Thinking about making the switch to FAR.

[capi3101]

Alright - did some design tweaks last night, gunning for relatively low wing-loading (0.3 t/m^2) and moderate wing aspect ratio (~3.5:1). I was also wanting to see what 25% of the wing area for the tail looked like with a relatively short aspect ratio (wound up going 3:1) and 10% of the wing area for the fin.

Came up with this, the Phoenix 7c:

Flew pretty well - takeoff speed was low, around 95 m/s, and yes it did have issues at 30k with getting above Mach 4. FAR greenlit everything except N_ß at 30/4 and that was unstable, but the plane flew straight and true the whole time. Didn't even oscillate much until I got up above 20k and began leveling out to gain speed. Had about 18% rocket fuel reserve after orbital insertion. Same 6.1 tonne probe the 7b took to orbit. I'd have SPH screenies for y'all, but the game crashed on me prior to the actual flight and LoadOnDemand didn't want to work right away after I fired the game back up, so all my parts were purple and grey until about midway through the flight...

I'm thinking I probably could stand to up the wing loading a hair - the Concorde, Tu-144 and SR-71 had loadings up around 0.45-0.5 or thereabouts. 25% also seems a bit much for the tail still but it's hard to judge. I might try this design again as a tail-less delta with canards just to see what that's like. 10% definitely seems to be too much for the fin (and yet I've got a yaw instability, which in general you're supposed to correct with more fin, right?). I'm guessing the fin root chord needs to be longer than I've got it so I don't need so much span-wise, but then I run into the issue of insufficient mounting surface for where I've got it; I've already got issues with the design there as is. I've also noted that the fins on the SR-71, Concorde and Concordski are maybe 1.5 times the height of the fuselage and the fin tip chord is maybe half to two-thirds the length of the fin root. Shorter aspect ratio there too.

I need to figure out a good way to build a cranked delta wing too; just cuz. One of y'all mentioned I should be looking at fourth-generation fighters and that particular planform is used in a lot of the ones I've looked at. Pretty sure B9 Procedural Wings won't do them (at least not as a single wing piece). No idea if any mod in KSP handles ogival deltas either...

[Van Disaster]

Your vertical stabilizer is not far behind CoL ( or CoM ) so it needs to be bigger to compensate. Area * distance, not just area. You could raise wing loading a bit, just remember you're not actually building Concorde, it's just the nearest reference point we have so far.

[panzer1b]

Im actually wondering if FAR (or at least the simplified NEAR) will become redundant after 1.0. Perhaps not as im guessing even stock 1.0 wont be perfectly realistic, but then again, neither is FAR, you cant really have 100% realism in a game where you make lego styled planes/rockets that arent exactly copies of proven real life designs.

[Van Disaster]

Im actually wondering if FAR (or at least the simplified NEAR) will become redundant after 1.0. Perhaps not as im guessing even stock 1.0 wont be perfectly realistic, but then again, neither is FAR, you cant really have 100% realism in a game where you make lego styled planes/rockets that arent exactly copies of proven real life designs.

Not even remotely. Check out what will be version 2 of FAR sometime if you don't mind running dev software. NEAR perhaps, but that is well out of date as it is anyway and wasn't ever anything more than a joke.

From the various streams - which admittedly aren't a great thing to be judging from - the new aero is quite some way from realistic. Given I'm reasonably sure we can't actually compute realistic airflow in realtime yet, nothing is going to be realistic but stock is always going to be compromised in the way all the other bits of the game are, so I think if you're expecting something close to FAR you're in for disappointment.

Edited April 27, 2015 by Van Disaster

[Wanderfound]

Im actually wondering if FAR (or at least the simplified NEAR) will become redundant after 1.0. Perhaps not as im guessing even stock 1.0 wont be perfectly realistic, but then again, neither is FAR, you cant really have 100% realism in a game where you make lego styled planes/rockets that arent exactly copies of proven real life designs.

Based on what I've seen of the previews, FAR will still have a large fanbase. The new aero looks better, but it's nowhere near FAR grade.

[Hodo]

Based on what I've seen of the previews, FAR will still have a large fanbase. The new aero looks better, but it's nowhere near FAR grade.

This is pretty much my feelings as well.

[capi3101]

From the early reviews, it sounds like learning how to work with FAR was a good decision...bad news is I had a rocket fail to make orbit yesterday.

My own fault, though - it had a short, fat stock-origin asparagus booster that I just didn't feel like re-designing aside from adding an oversized fairing, clipping on some fins and adding some (undersized) nose cones. Furthest I made it was about 14k up; damn thing flipped over on itself each time. By which I mean I'd have it at 45 degrees on course 090 and it'd suddenly tilt up (back towards vertical, then spinning end-over-end out of control). CoM too close to the engines - lawn dart behavior. It's what I get for being lazy.

Sounds like I could've stood to learn DRE too...

[DaMichel]

The new stock aero is a nice improvement i suppose. At least infiniglide is gone. But i still prefer FAR. Why? Here are some brief observations from a quick f-15 lookalike:

* Stock control surfaces move awefully slow.

