thing to orbit on rescaled Kerbin before making performance claims !  The craft I built started life as an SSTO - in Kerbin orbit it used to have 3k delta V left over,  but the increased losses in rescaled Kerbin meant it just failed to orbit.      The TSTO has better nuke TWR and fuel fraction.

 

edit - craft file https://www.dropbox.com/s/v3enet77pm1jl5i/Starion2.craft?dl=0

[Spricigo]

3 hours ago, AeroGav said:

 As for carrying the LFO engine to orbit if choosing an expensive one like a Dart, that kind of defeats the purpose of mass shedding.

The purpose is not mass shedding, its to get an effective TWR (I guess also for an acceptable price). You can either do it by reducing the mass or by increasing the thrust.(yea, Moar boosters!!!)

The actual problem with replacing the Nerv with a chemical rocket is not so much the added weight of the engine+tanks (that may even be less than the weight of the Nerv). But rather that after you run out of oxidizer you have one less nerv at a flight phase the gravity/aero losses are still relevant.

In any case there is a 'philosophical divergence' there, you like to design your SSTO to smooth ride to orbit while for mine I trust in thrust. Your use less fuel, mine reach orbit faster, nonetheless both get the same job done, and probably neither is the ideal for anyone else.

Spoiler

 

 

[AeroGav]

58 minutes ago, Spricigo said:

In any case there is a 'philosophical divergence' there, you like to design your SSTO to smooth ride to orbit

Can't eat Snacks while pulling heavy Gs.       Not sure what the structural integrity of a Pretzel or Cadbury's flake is but I don't want to take any chances.    Main goal is to reach orbit in a way that minimises loss of M&M's  

1 hour ago, Spricigo said:

or mine I trust in thrust

Most people would consider one Poodle rather little for 12 Kerbals ! Good ship, looks like an efficient mid career plane (modest tech level)

[Spricigo]

29 minutes ago, AeroGav said:

Most people would consider one Poodle rather little for 12 Kerbals ! Good ship, looks like an efficient mid career plane (modest tech level)

I confess: considered the Skipper for a moment.   OTOH I have 4 Phanters for 20t of ship.

[FleshJeb]

On 11/10/2017 at 7:20 AM, AeroGav said:

Just remember to actually fly the thing to orbit on rescaled Kerbin before making performance claims !  The craft I built started life as an SSTO - in Kerbin orbit it used to have 3k delta V left over,  but the increased losses in rescaled Kerbin meant it just failed to orbit.      The TSTO has better nuke TWR and fuel fraction.

 

edit - craft file https://www.dropbox.com/s/v3enet77pm1jl5i/Starion2.craft?dl=0

I WAS getting a bit ahead of myself. I only upgraded from v0.25 a couple of weeks ago, and I've been experimenting like mad. I've gotten pretty good at designing for drag and thermal.

I flew your craft on 1x Kerbin--It's beautifully balanced and tuned. Very, very elegant.

So, I just downloaded 3.2x (Galileo88 - is there another version?):

HOLY CRAP, YOU WERE *NOT* JOKING. I did not get to space today. I can get my Apo out of atmosphere, but I don't have enough left to circularize.

The loss of comparable centripetal force is killing me.

Just based on that, I'll need about 175% of the lift that I've currently got at the switch from jet speedrun to closed cycle. I calc'd [g(altitude) - a(v_tangent)] for both scalings.

Based on my necessary pitch angle (5deg up craft pitch = AOA 5deg) (35km @ 2450m/s), I'll need about 200% of my current lift to maintain a zero craft pitch [sin(10)/sin(5)] ~ 200%

I want to verify that the atmo scaling is 16% higher. Is this correct? In 1x, I like to do my speedrun at 21km => 24.4km, and get above 32km => 37.1km before going on pure nukes, to clear the worst of the atmosphere.

Speedrun: 1x Kerbin = 21.5km @ 1667m/s,  3.2x Kerbin = 21km @ 1667m/s.

So, I know you want to answer what an optimal range for TWR is--I think the another pertinent number is L/D.

11.32 "lift units" @ 5deg for 26.375tons = 0.42 lift / ton. I'm going to try for 0.84 lift/ton and see how that works.

You're at about 27.75 lift/36.638tons = 0.75 lift/ton

This is assuming we're both close to minimum drag for a "3-column" design with an Inline Cockpit, and two Crew Cabins. More wing will be a little worse, but we'll see what happens.

For the TWR question, you want the Nervs to be a least better than a good LFO Isp.

In your top pic, you've got 180kN thrust - 65kN drag = 115kn. This goes to an effective Isp of 800*115/180 = 511

My effective Isp with one Nerv was between 266 and 133.

These are not counting the work the wings are doing, but I don't know how to math the radial out contribution. I guess you could integrate how fast the drag falls off to figure how it affects the average effective Isp for the whole run.

