1.05 Intakes - Lets figure them out

[Right]

So intakes are behaving differently. I set up a test rig with all the intakes to map out their performance. Here are my findings:

All measurements taken at sea level, 0° AoA +/-0.02°, all intakes simultaneously.

We can see four type of curves:

• 1. Maximum at mach 1.5 (Engine Nacelle, Circular Intake, Small Circular Intake, XM-G50 Radial Air Intake)
  
• 2. Maximum at mach 2.5 (Mk1 Diverterless Supersonic Intake)
  
• 3. Maximum at mach 3 (Adjustable Ramp Intake, Adjustable Ramp Intake (Radial), Engine-precooler)
  
• 4. No maximum (Shock Cone)
  

Next step will be to divide each intake by its drag values to see what has the best performance for the least drag.

https://drive.google.com/file/d/0BwgWaXQEa1uwb0lQYV9HSkZTQjA/view?usp=sharing

/Discuss

[paul23]

A hint: Don't ever smooth a curve. It adds maxima that you are not sure off.

Anyways, this would mean the shock cone is simply the best in all cases?

[LaytheDragon]

If I remember correctly, the Goliath has a built-in intake too, which you forgot in the chart.

[GoSlash27]

Right,

I'm glad you've started tackling this. This is the root of all the mystery drag problems people are running into in 1.05 IMO.

Rather than looking at which intake can get the most air for the least drag, we need to look at which intake setup can keep an engine lit throughout it's envelope for the least drag.

My early testing confirms that any air breather up to and including the RAPIER can be supplied adequately with just 0.001 m^2 effective intake area. I have run all of the spaceplane engines up to their operational limits using just 2 structural intakes, and this is *much* less than what was necessary in 1.04.

My testbed Panther hybrid SSTO is successful using a plain old nacelle, and that's not a particularly good intake.

The Shock cone is technically the second-most drag effective intake, but it's advantage is wasted if you're not driving a large number of engines.

The Mk1 radial "precooler" intake is going to be the best intake available because it feeds an engine and adds absolutely no drag.

Anywho... I look forward to discovering the new wrinkles of 1.05 and I'm glad we've got your help on this.

Best,

-Slashy

Edited November 12, 2015 by GoSlash27

[Right]

@Paul I've plotted the actual float curve for mach on unity3 and they are smooth in this case...so smoothed curves are more accurate for this.

It is the best for raw intake air. Most intakes outperform it at mach 0 though due to its low suction rate though. If I included the Goliath engine's intake on this graph, it would probably do better than everything up to Mach 1.5-ish.

@LaytheDragon Thats true. I thought about it but it was going to make my test have problems so I left it out. It will perform just like a normal circular intake just with a larger intake area/more flow.

@GoSlash I agree, that will be important to know for single engine set ups. Knowing what produces the most output to drag ratio will give us a good window for starting points. I agree about the shock cone, its going to be overkill when your engines are forced to spool down at high altitudes/speeds.

Edited November 11, 2015 by Right

[GoSlash27]

A hint: Don't ever smooth a curve. It adds maxima that you are not sure off.

Anyways, this would mean the shock cone is simply the best in all cases?

paul23,

Definitely not.

Remember that airflow that isn't used is wasted. You have to look down into the chart and see how much airflow you actually *need*, and then figure out which intake can supply it for the least drag.

Best,

-Slashy

[NathanKell]

Numbers are good

Also GoSlash27, I think you'll find that the areas reported by the intakes are sane now, on the tooltips. If not, I must have screwed up a commit. (I.e. area is now actually in m^2, not just 'units')

[Teilnehmer]

What about the air pressure? Does the picture change at, say, 10 km of altitude, or it is just scaled linearly?

[Alshain]

What about the air pressure? Does the picture change at, say, 10 km of altitude, or it is just scaled linearly?

It's irrelevant for this case because it will be consistent for all intake types, regardless of what altitude you are at. As mentioned in the OP, the testbed was at Sea Level. If you were to climb to say 20km, drag would be less but so would air. However that effect is true of all intakes at 20km.

Edited November 11, 2015 by Alshain

[KerikBalm]

Well, shock cones used to be the lowest drag type of nosecone one could add to a part... better than any of the nosecones... so you should spam them on all open 1.25m nodes... is that still the case?

"Remember that airflow that isn't used is wasted. You have to look down into the chart and see how much airflow you actually *need*, and then figure out which intake can supply it for the least drag."

Well, if the shock cone is the least draggy thing you can put on a node, period, its shock cones all day every day... indeed in light eve ascent vehicle challenges, Shock cones were used despite having no airbreathing engines at all (no breathable air either).