* Stock CoL indicator is imprecise. I added a bit of dihedral to the wings and the CoL indicator jumped over half the plane. FAR is not perfect but usually makes more sense than this.

* With FAR you can tweak the control surface deflection which is quite nice.

* FAR has the analysis gui. The stability derivs. may be confusing, but knowing the c_m curve is on its own just so useful.

* In FAR wings can stall. i.e. there is a sudden loss of lift. In stock it seems to be very gradual like before.

* IMO parts are easier to configure for FAR. There is no "lift rating" rating. FAR figures something reasonable out on its own, mostly.

Edited April 27, 2015 by DaMichel

[vosechu]

Based on what I've seen of the previews, FAR will still have a large fanbase. The new aero looks better, but it's nowhere near FAR grade.

This was the point from what I read. It wasn't intended to replace FAR, it was designed to make the game more understandable and to cover the basics a little better. I think of 1.0 aero as high school funtime and far as college physics. Sometimes I just want to stand around in the halls and act like a teenager though, and I'm glad to have both options.

[capi3101]

Haven't got FAR for 1.0 yet (because to my knowledge it doesn't exist for 1.0 yet); mainly just bumping the thread so it doesn't get lost in the shuffle.

I did try out the new version yesterday with no mods installed except for KER (and I'm not convinced it's working properly; I'd get lower TWR readings on the pad than I did in the VAB). Got up to the 45 point tier before the end of the night and unlocked the general aviation parts, enough to do mid-air parts testing and maybe a few tourist contracts. Based on y'all's observations so far, how close is the new aero model to NEAR/FAR?

Did build a general aviation plane last night to do a radial decoupler mid-air test. Let's just say I've gotten a bit used to having airbrakes, and I haven't unlocked the dedicated airbrake parts yet. Flight was successful - I just wound up coming to a stop well past the end of the gravel strip that passes itself off for a runway...

[Hodo]

I did the samething. I found the new aero model an improvement but it is far from being FAR. The plane I made took off at 34m/s, which is STUPID slow but hey it flew at that speed, and it wasn't even a small craft.

[Van Disaster]

Unfortunately new-stock still feels like soup :S it's like there's 5x the drag, so there's also 5x the thrust to compensate. Little AoA changes can lose humongous amounts of speed almost instantly, it's very odd.

[capi3101]

Alright...just downloaded the new FAR version for KSP 1.0.2; I'll give it a shot next time I can play. The weather where I live has been crappy of late. Tonight is no exception.

Edited May 8, 2015 by capi3101

[capi3101]

Haven't had much of a chance to play with the new FAR yet. Did a little the other day, and I noticed green and yellow squiggly lines are now showing up in the SPH. What are those things all about?

I've also read that new FAR and B9 Procedural Wings aren't getting along with one another very well. Any pitfalls to watch for there?

[Wanderfound]

Haven't had much of a chance to play with the new FAR yet. Did a little the other day, and I noticed green and yellow squiggly lines are now showing up in the SPH. What are those things all about?

I've also read that new FAR and B9 Procedural Wings aren't getting along with one another very well. Any pitfalls to watch for there?

New nuFAR released today; fixes most of the initial bugs.

The squiggly lines are about supersonic drag and area ruling. Green line shows how the cross-section area of the craft changes as you go from front to back, yellow line is the second derivative of that, showing how smoothly the first curve changes.

Basically, if you're planning on going supersonic, you want the green one to be as close as possible to a straight line (inclination not that important) and the yellow one as close as possible to a zero-inclination straight line. See http://en.wikipedia.org/wiki/Area_rule for why.

[capi3101]

Okay; I'll check out the new version. Bit of a relief actually - I had just gotten StageRecovery working before installing new FAR, and it had quit working again (which was aggravating). So I want straight lines, green can be inclined or not but yellow needs to be level with the ground - got it.

I'd actually seen that thing about the area rule the weekend before 1.0 came out (still trying to see if there was a way that could tell you when you had too much wing) and I was going to ask y'all about it because I really didn't understand it. It's just "make sure the area of the cross section doesn't change over the length of the plane", right? And since the wing contributes x amount of area, the area of the fuselage needs to be reduced by the same amount to ensure it stays stable in the trans-sonic range, right? Assuming that's right, my question was going to be one of how do you that with stock parts, and is it necessary to do so?

[Wanderfound]

It's not "make sure area doesn't change", it's "make sure that any changes in area happen smoothly". One way to do this is to have a featureless tube, sure, but it's not the only way. And the key point of area rule is that the distribution of area within a given bit of cross section doesn't matter; you can get away with big wings so long as the fuselage narrows a touch at that point.

You can massively reduce or increase supersonic drag just by tiny adjustments to the position of parts; there's scope for endless fiddling. Fun for those who like tweaking their designs, huge PITA for those who don't enjoy such things.

See here for example:

The thing that you're actually targeting is the Mach 1 Wave-Drag area; you want it as low as possible. The green and yellow lines give you clues as to where to concentrate your efforts. This one is fairly good already (although the yellow line is wiggly, the amplitude of the wiggles is low), but you could smooth it a bit further by adding a bit of bulge to the fuselage between the canards and wings (for example). I expect that tricks like "stack of empty Oscar-B's with a mini nosecone on the front, partially sunk into the fuselage" are going to see a lot of use.

