Pre-cooler vs. Shock Cone

[Raideur Ng]

I know this question has been asked time and again, but I have dug deep and yet to find a clear cut answer to the question:

Are pre-coolers capped with an aerodynamic nose (advanced nose or tail connector) less draggy than the same amount of shock cones per intake area?

That seems to be the two greatest issues is intake area vs drag on any high speed vehicle. Testing the same vehicle with the same weight, does intake area even count towards drag on the pre-coolers that are streamlined? I guess what I'm asking is how the heck drag is calculated, especially if you can't directly tell from the debug part menu? This is the single most difficult issue preventing an SSTO to the outer planets is the drag on Kerbin.

Edited February 20, 2016 by Raideur Ng

[GoSlash27]

Raideur NG,

 I suppose the precooler is lower drag per intake area, but the important question is whether it's lower drag in any practical installation in sufficient quantity to feed your engines.

 Happily, the answer here is "yes" also. The precooler is always lower drag than the shock cone when preceded by a tailcone or a cockpit. In fact, in most cases the radial VRI intakes also make for a cleaner installation than the shock cone under these same circumstances.

 They will be cleaner, but not necessarily lighter or more heat tolerant. All of these factors need to be taken into account, but for spaceplanes low drag is usually the #1 most important consideration.

Best,
-Slashy

 

Edited February 20, 2016 by GoSlash27

[GoSlash27]

A more in- depth review of how it all works:

 If you look in the .cfg files for the intakes, you will see their intake area and mach curve for each intake.

 For example, the VRI radial intake:

area = 0.001
        intakeSpeed = 10
        intakeTransformName = Intake
        machCurve
        {
            key = 0 0.85 0 0
            key = 1 1 0 0
            key = 2 0.95 -0.08751557 -0.08751557
            key = 4 0.5 -0.4034287 -0.4034287
            key = 8 0.01 0 0

This means that the intake has .001 area and at Mach 4, it's effectiveness is at 50%. Each intake has a unique mach curve and you need to know roughly how much *effective* area it has at your desired speed.

 A good rule of thumb is that a Panther or RAPIER will require about .0015 effective intake area, while a Whiplash will need about .002. This means that you would need 3 of these intakes to feed a Panther or RAPIER, 4 to feed a Whiplash.

Extrapolating, a precooler is sufficient to feed a Panther or RAPIER, but you'd need to add a radial VRI to feed a Whiplash. The shock cone has an insanely high area. A single shock cone could feed 3 or 4 engines all by itself (depending on type), but you get no added benefit from excess airflow; it's either enough or not enough.

Moving into the "drag" side of it, that info can be found in the drag cubes in the part database.cfg. An example:

    url = Squad/Parts/Aero/circularIntake/intakeShockCone/shockConeIntake
    DRAG_CUBE
    {
        Default = Default, 0.7486105,0.6887614,0.7221569, 0.7486105,0.6887677,0.7221569, 1.213026,0.3,1.183155, 1.213026,1,0.100694, 0.7486105,0.6903304,0.7221569, 0.7486105,0.6871722,0.7221569, 0,-0.08307549,0, 1.25,1.083849,1.25

This means that the part has a frontal area of 1.213m^2 and a drag coefficient of .3. If you multiply them, it gives an equivalent flat plate area of .364 m^2. The idea is to come up with the lowest flat plate area you can while still supplying enough air to feed your engines.

The rules: A smaller part will occlude it's area of a part attached behind it. A radially attached part adds it's flat plate area to the total.

So for example, a tailconeA, precooler, and a radial VRI.

 Tailcone A area 1.213, Cd .132

 Precooler area 1.213, Cd .972

 Radial VRI area .120, Cd .351

 In this case, the tailcone completely occludes the precooler, so the resultant area is 1.213x .132. The radial VRI is hanging out in the breeze, so it's area is added seperately (.120x .351). Total flat plate area is .202.

Contrast this with the shock cone, 1.213x .3; .364. The tail cone/ precooler is a little over half the drag, even with the radial VRI thrown in.

 You can analyze all manner of adapters/ cones/ intakes/ etc in the same manner.

HTHs,
-Slashy

[Val]

To add to what @GoSlash27 said.