[Octa]

"Remember that airflow that isn't used is wasted. You have to look down into the chart and see how much airflow you actually *need*, and then figure out which intake can supply it for the least drag."

Is that displayed by the info "Intake air provided"/Intake air needed" (both kg/s^2) by Mechjeb?

Can't remember my mach numbers (must be mach 4 or 6), but at 20-24km and 1200-1400km/s, my thrust startet to decrease. 4 Rapiers, 6 Shock-Cones and a few structural radial intakes, because i had no attachment nodes left and needed to counteract a bit on the fact that 2 engines were lacking one shock-cone.

Since SSTOs have high speed at high altitudes, i can't see chances that other intakes perform somewhat better for this.

[Alshain]

Well, shock cones used to be the lowest drag type of nosecone one could add to a part... better than any of the nosecones... so you should spam them on all open 1.25m nodes... is that still the case?

"Remember that airflow that isn't used is wasted. You have to look down into the chart and see how much airflow you actually *need*, and then figure out which intake can supply it for the least drag."

Well, if the shock cone is the least draggy thing you can put on a node, period, its shock cones all day every day... indeed in light eve ascent vehicle challenges, Shock cones were used despite having no airbreathing engines at all (no breathable air either).

I think the precoolers have almost no drag at all.

[Right]

What about the air pressure? Does the picture change at, say, 10 km of altitude, or it is just scaled linearly?

Air intake scales with air pressure. If air pressure is 1/10th, then air intake is 1/10th. This applies to all intakes.

Well, if the shock cone is the least draggy thing you can put on a node, period, its shock cones all day every day... indeed in light eve ascent vehicle challenges, Shock cones were used despite having no airbreathing engines at all (no breathable air either).

Drag, mass, and cost are all concerns. Shock cones represent a very good optimization of drag and mass but with high cost. Those alternatives which have a better drag (like the tail connector A) tend to have a higher mass. This makes them good for SSTO applications, but often not for rockets.

Can't remember my mach numbers (must be mach 4 or 6), but at 20-24km and 1200-1400km/s, my thrust startet to decrease. 4 Rapiers, 6 Shock-Cones and a few structural radial intakes, because i had no attachment nodes left and needed to counteract a bit on the fact that 2 engines were lacking one shock-cone.

Since SSTOs have high speed at high altitudes, i can't see chances that other intakes perform somewhat better for this.

Decreasing thrust does not always indicate a lack of air (in fact, I'm not sure if it ever does). You can spam dozens of intakes and you will still see the thrust decrease at certain speeds and altitudes.

[GoSlash27]

Well, shock cones used to be the lowest drag type of nosecone one could add to a part... better than any of the nosecones... so you should spam them on all open 1.25m nodes... is that still the case?

"Remember that airflow that isn't used is wasted. You have to look down into the chart and see how much airflow you actually *need*, and then figure out which intake can supply it for the least drag."

Well, if the shock cone is the least draggy thing you can put on a node, period, its shock cones all day every day... indeed in light eve ascent vehicle challenges, Shock cones were used despite having no airbreathing engines at all (no breathable air either).

KerikBalm,

Actually, the shock cone is not the least- draggy thing you can put on a node. The tailcones and Mk1 cockpit are much cleaner. The shock cone often gets used because it's lighter than the other options, but for spaceplanes "light" is a much lower priority.

If you want the ultimate low drag intake, the tailcone A + precooler is the way to go. Total frontal drag is less than half that of a shock cone.

I should also point out that you don't necessarily want to adapt your structure just so it'll have a node for a shock cone. Often the additional drag of the adapters offsets any gain from the shock cone itself. For example, you wouldn't want to (and I see this a lot) use an inline Mk2 cockpit preceded by a long Mk2-Mk1 adapter and a shock cone. You'd have much less total drag with a plain old Mk2 cockpit.

Decreasing thrust does not always indicate a lack of air (in fact, I'm not sure if it ever does). You can spam dozens of intakes and you will still see the thrust decrease at certain speeds and altitudes.

Right,

Agreed. Your thrust and airflow are going to taper off with speed and altitude no matter what you do. This is a function of the engine rather than the intake. So long as it doesn't flame out before you max out, you've got enough intake.

example:

1600m/sec at 20 km altitude using just 2 structural intakes. In 1.04 I would've needed 10 of these intakes to do the job.

This sets the upper limit of how much intake you actually need for an engine, and nearly all of them can do the job no problem.

*edit* After further testing, the Mk.1 diverterless intake and the XM-G50 are inadequate to keep a RAPIER fed at Mach 5 by themselves. This should set a lower bound for what's required.

Best,

-Slashy

Edited November 11, 2015 by GoSlash27

[Foxster]

Hasn't the shockcone increased in mass quite a lot in 1.0.5?

