Why use the arospike

[herbal space program]

Cuz it's shoart, so your spaceplane won't leave it on the runway trying to take off!

[UmbralRaptor]

I don't think the thrust loss results in a lower ISP. I thought it was just a thrust scaler, and then ISP considers how much fuel is needed to produce that thrust.

Can anyone confirm?

That 0.02738 U appears to be LF+IntakeAir in tonnes per second. o_O

Isp = Thrust/(mass flow * 9.82) = 112.5/(0.02738 * 9.82) ~= 418.4 s

418.4 s is conveniently half of what is displayed, and this is a stock turbojet that would be getting half thrust while sitting on the runway...

(The fuel consumption in 0.34 units/second appears much the same, though mind the 16x from the "free" intake air.)

For clarity: 0.24 unit/s of LF into the above equation:

Isp = 112.5/(0.24*0.005*9.82) ~= 6738.9 s. Divide by 16 (since LF is specified to be 1/16 of the propellants in the part.cfg), and you get ~421.2 s. I'm going to say that's the same to within measurement error.

Edited October 24, 2014 by UmbralRaptor
Clarity of explanations

[numerobis]

Basic equation:

Ve = thrust / mdot

Where Ve is the exhaust speed, i.e. the speed at which molecules come out the tail end; and mdot is the mass flow rate, i.e. the number of molecules going out the tail end. The Isp is exhaust speed divided by g0.

If you halve the thrust without changing mdot, then you've halved the Ve (and thus halved the Isp).

In the game, the Isp that is displayed is used to calculate mdot; it isn't the actual Isp, it's just a lookup in the altitude curve. The fuel flow (mdot) that is displayed is accurate (but note that it takes into account the intake air, not just liquid fuel). The thrust that is displayed is accurate. With the mdot and thrust, you find the actual Isp. To calculate the effective Isp counting only the mass of liquid fuel, divide mdot by 16, which means multiplying Isp by 16.

That means there's a vary narrow envelope where a rocket is more fuel-efficient than a jet. Using the KSP-scripts:

python
import jets
def jetIsp(v, alt):
     mdot = jets.turbojet.fuelRequired(alt)
     T = jets.turbojet.thrust(v) * 1000
     return T / (mdot * 9.82)

Now we can check the Isp of a jet in terms of liquid fuel usage. Let's put it at stupid-high elevation and find the speed we need to go until the jet is less efficient than an LV-N:

>>> jetIsp(2300, 69000)
1500.005501351666

Nope; faster...

>>> jetIsp(2329, 69000)
800.0096340787961

Aha! So if you have a jet going 2329 m/s at 69km altitude, you're doing no better than if you were using an LV-N. At that speed of course, you can easily have an apoapsis above the atmosphere.

In order for the aerospike to beat the jet on fuel efficiency, you need to be going about 2351 m/s. You can't go that fast with jets alone.

[Northstar1989]

@numerobis

Is there a reason you have:

T = jets.turbojet.thrust(v) * 1000

Note the emphasis on the *1000. That would, it appears at first glance, lead to a calculation where the turbojets perform 1000 times better than in KSP.

As for all those talking about the RAPIER vs. Turbojet, remember this- the higher your TWR, the less TIME it will take your spaceplane to break atmosphere and make orbit, and the less time you spend fighting gravity. When you start your burn-to-orbit, you can switch your decativated RAPIER engines to rocket mode to get a LOT more thrust than if you left them inactive or used trubojets (remember, near altitude ceiling you can only run one jet at full thrust without intake spam...)

I'm also going to completely ignore all this bit about getting an apoapsis outside the atmosphere with just turbojets, because that's not actually possible without either (a) SERIOUS intake-spam (by the way, if you have 9 intakes and 3 engines, it's still intake spam if you shut down 2 of the engines but none of the intakes) or ( an EXTREMELY small payload relative to the thrust, in which case what you have isn't a plane anymore (it probably can't even land like a plane), it's an air-breathing rocket...

Jet engines are OP'd enough in KSP as is (due to ridiculous TWR values, the improper inclusion of IntakeAir in ISP calculations, the lack of compression-heating, and overly-generous speed and altitude curves), please stop abusing them further by hooking up 12 or 15 intakes to a single jet engine. I'd play with Advanced Jet Engines and laugh at you all, except that mod goes too far by nerfing most of the stock jet engines down to 1970's tech-levels (excepting the RAPIER, the only way to get a MODERN jet engine level of performance with AJE is to install a mod like B9 Aerospace- which AJE assumes to have modern engine performance levels for many of the jets... I wish it would just institute "tech levels" like the "Stockalike" RealFuels engine config.)