Yes, a single Shock Cone can feed 4 Whiplashes. That's what I do on this craft.

8 Whiplash fed by 2 Shock Cones.

The only catch is that it has to start at half throttle until speed is above 15 m/s. Then it can go full.

If it's started with full throttle from standstill, it gets asymmetric flameouts and veer off the runway before it gets up to 15 m/s.

[Raideur Ng]

By that logic, with it entirely occluded, you can stack precoolers and rack up just as much intake with no drag penalty besides what is stacked forward of it. If so, why wouldnt EVERYONE use stacked precoolers in occluded areas like just behind engines and have highly aerodynamic cones instead of shock cones, besides having heat resistance, which you can get just as much from the aerodynamic nose or the shielded docking port.

There has to be some kind of catch...

 

Additionally, two questions for Val, why would you ever use normal wings when you can use ones that add fuel and why are the engines not all lined up and your CoM adjusted so the torque is nearly zero?

Edited February 20, 2016 by Raideur Ng

[GoSlash27]

33 minutes ago, Raideur Ng said:

By that logic, with it entirely occluded, you can stack precoolers and rack up just as much intake with no drag penalty besides what is stacked forward of it. If so, why wouldnt EVERYONE use stacked precoolers in occluded areas like just behind engines and have highly aerodynamic cones instead of shock cones, besides having heat resistance, which you can get just as much from the aerodynamic nose or the shielded docking port.

There has to be some kind of catch...

 Raideur NG,

 I often use the radial VRIs over the precooler because stacked precoolers make for a long and structurally flimsy fuselage. Plus, it's very rare that you're going to have multiple engines stacked behind a 1.25m node.

*edit* and the advanced nose cones/ shielded ports erase the advantage of precoolers and radial VRIs. The shock cone has a lower Cd than these parts.

 Mainly... people don't have a good working sense of how the drag accumulates, so word never really got out about how superior the tail cone/ precooler system is for spaceplanes. If you look at the super- efficient designs, you won't see the shock cone used very often, except where you have 3 or 4 engines in a single stack (like Val's example).

 I take the blame for this personally. I worked it out shortly after 1.05 dropped, but I was focused on stock career progression and super- cheap disposable lifters at the time and never did a write- up on the subject.

I haven't bothered since, as 1.1 is liable to drop soon and upset everything again. But in the meantime, feel free to take advantage of this info. It's very powerful mojo for spaceplanes.

Best,
-Slashy

Edited February 20, 2016 by GoSlash27

[Raideur Ng]

Diabolical. One further question is if your ship is underpowered and you fly at more than a few degrees AoA, that basically invalidates everything concerning your saved drag, does it not?

[Val]

15 minutes ago, Raideur Ng said:

why wouldnt EVERYONE use stacked precoolers in occluded areas

Nothing is stopping that. It does take up space, though, and part count.

17 minutes ago, Raideur Ng said:

Additionally, two questions for Val, why would you ever use normal wings when you can use ones that add fuel and why are the engines not all lined up and your CoM adjusted so the torque is nearly zero?

I only use the wet wings when I can get away with using them as separate wings. I don't like how it looks if multiple are used together, which I would have needed for the big compound deltawing look I was going for on that craft.

The Whiplashes are pointing outwards because there's a Skipper clipped in between them, and I didn't want it to clip the nozzles and engine bell too much.

They are also thrusting very slightly above CoM, to counter aerodynamic forces during ascent. The net torque should be very small.

The Skipper is aligned with CoM, though.

 

[Val]

14 minutes ago, GoSlash27 said:

...so word never really got out about how superior the tail cone/ precooler system is for spaceplanes.
...
I take the blame for this personally. I worked it out shortly after 1.05 dropped, but I was focused on stock career progression and super- cheap disposable lifters at the time and never did a write- up on the subject.

I remember you posting this when 1.0.5 was new. It does work very well, but it also takes up a lot of space, which is why I haven't used it often.

[GoSlash27]

23 minutes ago, Raideur Ng said:

Diabolical. One further question is if your ship is underpowered and you fly at more than a few degrees AoA, that basically invalidates everything concerning your saved drag, does it not?

