[1.3.1] Ferram Aerospace Research: v0.15.9.1 "Liepmann" 4/2/18

[SymbolicFrank]

All the instruments fit into a single service bay, so that is cool:

And it does look a lot better. But the center of gravity is probably still too high, even with the increased fuel at the bottom. I'll look into the Infernal Robotics. And a mod for better lights.

[AndreyATGB]

You have a lot of small objects sticking out of the craft, this creates turbulence (I forgot the technical word for it) behind them which can make the rocket unstable (usually it's fine but you have a lot). The new voxel FAR might simulate these effects, so if it's unstable the first thing I'd do is make it smoother.

[Darkway]

The new FAR Colibri has a wave drag area of 0.3923 m^2.

I hope it serves as a good example on techniques to help people handle area ruling using stock parts.

Oh, and since one stationary BasicJet engine is too weak to lift it, it's no longer a VTOL.

I guess I could squish some more out of it (Like getting rid of the ladder on the cockpit ), but I have other things to do, haha.

Oh god, have some mercy on those studying for finals. Seeing this, I want play so badly!

[Shazbot]

It seems to me that the basic jet engine's shape is as terrible as it gets in terms of drag. Any workarounds?

[soulsource]

The point was that the drag was going to be uneven and also would create drag at the top of the craft in the first lunch so that was kinda the point that it would be unstable, aerodynamic efficiency was just a nice bonus.

if you look at the image the craft goes on for what I assume is the orbital injection stage.

Also... it looks more neat that way so win win win.

Don't get me wrong, I fully agree with you. All I wanted to say is that mounting the landing legs on the surface without shielding probably wouldn't be a catastrophe, not that it wouldn't be better to properly shield them.

[Wanderfound]

The new FAR Colibri has a wave drag area of 0.3923 m^2.

I hope it serves as a good example on techniques to help people handle area ruling using stock parts.

Oh, and since one stationary BasicJet engine is too weak to lift it, it's no longer a VTOL.

*slightly outdated image*

http://i.imgur.com/pwrZ2fA.png

I guess I could squish some more out of it (Like getting rid of the ladder on the cockpit ), but I have other things to do, haha.

Just to confirm that I understand [1] the area-rule stuff right: it's about making the green line as close to a straight line as possible? Overall angle of the line less important?

The yellow line is somehow expressing a measurement of the straightness of the green line? What would the yellow line look like if the green was perfectly straight? Also straight, or the high-frequency low-amplitude regular oscillations that it appears to be tending towards here?

[1] Note: full-throttle math-phobic here, despite [2] having a science PhD. If your answer requires any understanding of calculus, I'm unlikely to get it.

[2] Actually in part because of my science training. I have never encountered teaching so incompetently done as it was in the math faculty of the University of Sydney.

Edited May 5, 2015 by Wanderfound

[Darkway]

Just to confirm that I understand [1] the area-rule stuff right: it's about making the green line as close to a straight line as possible? Overall angle of the line less important?

The yellow line is somehow expressing a measurement of the straightness of the green line? What would the yellow line look like if the green was perfectly straight? Also straight, or the high-frequency low-amplitude regular oscillations that it appears to be tending towards here?

[1] Note: full-throttle math-phobic here, despite [2] having a science PhD. If your answer requires any understanding of calculus, I'm unlikely to get it.

[2] Actually in part because of my science training. I have never encountered teaching so incompetently done as it was in the math faculty of the University of Sydney.

My understanding is that the yellow line is the (second derivative?) of the green line, indicating how rapidly the cross section changes (and whether the rate of the change is "accelerating" or "decelerating"). If the green was straight, yellow would be zero all along, right on the gray line.

http://en.wikipedia.org/wiki/Area_rule

The cross section area should change as smoothly as possible.

[cantab]

You can easily go supersonic with the basic jet engine

Does this mean nuFAR will no longer be modifying the jet engine performance like FAR used to? Or is that just something Ferram hasn't looked at yet?

[tetryds]

@Shazbot: Just leave them alone, it's not as bad as it seems.

If you want, you can try to add area to the bottom of it, or half-clip it inside of a nosecone.

But be warned that if you clip it too much or wrongly, it will generate no thrust.