How much range do you want in space? That affect the choice of discarding the Nerv entirely or not? If we're just meeting a space station, it could be minimal.

FURTHER THOUGHTS:

Did another jet speedrun with my design: 1669m/s @ 21km => 100.5kN drag.

This tells me a few things about that moment in the flight :

• If your design experiences similar drag at similar speeds, you'd have an effective Isp of 356. [800 * (180kN -100kn)/180kN]
• When I light my two Rapiers and the Nerv, my effective Isp is 254, but that changes QUICKLY. In zero drag = 336
• If I had two Nervs, and one Rapiers quit, I'd only have an effective Isp 0f 270. In zero drag = 405

In short, yes, it's possible to establish an optimum TWR/engine choice if we can graph drag, speed, and altitude for some reasonably optimal zoom climbs for the higher TWRs. I think I'll write a KRPC program to log data. That'll be later this week.

Edited November 12, 2017 by FleshJeb
Unretarded myself

[Spricigo]

1 hour ago, FleshJeb said:

In short, yes, it's possible to establish an optimum TWR/engine choice if we can graph drag, speed, and altitude for some reasonably optimal zoom climbs for the higher TWRs.

Not an easy task by any means. Good luck.

[FleshJeb]

I just got this to orbit with over 2150m/s left in the tank. It could still use a lot of tuning.

The Wet-Mode TWR is about 1.2, when the Nerv starts running solo, it's 0.25-ish.

1.5deg up on the runway.
10 deg up, slowly leveling off to 0 at 22km.
Run to about 1635 @ 23.4km.
Hit Srf+0, engage Closed Cycle and Nerv.
Around 40km, the Rapiers burn out, and the drop tanks stage.
Keep Srf+0 until pitch = 0 (around 47km?), then go to Srf+ 2.
It'll do a bit of a descent (to <40km), return to Srf+ when it wants to climb again.
Shutdown Nerv when Apo = 95km min. I shut down at Apo = 98km, and the wings lofted it to 105km.

I'm pretty sure you could SSTO with 3 Rapiers + 2 Nervs. Maybe a Whiplash/Panther and 2 Rapiers. I tried to make a 1 Rapier + 2 Nerv plane work, and it would not.

As a general rule, I think optimum TWR is dependent on L/D, but more importantly, Thrust/Drag. T/D is dependent on altitude. The zone between the top of the speed run and about 37km is worth getting out of quickly.

I believe jet-mode Rapiers will always be superior choices due to the fact that they're they ONLY engine than can traverse a certain region at such a high ISP and reasonable TWR. Basically, you're running on ion engines for a 3-5km band. I'd take them even if they didn't have closed cycle.

What's the optimum L/D pitch for a wing? In the old Aero, it was around 26deg (which is ridiculous)

https://www.dropbox.com/s/w2i9uk93c99nqv0/AeroGav 2STO v2.craft?dl=0

[AeroGav]

3 hours ago, FleshJeb said:

What's the optimum L/D pitch for a wing? In the old Aero, it was around 26deg (which is ridiculous)

About 5 degrees when supersonic,  so if you tap ALT and one of the direction keys (can't remember which) when attaching wings it gives you that much as an incidence angle.    As Editor Extensions no longer works in the current KSP  version this is how i do it.

However, 3-8 degrees is not much worse.   Fuselage drag dwarfs wing drag, so the most important thing is to have a fuselage AoA close to zero (but preferably nose still above prograde, slightly) when on NERV power.     I used the stability analysis in CorrectCoL to balance my plane so it flies *slightly* nose up without control input.   RCS build aid to correct for thrust torque and fuel burnoff.    ACtion groups 1,2 and 3 deploy trim flaps that raise and lower the nose a degree or two.

The reason my craft has so much wing is because it lowers fuselage drag.   Eg.   Mach 6,  body on prograde, wings on 5 degrees due to incidence.    Your design might be flying at 33km,  mine may be 42km because it has twice as much wing relative to its weight.      The drag from the wings from both designs will be the same, since we weigh the same and our wings produce the same lift drag ratio (we're both at same AoA and mach number). But my fuselage will make less drag at mach 6 because it is at higher altitude than yours.    That said, you're using oxidizer to achieve much the same effect. 

I'd not have used any fuselage tanks at all, but i needed them to block the heat soak from the nose cone/main stack nuke.  If not for that, i'd have just enlarged the wings till all the fuel was in there.   Well, I think i also needed a trim tank up front too.   Just found out the 1.25m reaction wheels are quite draggy, but not sure if there's much alternative.    Perhaps the side stacks can just have ncs adapters instead of mk1 tanks to mount the nukes with,  get their fuel put in an enlarged wing instead.

I'm considering a 3 stage version for RSS scale.   Two whiplash, then two nukes and a terrier and extra tanks.   When they are dumped, it's just two nukes.