Yes. This is why wing incidence is so critical for spaceplanes, especially right around 360 m/sec. I don't take the blame for that not being common knowledge; Val was the progenitor of that critical bit of info

 See what I did there, Val? Chucked ya right under the bus!

Seriously, though... that's the whole key to efficient SSTO spaceplanes. Low front plate area and zero attitude in the transsonic region. If you sort that out, you're into "SSTO ninja" territory.

Best,
-Slashy

 

Edited February 20, 2016 by GoSlash27

[Raideur Ng]

I already have. The issue is finding the sweet spot between having enough thrust that you can climb with zero body AoA and not too much that you overhead, which seems to be between .75-1.0 TWR. Also, massively huge vehicles Val has created are not practical at all. One !@#$ up during a manual ascent and the entire fuselage will come apart. Can a stock super heavy lifter be built that doesnt naturally want to tear itself apart during any control inputs?
 Sorry to pick your brains but trial and mid air explodely painful errors later, I can only gather so much information.

[GoSlash27]

19 minutes ago, Raideur Ng said:

I already have. The issue is finding the sweet spot between having enough thrust that you can climb with zero body AoA and not too much that you overhead, which seems to be between .75-1.0 TWR. Also, massively huge vehicles Val has created are not practical at all. One !@#$ up during a manual ascent and the entire fuselage will come apart. Can a stock super heavy lifter be built that doesnt naturally want to tear itself apart during any control inputs?
 Sorry to pick your brains but trial and mid air explodely painful errors later, I can only gather so much information.

Well... efficient spaceplanes can be built with much lower t/w than that. Most of my designs are in the neighborhood of .3- .5 t/w in air breathing mode. The important part isn't t/w, but rather t/d. My "fringe" designs are in the neighborhood of 1 engine per 40t of spaceplane.

For the Panther and RAPIER, so long as it clears Mach 1.5, it's all downhill from there. The Whiplash is a little different; it prefers operating in an overpowered state because the thrust falls off so sharply with speed. Max efficiency with those engines requires a higher t/w, with some throttling back during the supersonic climb to avoid going all 'splodey.

 As for structural stresses, I (infamously) use tiny amounts of control authority in my spaceplane designs. I occasionally use no control surfaces whatsoever, instead relying on CoG stability and static balance and reaction wheel torque.

It's not really an issue if the plane naturally wants to fly where it's already headed anyway

 For gimongous designs, the limit isn't how much mass you can get to orbit (that's pretty much limitless so long as your CPU can handle it), but rather how big a plane you can safely land at KSC at the end of the mission.

Best,
-Slashy

 

Edited February 20, 2016 by GoSlash27

[Raideur Ng]

I just delivered 60T to a 100 orbit with a TWR of 0.61 - 210/118 Wet/Dry.  The AoA is still too much for my likings so trying to scale it further will be problematic. No Jettison spaceplane is the -only- way I deliver anything to orbit and beyond in KSP because it seems to be hard / fun mode. I don't think reducing AoA is a lift issue, or a matter of balance, but just TWR. Anything under 0.5 TWR can barely clear the ocean and it drags the tail horrid during the entire flight. That HAS to be avoided otherwise its wasted fuel.

I also prefer 1.0 TWR smaller rescue vehicles that are far more forgiving on ascent / descent. What is your opinion on practicality (one control input doesn't end you during climb / re-entry) vs efficiency? Having the vehicle not fight me the entire way is a good feature for vehicles, which requires decent thrust.

[GoSlash27]

8 hours ago, Raideur Ng said:

I just delivered 60T to a 100 orbit with a TWR of 0.61 - 210/118 Wet/Dry.  The AoA is still too much for my likings so trying to scale it further will be problematic.

I guess I'm not following. If you have sufficient wing area and incidence, you will have zero AoA at Mach 1.5. AoA at lower speeds will use up some fuel, but it's just jet fuel at 4000s Isp so you can afford to waste it.

8 hours ago, Raideur Ng said:

What is your opinion on practicality (one control input doesn't end you during climb / re-entry) vs efficiency? Having the vehicle not fight me the entire way is a good feature for vehicles, which requires decent thrust.

 Having the vehicle fly well has nothing to do with thrust. That's just a matter of balance and using minimal control surfaces.