[Right]

Yeah, its more than 4x heavier (.0295 vs .13). We'll probably start seeing more of the stubby aerodynamic nose cone and advanced nose cones thanks to their lower mass. (In rocket booster stages for example)

[g00bd0g]

So how do I analyze the number of intakes I need? Let's say I'm cruising at 20k alt on 2 rapiers. By what metrics do I determine how many and what kind of intakes I need? Do I use the "intake air" resource, and verify I need some minimum amount? Do I look at the "flow rate" of the intakes themselves? Just wonder what the actual "workflow" to better engineer my vehicles would be?

Thanks!!!

[A_name]

Right,

I'm glad you've started tackling this. This is the root of all the mystery drag problems people are running into in 1.05 IMO.

Rather than looking at which intake can get the most air for the least drag, we need to look at which intake setup can keep an engine lit throughout it's envelope for the least drag.

My early testing confirms that any air breather up to and including the RAPIER can be supplied adequately with just 0.001 m^2 effective intake area. I have run all of the spaceplane engines up to their operational limits using just 2 structural intakes, and this is *much* less than what was necessary in 1.04.

My testbed Panther hybrid SSTO is successful using a plain old nacelle, and that's not a particularly good intake.

The Shock cone is technically the second-most drag effective intake, but it's advantage is wasted if you're not driving a large number of engines.

The Mk1 radial intake is going to be the best intake available because it feeds an engine and adds absolutely no drag.

Anywho... I look forward to discovering the new wrinkles of 1.05 and I'm glad we've got your help on this.

Best,

-Slashy

Slashy, is there a way to make it so that a single intake will feed multiple engines? I seem to remember this was impossible but maybe it's changed,

- - - Updated - - -

So how do you guys even know how much drag a particular part has? I.e. how can you tell if a shock cone or nose cone will be more draggy.

[NathanKell]

That's not quite correct. Let's quickly review how KSP resources work, and indeed modules on parts.

1. Each frame, all parts are processed in tree order.

2. Each part processes its modules.

3. All resource requests (either pushing or pulling a resource) are first-come-first-served.

Now, consider

2 air intakes

2 jet engines

If the air intakes produce more air than is required by your two jets, and if there's enough buffer space (as of 1.0.5 there should be, so let's ignore that), then everything's fine. Both intakes add 1 air each, then each jet removes 1 air each, and you're all good.

Now consider what happens if each jet still wants 1 air each, but your intakes are only intaking 0.5 air each.

Air starts at 0

Intake 1 adds its 0.5 air. There's now 0.5 air.

Intake 2 adds its 0.5 air. There's now 1.0 air.

Jet 1 pulls its 1 air. There's now 0 air.

Jet 2 tries to pull its 1 air, gets nothing. Oh dear.

Say the intakes produce 0.75 air each. Jet 1 will fully pull its 1 air, but jet 2 will only get 50% of what it asked for.

Now you can see why people mess with order. Let's try interleaving.

Intake 1 produces 0.5 air.

Jet 1 pulls 1 air but gets only 0.5. It runs at 50% thrust.

Intake 2 produces 0.5 air.

Jet 2 pulls 1 air but gets only 0.5. It runs at 50% thrust.

Hopefully that clarifies the whole mythos of "intakes can only feed one engine" thing.

[paul23]

Can we see the order ourselves? - When things get attaches in parallel it's hard to know the order of processing.

[Alshain]

That's not quite correct. Let's quickly review how KSP resources work, and indeed modules on parts.

--Snip--

I thought asymmetrical flameout was fixed several patches back. Are you saying that isn't the case?

[NathanKell]

If you have 'enough' intakes you won't get asymmetric flameouts, because the engines will flame out together due to the flow multiplier being too low, a good bit before the air intakes cease being able to shove enough units at you. If you don't have enough intakes you'll hit that problem above.

However, my noting it publicly ought be taken as indication that I'm aware it's a problem.

[Alshain]

If you have 'enough' intakes you won't get asymmetric flameouts, because the engines will flame out together due to the flow multiplier being too low, a good bit before the air intakes cease being able to shove enough units at you. If you don't have enough intakes you'll hit that problem above.

However, my noting it publicly ought be taken as indication that I'm aware it's a problem.

Oh I see, so asymmetric flame-out if you lack enough air intake but NOT from simply climbing in altitude where there isn't enough air to be in...took?

[322997am]

Also is there a limit where the intakes have no intakeair. Please inform me if you know and if no one knows then I'll run some tests.

[g00bd0g]

What is the new "effective base speed" metric in the part description? I seems like lower is better for higher speeds ops?