Regards,

Northstar

Edited October 25, 2014 by Northstar1989

[Red Iron Crown]

Single intake per jet can get the apoapsis out of atmosphere, it's just a bit more difficult than with more intakes. Beyond a certain point the TWR doesn't really matter anymore, the craft is already coasting to its circularization burn and only needs enough thrust to overcome drag, so why not use the highly efficient jets at low throttle rather than eating up rocket fuel?

I won't stop "abusing" them until the game mechanics change, until then I'll just cope with the laughter.

[Jouni]

I'm also going to completely ignore all this bit about getting an apoapsis outside the atmosphere with just turbojets, because that's not actually possible without either (a) SERIOUS intake-spam (by the way, if you have 9 intakes and 3 engines, it's still intake spam if you shut down 2 of the engines but none of the intakes) or ( an EXTREMELY small payload relative to the thrust, in which case what you have isn't a plane anymore (it probably can't even land like a plane), it's an air-breathing rocket...

Two comments:

1) It's easy to get the apoapsis outside the atmosphere with just one ram intake, if the craft is small enough. The real measure of intake spam is intakes/tonne, because the number of engines is unimportant at high altitudes. During the interesting part of the ascent, thrust is proportional to intake air, while drag is proportional to mass. As long as you can generate enough thrust to overcome drag, you keep climbing and accelerating.

2) Air-breathing rockets can have payload fractions over 70%. Assuming TWR 1.05 at liftoff, each turbojet can lift 9.72 tonnes in addition to the engine itself, out of which around 8 tonnes can be payload.

[Northstar1989]

1) It's easy to get the apoapsis outside the atmosphere with just one ram intake, if the craft is small enough. The real measure of intake spam is intakes/tonne, because the number of engines is unimportant at high altitudes. During the interesting part of the ascent, thrust is proportional to intake air, while drag is proportional to mass. As long as you can generate enough thrust to overcome drag, you keep climbing and accelerating.

I assume you're playing with the horrible stock aero instead of FAR? (or even NEAR- which isn't half as good, as it cuts a lot of good stuff like body-lift) Because what you just said is COMPLETELY invalid when playing with FAR, where drag is related to shape and size, not mass.

2) Air-breathing rockets can have payload fractions over 70%. Assuming TWR 1.05 at liftoff, each turbojet can lift 9.72 tonnes in addition to the engine itself, out of which around 8 tonnes can be payload.

You completely missed my point... It's not a plane if it's an air-breathing rocket. IF you just have a tiny, tiny payload, and intake, and a jet engine, with almost no wing area, what you basically have is an air-breathing rocket, not a plane.

The measure of a plane is if you cut the thrust, would it drop out of the sky like a brick? If the answer to that question is "Yes", then it's a rocket, not a plane (although even some wingless rockets can have decent gliding capabilities in FAR due to body-lift...)

And air-breathing rockets that can make orbit are COMPLETELY unrealistic (unless you're talking air *augmented*, rather than air-breathing, rockets), as what basically equates to their velocity curves (tough there is no such thing in real life) completely disallow it. Currently, jets have far too generous of a velocity curve, and orbital velocity is far too low (Kerbin has too small of a radius), which adds up to jets being able to take you almost all the way to orbit with a rocket- which is completely unrealistic. Also, the jets have far too high of a TWR- many jets can barely lift their own weight in real life... (even military fighter jets barely have a TWR of 1 at optimal altitude AND speed)

Regards,

Northstar

Edited October 25, 2014 by Northstar1989

[KerikBalm]

Interesting, so the thrust scalar effectively scales ISP as well, I did not know that!

I had begun to suspect something like that with my fuel consumption rates in NEAR/FAR, when my jets produce very little thrust at ~1,500 m/s, but I thought it was specific to the mod.

I think that *1000 is to convery kN to N

Can you do the calculations as to when a 390 ISP engine is more efficient that Turbojets in FAR? when the LV-N is more efficient in NEAR/FAR?

Assuming 16x1200 ISP (the rated ISP for Turbos at very high altitude, ISP won't change much between 30km and 60km) for the turbos, and 800 ISP for the LV-N (again, I'm assuming one gets nearly vacuum numbers at 30km) or 390 for the aerospike.

Jets get 19200 ISP at their peak thrust levels. (1000 m/s stock, 900 for FAR, 1100 on the rapier)

390/19200= 0.02

800/1920= 0.042

So... at what speed will turbotjets or rapiers give 2% or 4% thrust , given these velocity curves:

Turbojet

key = 900 1 0 0

key = 1800 0 0 0

Rapier

key = 1100 1 0 0

key = 1700 0 0 0

Do I do linear interpolation, or is it logarithmic or something?