 

[GoSlash27]

Better yet... pictures.
 

Spoiler

 

 

 These principles are infinitely scaleable so long as your wings don't become too noodly and your CPU holds up. Efficiency jumps dramatically when you start using RAPIERs.

Best,
-Slashy

[GoSlash27]

Applying the same principles to small scale crew shuttles (or rescue vehicles)...
 

Spoiler

 

Tech level 6 crew shuttle

A tech level 6 shuttle could be scaled down even further to act as a pure rescue vehicle by using a single Panther and twin LV-909s. Something like this...

 Except without the crew compartments and docking collar.

 

Best,
-Slashy

[sal_vager]

17 hours ago, Raideur Ng said:

does intake area even count towards drag on the pre-coolers that are streamlined?

I can answer this, no, it doesn't.

Intake area only effects the amount of air that can be drawn in at a given speed and air pressure, the larger your area the more air you can get, if there is enough to begin with.

Drag is controlled by the parts drag cube, parts will have different sized front, side and rear drag cube surface areas depending on their size and shape.

Parasitic (skin) drag is a thing, so longer and larger fuel tanks will have more skin drag than short skinny tanks, even precoolers behind nosecones will contribute to drag, though it's a lot less than a nosecone.

Different nosecones (and intakes) also have different drag depending on the sizes of their drag cube faces, the tailcone is so long and has such a small rear face that it's pretty much all skin drag, which is why it's so good for this.

Notice that there is actually drag on the precooler, it's just lower than the other parts..

[Starhawk]

2 hours ago, GoSlash27 said:

Better yet... pictures.
 

 These principles are infinitely scaleable so long as your wings don't become too noodly and your CPU holds up. Efficiency jumps dramatically when you start using RAPIERs.

I must have missed an important part of the incidence discussion.

Can you pleas explain the negative incidence on the rear control surfaces.

Thanks in advance.

Happy landings!

 

[GoSlash27]

Sal,

 I've never looked at the drag directly like that. Pretty cool!

 Yeah, skin drag and tail drag are still a thing, but they fall away dramatically after you hit Mach 1, so pretty much all of the effective drag is due to front plate area and induced drag from the wings.

Starhawk,

 The negative incidence on the rear control surfaces is to adjust the craft to neutral trim around Mach 1. This allows you to have full control authority from weak control surfaces. It also gives you more pitch authority during takeoff, reentry, and landing. These regimes can be difficult to handle with low control authority, but it makes takeoff and climbout much smoother (less bobbling, so less drag).

TL/DR...

 It's to help make the aircraft stable and flyable in all flight regimes.

Best,
-Slashy

[Raideur Ng]

Having tested pre-coolers vs shock cones, even with the SAME amount of intake area, an engine will flame out at a lower altitude than with shock cone-fed air. Two shock cones vs 3 pre-coolers SHOULD be the same intake air, but it very obviously is not. Why.... I have no idea. The pre-cooler also has "5.0" vs "2.0" air amount, whatever those numbers even mean.

The drag conversation was very informative, but getting an extra few thousand feet of jet powered flight is also very critical, since you have a max speed before you simply cannot go faster, so you hit a 'wall' of sorts around about 1500 m/s that no sane vehicle can really pass. That's also with absurd TWRs. Getting more altitude may be the best bet?

 

EDIT: Disregard. Exact weight and dimensions with 3 coolers vs 2 Rapiers gives the same peak altitude performance. Not sure if you cut the numbers in half, since you cant cut intakes in half, I will add more engines and report back. The numbers associated with "Air Amount" are still a mystery.

Edited March 6, 2016 by Raideur Ng

[Nich]

Doesn't airflow for precoolers and ramp intakes drop off at some point?  Where as for shock cones it remains constant.

Also I am not a big fan of tail cones/precoolers as they add a lot of weight.  When you spend 75% of your fuel in vacuum every kg you can shave off is dv for getting to orbit.  Finally tail cone has a speed limit of about 1350 below 20km and 1800 below 30km otherwise it will blow up.  My most efficient SSTOs tend to get 1590-1620m/s off air breathing and hohmann transfer from 30km (2450ish) to AP.  This requires a shielded docking port on the front of the craft.