@everybody: Writing a "hopefully understandable" way to explain area ruling right now.

@cantab: nuFAR does not touch jet engines... yet? (turbojet OP)

[Jovus]

Wanderfound, you have that exactly right. If the green line were straight (regardless of slope), which is what you want, the yellow line would be a horizontal line through the origin.

[SymbolicFrank]

You have a lot of small objects sticking out of the craft, this creates turbulence (I forgot the technical word for it) behind them which can make the rocket unstable (usually it's fine but you have a lot). The new voxel FAR might simulate these effects, so if it's unstable the first thing I'd do is make it smoother.

Yes, that's the problem.

Btw, engine fairings (over which you have no control) are problematic:

Edited May 5, 2015 by SymbolicFrank

[tetryds]

Just to confirm that I understand [1] the area-rule stuff right: it's about making the green line as close to a straight line as possible? Overall angle of the line less important?

The yellow line is somehow expressing a measurement of the straightness of the green line? What would the yellow line look like if the green was perfectly straight? Also straight, or the high-frequency low-amplitude regular oscillations that it appears to be tending towards here?

[1] Note: full-throttle math-phobic here, despite [2] having a science PhD. If your answer requires any understanding of calculus, I'm unlikely to get it.

[2] Actually in part because of my science training. I have never encountered teaching so incompetently done as it was in the math faculty of the University of Sydney.

Good that you ask, so you were right about the green line having to be straight, but got confused about the yellow one.

Yes he yellow one is the "straightness" of the green line, and the green one does not really matter, it mostly assists you on knowing where to position stuff.

A perfectly area ruled shape would have the yellow line = 0 for any point, but that is impossible.

If you opened this you certainly know what a derivative is.

Well, the yellow line is the second derivative of the green line.

The area rule dictates that the cross-sectional area must vary smoothly...

...Which means that the second derivative of the cross-section area must be as close to zero as possible (not just that, well, ugh.)

*do not insert math here*

After much thought I came up with simpler ways to make this at least understandable.

These explanations may not be (certainly aren't) 100% accurate, also because they are simplified.

But I hope that they at least help you to not get on a blind fight against lines which you have no idea what they mean:

See that yellow line? You have to make it as close to zero as possible.

The yellow line means that the green line does not change abruptly.

The green line is the cross section area, it's how "fatty" the craft is at that point, no matter where the parts are.

If you add the [Nosecone][Oscar B][...][Oscar B][Nosecone] combo or similar (1.5m tanks, etc) you create bumps on the green line, you make it "fattier" on that part.

Try to add those bumps where your green line drops (or around where it peaks) to smoothen it out.

Do not have straight wings, always have both edges swept inwards or outwards (Blitzableiter is an example plane using straight trailing edges, was a PAIN to get right).

After you do that and find its good enough, open the FAR GUI and move everything around and try to get as small "Mach 1 Wave Drag Area" as possible.

Your plane:

I assume that you already know what the cross-sectional area means, if not check this first: http://en.wikipedia.org/wiki/Area_rule

The very first thing here is that your airplane shape (as of squared corners, or where to place ballasts) does not matter, it matters for other effects, but not for this one.

Area is area, does not matter if it's on the wingtips or hull, and does not even need to have a cross-sectional connection to the main body (yep floating stuff counts too).

Now to what is going on:

I came up with a way to explain this, it may not be exactly what the theory says, but it's a good way to visualize it.

Imagine that the air is heavy, and does not want to change its direction.

So, if your airplane cross-sectional area starts to increase, the air will start moving outwards.

If your area continues to increase, the air will keep going outwards like it was, generating little to no wave drag.

Same applies if your cross-sectional area is decreasing, or is continuous.

If it changes abruptly, the air will have nowhere to go, shockwaves on that section cause drag, a lot of it.

Now you know that what you want is not vary your cross-sectional area abruptly.

Ingame:

The mathematical way to express "continuous variation" is the second derivative, represented by the yellow line.

The cross-sectional area itself is the green line.

So, in a nutshell, the yellow line means how smooth the green line is.

The smoothness of the green line means how much wave drag you have.

If you have some part, like the wings, that adds a lot of area, you will have to balance it out.