[kujuman]

Also, the jets have far too high of a TWR- many jets can barely lift their own weight in real life... (even military fighter jets barely have a TWR of 1 at optimal altitude AND speed)

That's not really evidence of the engines being OP. Since we don't have the kind of machines people build in KSP, we can't compare them against real life machines. F-15s and Su-27s aren't designed to do the spaceplane thing so we can't compare them to KSP spaceplanes: just as we can't really compare a moon lander with a heavy lift rocket.

The engines themselves in KSP have a TWR in the low teens at reasonable speeds (so we ignore any ram effects). This isn't too far from high performance engines we have for conventional use. But the lack of existence of a such an engine is evidence that such an engine is not being built, not that it can't be built.

So... at what speed will turbotjets or rapiers give 2% or 4% thrust , given these velocity curves:

Do I do linear interpolation, or is it logarithmic or something?

My understanding is that the floatCurve KSP builds from those keys is a Unity AnimationCurve, which is a Bézier curve. If the keys didn't have the last two 0s on them they would be linear. But the two 0s change the slope at the extremes making it a bit more complicated, assuming I am fully understanding of how unity is interpreting the curves.

Edit: Played around in unity...

A turbojet is about 2% at 1725 and 4% at 1690

Rapier is about 2% at 1650 and 4% at 1630

using UnityEngine;
using System.Collections;


public class NewBehaviourScript : MonoBehaviour {
    AnimationCurve aC;
    float time = 0;
	// Use this for initialization
	void Start () {
        float time = 900;


        Keyframe[] keys = new Keyframe[2];
        keys[0] = new Keyframe(1100, 1, 0, 0);
        keys[1] = new Keyframe(1700, 0, 0, 0);


        AnimationCurve aC = new AnimationCurve(keys);


        do
        {
            time += 10f;
            Debug.Log("Key at " + time.ToString() + " : " + aC.Evaluate(time));
        } while (time < 1800);
	}
}

Edited October 25, 2014 by kujuman

[Jouni]

I assume you're playing with the horrible stock aero instead of FAR? (or even NEAR- which isn't half as good, as it cuts a lot of good stuff like body-lift) Because what you just said is COMPLETELY invalid when playing with FAR, where drag is related to shape and size, not mass.

I'm playing one-off challenges with different mod combinations. Sometimes I use FAR, sometimes not. This is a KSP forum, however, so all discussions are about the stock game, unless explicitly declared otherwise. KSP+FAR is not KSP.

[numerobis]

@numerobis

Is there a reason you have:

T = jets.turbojet.thrust(v) * 1000

fuelRequired gives kg/s whereas thrust gives kN. Without that 1000 factor, you get an Isp in ks rather than s.

I'll have to complain to the author of the scripts for not being consistent. Self a year ago, you're an idiot.

[numerobis]

That's not really evidence of the engines being OP.

The Isp is close but a bit high compared to modern jets. The TWR is very high compared to modern jets, and they can operate at much higher speeds.

About TWR: the SR-71 engine and the engine on the A380 each get about 55 N per kg of engine mass, whereas the KSP basic jet gets 150 N/kg and the turbojet gets 187.5 N/kg.

The fastest highest jet ever in production on Earth is, far as I know, the SR-71. It only got to a bit over 1000m/s and the equivalent of 13,500m altitude on Kerbin (the scale height of Earth is about 9km whereas Kerbin is 5km, so you need to multiply Earth altitudes by 5/9 to get equivalent Kerbin altitude).

The Scaled Composites plane only gets to the equivalent of 8,500m to drop its rocketship payload.

So the jet aircraft we regularly build in KSP have much higher performance than anything we build on Earth, and a large part of that is that the engines have much better performance envelopes than what we can build on Earth.

My understanding is that the floatCurve KSP builds from those keys is a Unity AnimationCurve, which is a Bézier curve. If the keys didn't have the last two 0s on them they would be linear. But the two 0s change the slope at the extremes making it a bit more complicated, assuming I am fully understanding of how unity is interpreting the curves.

The KSP-scripts AnimationCurve reproduces the Unity curve accurately. I got that done with help from K^2 and a random link on the internet that is totally unsourced (but it works).

[Red Iron Crown]

The Isp is close but a bit high compared to modern jets. The TWR is very high compared to modern jets, and they can operate at much higher speeds.

The Isp is close to real world values, but it is applied incorrectly. It includes the mass of the intake air in the thrust calculation, which yields approximately 15x the thrust for a given fuel mass-flow rate.

[numerobis]

You made me look. The basic jet gets about 17ks Isp at cruise (300m/s and 10km) according to my script above. Wikipedia claims a representative 1970s-era turbofan gets 11.7 ks Isp. Engines have improved quite a bit since then, so the basic jet isn't ridiculously off on Isp.

The turbojet gets slightly better performance on most of its envelope, whereas real turbojets get much worse performance (that's one of the major reasons airlines use turbofans instead). So the turbojet is pretty far off.

[numerobis]

Anyway, this is getting OT, since none of this has anything to do with the aerospike.

Why use the aerospike? Because it's there. It's attractive. It's short, which is handy in some circumstances. It makes sense at the story level, even if KSP physics is biased against it.

[Red Iron Crown]

You made me look. The basic jet gets about 17ks Isp at cruise (300m/s and 10km) according to my script above. Wikipedia claims a representative 1970s-era turbofan gets 11.7 ks Isp. Engines have improved quite a bit since then, so the basic jet isn't ridiculously off on Isp.

The turbojet gets slightly better performance on most of its envelope, whereas real turbojets get much worse performance (that's one of the major reasons airlines use turbofans instead). So the turbojet is pretty far off.

Your link doesn't support your assertion, it gives a typical turbofan specific impulse of 3000m/s.

Edit: Whoops, missed that part of the chart where I was on mobile. Let me think about that a bit.

[Laie]

You made me look. The basic jet gets about 17ks Isp at cruise (300m/s and 10km) according to my script above. Wikipedia claims a representative 1970s-era turbofan gets 11.7 ks Isp. Engines have improved quite a bit since then, so the basic jet isn't ridiculously off on Isp.

Hold on a second (derailing this thread more and more): real-life turbofans are for (comparatively) low speeds and altitudes, where an engine can easily get more air than it needs to burn. The air that gets into the combustion chamber drives a turbine, which in turn drives a fan, which moves a lot of air. The turbine also creates some thrust, but that's mostly a side effect. Most of the propulsion comes from the fan (really, not all that different from an oldey-timey propeller).

KSP "TurboJets" are misnamed: given their mission profile and stats (more output at high altitudes and speeds), they should be ramjets. Keep that in mind when looking up real-life stats for comparison. IIRC anything with "fan" in it's name has a much better ISP than the "ram"-whatevers.

[Jouni]

You made me look. The basic jet gets about 17ks Isp at cruise (300m/s and 10km) according to my script above. Wikipedia claims a representative 1970s-era turbofan gets 11.7 ks Isp. Engines have improved quite a bit since then, so the basic jet isn't ridiculously off on Isp.

The CF6 turbofan gets 11.7 ks at sea level and 5950 s at cruise. Modern turbofans are 10-15% more fuel efficient. If the difference between 6-7 ks and 17 ks isn't too bad, then perhaps KSP rocket engines should also have their specific impulses increased to 700-1100 s.

[numerobis]

The CF6 turbofan gets 11.7 ks at sea level and 5950 s at cruise. Modern turbofans are 10-15% more fuel efficient. If the difference between 6-7 ks and 17 ks isn't too bad, then perhaps KSP rocket engines should also have their specific impulses increased to 700-1100 s.

So the basic jet is 2-3x the Isp it should be, I think is what you're saying?

[Jouni]

So the basic jet is 2-3x the Isp it should be, I think is what you're saying?

Yes, based on that single data point. The performance curves of real turbofans and KSP basic jets are probably quite different.

[KerikBalm]

So the basic jet is 2-3x the Isp it should be, I think is what you're saying?

2-3 * 16 times the ISP it should be.

Intake air should either be treated as massless for the ISP equations (like electric charge), or the ISP rating should be nerfed to 1/16th its value, due to the 15:1 mass ratio of intake air to jetfuel being consumed as reaction mass

[numerobis]

We already took account of that when calculating the KSP jets' Isp values.

[Hodo]

I use aerospikes for a number of things. When I need a higher thrust VTOL rocket engine for some of my SSTO space planes that have VTOL abilities. I also use them for some of my landers because of their higher ISP and decent thrust rating.

I use them quite a bit in my RSS install, and even have the linier aerospike in that install for some of my space plane attempts.

I love aerospikes for their dozen or so uses with anything. I think they are used almost as much as my LV-1Ns and SABRE engines.

[KerikBalm]

We already took account of that when calculating the KSP jets' Isp values.

The problem is the game's ISP value takes it into account twice....

If it treats intake air as part of the propellant, then it should have a much lower ISP value.

You'd only use the game's ISP values to get a close to realistic, if intake air was considered massless as far as the game's physics were concerned.

[numerobis]

The numbers I give are based on this code, which calculates Isp the way you want it to be calculated: figure out the mass flow, and take just the liquid fuel fraction of it.

The conclusion is that the KSP basic jet is 2-3x the Isp of modern Earth turbofans (my initial comparison had bad numbers for Earth jets). The KSP turbojet is yet more efficient at high speed, whereas Earth turbojets are less efficient than turbofans, so that's totally out of whack.