The best way to balance it is often adding more area.

Combining the shapes through the airplane to make it so that when your wing starts/ends it does not come from a low area to a big area.

That is why you often add bumps in front and behind of the wings or whichever feature you want.

Tips:

The more you derive a line, the higher the noise, you will notice that single parts have a super massive yellow line bumpiness.

Those are the triangles and imperfections of the shape coming into play, on the top of the approximation the mod uses (not that it's bad, but as I said, shape noise is real).

So, making your craft longer makes it smoother not just because of how the mod does it, nor just because that stretches and separates the quantized sections, but also because you are easing up on the derivative noise.

So, in resume:

I did not explain how the yellow line is actually handled to output the wave drag, but that does not matter.

Go smoooth.

Yellow line -> as close to zero as it can get

Hooray for radially placeable Oscar B tanks.

Hooray for Communotron 16.

Hooray for ferram4.

No worries, ferram4 forced me to make video tutorials about it.

I am just waiting for the derivatives table to get fixed to release the basic one and open my new yt channel and will go straight to area ruling depending on how demanding it is.

Edited May 5, 2015 by tetryds

[Svm420]

Thanks for the explanation tetryds! Still no the edge of my seat for release. The wait is very worth it though. Wish squad had felt that way...

[Van Disaster]

I'm with Wanderfound - I had calculus almost entirely messed up by completely incompetent teaching, but computers are natural differentiators/integrators so if you code at all you can look at basic calc from a different practical angle.

If you loop down the plane taking slices through the cross-section:

* The point on the green line for the current slice offset is the slice size

* Hidden away is the difference between the current slice and the previous slice - call that X

* The yellow line is the difference between X for this slice and the X from the previous slice - this is the rate of change of cross-section.

Visibly, if you clip a goo canister halfway into a cylindrical fuselage, when your slicing reaches the start of the canister the cross section is going to go up very quickly, so the rate of change is fast & you get a bump on the yellow line. At the start of the canister the cross-section is changing fastest so the bump is highest, but as you work down the hemispherical bit of the canister the difference between slices becomes less ( I think? actually how does the area of a demi-hemisphere vary with slices... note to self, do the maths ), so the yellow line gets closer and closer to zero until you get to the cylindrical part.

That doesn't quite cover what I wanted to explain but I'm mentally rather visual, so it's a bit awkward to cook up a simple but comprehensive example :S it doesn't cover how the yellow line can be flat when the cross-section is changing, but I hope that's one of the easier bits to grasp.

Question: should and do half-cylinder inlets on fuselage sides count the entire area of the inlet?

Edited May 5, 2015 by Van Disaster

[jrandom]

I installed the dev build last night and I love the new control surface tweakables.

Say, is there a good rule of thumb for estimating what the wing mass/strength sliders should be set to for any given purpose? I usually wind up with either way too heavy wings or too fragile and since I don't have a mental model of what the range represents from a usage standpoint, I don't know how to... how to "guess" correctly when speccing out a new design.

[Wanderfound]

I installed the dev build last night and I love the new control surface tweakables.

Say, is there a good rule of thumb for estimating what the wing mass/strength sliders should be set to for any given purpose? I usually wind up with either way too heavy wings or too fragile and since I don't have a mental model of what the range represents from a usage standpoint, I don't know how to... how to "guess" correctly when speccing out a new design.

As a very crude rule of thumb:

0.25: very light, very fragile. Eggshell glider, handle with care, especially at low altitude. Roughly equivalent to how all wings were before the mass tweakable was introduced.

0.5: Thoroughly solid, good enough for anything but extreme aerobatics or wing loading. Can be shaved down a bit further for a sensible pilot.

0.8: Extra tough, good for flight trainers, superheavy Mk3 stuff, light aerobatics.

1.0: Fighter strength, unnecessary unless you're intending to pull extreme high-G low altitude aerobatics.

>1: Magic plane with Adamantium wings. You'll probably snap the fuselage in half before the wings break. Good for having a giggle in tight mountain passes.

(note: this is based on oldFAR, I haven't messed with the dev build yet)

[jrandom]

Perfect! That's exactly what I needed. Thank you.

[Wanderfound]

Perfect! That's exactly what I needed. Thank you.

Keep in mind that there's also a fair bit of influence from personal design and piloting style. Rough piloting demands tougher wings than smooth, high wing loading wants tougher wings, extreme low-altitude speed likewise, etc. etc. Wing strength of 0.5 is fine for most straight-line-to-orbit spaceplanes, but it won't cope with being flown like a dogfighting F-16.

Information provided for recreational purposes only, Kerbodyne accepts no responsibility for the consequences of following this advice, see your doctor if symptoms persist.

[jrandom]

I'm wondering if maybe the wing strength should default to 0.5, since 1.0 is too heavy and is overkill for most planes. At the very least it would make the initial designs faster to iterate over if you're adding and removing wings constantly.

[UnanimousCoward]

Last craft I built with wing strength at 0.5 destroyed itself immediately upon takeoff due to aerodynamic forces. Setting wing strength to 0.75 let me fly it to orbit - a first for me.

Course, it's probably because I'm doing something wrong. That's usually the case...

[tetryds]

On the new version it's pretty much like before, on the top of the heat, which makes it way more complex, it's fun.

Wander's rule of thumb is pinpoint accurate, I play on the obscure side of pulling as high g's as fast and as low as possible, and that's pretty much it.

The trick to lower the wing weight is optimal wing placement and stress management, most of suicides do not happen because of wing stress itself, but because of wrong maneuvering.

By simply making sure that your main wings center of lift is aligned with the center of mass you can already decrease the mass ratio by something between 0.1 and 0.05 (not linear).

AoA or lift peaks, high speed stalling and sideslip are the most common causes of wing stress death (esp. sideslip).

So, the major trick is to know your plane well, what it's capable of, and how it behaves at different speeds.

That is why I heavily recommend anyone attempting low altitude tricks to always put the navball indicator on Mach mode, and use a joystick.

Edit:

@jrandom: the default is 1.0 because when wing stress damage was first implemented barely anyone could even lift from the ground, I would go for a 0.75 default instead, but 1.0 seems to be very noob-friendly.

Edited May 5, 2015 by tetryds

[Vitasalato]

@Vitasalato: Even oldFAR was better than the current stock aero, if perhaps a little less robust. The soon-to-be-released-whenever-I-stop-finding-issues-that-I-really-should-squash version of FAR will be a whole lot more accurate than oldFAR was.

Vita - the current stock aerodynamics are much better than old, but FAR will still be way more realistic. It will model the transonic and supersonic regimes far better (better even than old FAR), and in the current stock model you basically can't stall. Essentially, stock gives a very cartoonified version of stability, lift and drag, where FAR gives you as close to the real deal as Ferram can give us.

Great, that what i was hoping!

No offense, but it would have taken less time to just look at the posts directly above yours to find the answer than asking it.

All right, when I asked the question, I knew the risk I was taking no reading older posts

[4plains]

As a very crude rule of thumb:

0.25: very light, very fragile. Eggshell glider, handle with care, especially at low altitude. Roughly equivalent to how all wings were before the mass tweakable was introduced.

0.5: Thoroughly solid, good enough for anything but extreme aerobatics or wing loading. Can be shaved down a bit further for a sensible pilot.

0.8: Extra tough, good for flight trainers, superheavy Mk3 stuff, light aerobatics.

1.0: Fighter strength, unnecessary unless you're intending to pull extreme high-G low altitude aerobatics.

>1: Magic plane with Adamantium wings. You'll probably snap the fuselage in half before the wings break. Good for having a giggle in tight mountain passes.

(note: this is based on oldFAR, I haven't messed with the dev build yet)

Seems alright. I've been doing acrobatics at the mountains west of KSC. I can do 25G and my wings don't snap off at wing strength setting 1. So solid andamantium used there. I'd say I can get away with 0.5 strenth wings.

[sam.johnson841]

The green and yellow lines/graphs, are they actually seen in the SPH or have you guys used ms paint?

- Sam

[Hodo]

The green and yellow lines/graphs, are they actually seen in the SPH or have you guys used ms paint?

- Sam

NuFAR they are in the SPH under the FAR tab.