Area ruling is part of why fighter jet fuselages tend to flatten at the rear of the aircraft; they're thinning the fuselage to compensate for the cross-sectional area of the wings. See how the cockpit canopy gives way to the intakes, then the fuselage flattens at the wings? And how the vertical stabilisers kind of fill the gap between the wings and the stabilators? Even though the distribution of area changes sharply, the change in total area flows smoothly.

Edited May 11, 2015 by Wanderfound

[tetryds]

@Wanderfound: Yes, you got it right about the area rule, but jet fighters are a different beast.

There is transonic area rule, but there is also supersonic area rule, where things change.

So, yes, they follow an area rule, but it's not the same as transonic, supersonic area rule is an average of the area rule of planes from different mach angles.

Same principle, just a bit more complex, you may notice that these supersonic fighters look very different from transonic airplanes like the Dassault Mirage.

They have enough thrust to overcome the transonic barrier, thus their main concern is being efficient at their supersonic cruise speeds, so their shape is different from transonic airplanes.

But don't worry too much about it right now, FAR still does not implement supersonic area ruling, it will probably do on the future, but that may take a while.

Anyway, good catch on the vertical stabilizers "filling the gap" between the main wings and horizontal stabilizers, that is also a good transonic aircraft practice.

You can also use the vertical stabilizer to smoothen out the main wings transition on delta-wing aircraft.

The stock FAR planes use many tricks for area-ruling, learning with them can help a lot, I recommend moving the ballasts around to see what happens.

Also: http://i.imgur.com/RPEKlW3.png

Edit: and ah, yes, with supersonic area rule comes "speed nicheing", when it's implemented you will have to focus on a more specific mach number to be efficient at, making trade-offs, the airplane can't be good at everything.

Edit2: I am posting it here and not on the main FAR thread to not scare everybody

It's not that much of a beast, but more like a spicy flavor.

Edited May 11, 2015 by tetryds

[Wanderfound]

Tetryds understands this much better than I do, Tetryds is wise. Listen to Tetryds.

In particular: as Tetryds says, have a fiddle and see what happens. Open up the graphs, move some bits about with the translation tools, and try to figure out how to minimise your wave-drag number. It's not that hard; the one of mine pictured above was my first attempt at it, and that ship roars up to Mach 5 on jets with no trouble at all.

[capi3101]

Might be worth my while to import all the wackiness I built for 0.90 and see how it all performs - and then see what can be done to fix things. Does mean more time in the litterbox, though.

So what exactly does the new FAR do over new stock aero (can somebody give me a summary)?

[tetryds]

@capi3101: Amram did it pretty well: http://forum.kerbalspaceprogram.com/threads/20451-1-0-2-Ferram-Aerospace-Research-v0-15-1-Fanno-5-10-15?p=1924777&viewfull=1#post1924777

[capi3101]

Alright then - first steps in 1.0.2. New design. First time I tried to build a plane totally from stock parts. Not trying to make orbit with this one - just trying to design a supersonic craft capable of reaching ~Mach 2 at 19k.

Javascript is disabled. View full album

FAR gave me greens up to 20k/Mach 2 - which is what I was going for - except for pitch-up tendencies (unstable M_w) in the subsonic range below 5,000 meters; I had a reaction wheel installed (in the aft Service Bay) which helped to keep the craft under control in that regime. Total listed wing area was a little over 28 square meters, so wing loading was somewhere around ~0.4 tonnes/square meter or thereabouts. It did struggle to get above Mach 0.8 under the jet's power alone; in the flight I lit the rockets at 15k to bring it up to the target speed range and help lift it to the target altitude. As I mention in the album, the plane actually did accomplish its mission but I was also running a test for the guy who runs the NavUtilities mod to see if it would work with blizzy78's toolbar (it's totally borked otherwise), and when I went to activate that mod's UI to bring up an ILS interface, it completely blanked my flight UI in Staging View (kinda like hitting F2, but no way to reverse it - I couldn't even bring up the game options menu with the ESC key). I was finally able to revert but I had to go to Map View to do it. That's beside the point...

I should probably mention that this was a career craft and I'm still very far down on the tech tree; I've got the entire 45-point tier unlocked but the only two 90-point techs I've got unlocked are Electrics and Fuel Systems (i.e. Solar Panels, the OKTO probe core - which this design utilizes in the forward Service Bay - Fuel Ducts and Large Fuel Tanks). Aerodynamics (Circular Intakes and Delta Wings) is next on my to-do list. The Runway is Level 2, the SPH is still Level 1 - so I've got the thirty part limit at work. Not far enough along to have B9 Procedural Wings unlocked...which is a drag (pun intended) as I know I could reduce the wave drag area by using a thinner wing.

Ways to improve this particular plane design, given my goals and current design limitations? Were I to have a less restrictive part number limitation, I'd want something up there towards the area where the fueslage meets the wing to smooth that out, right? And what do y'all think is going on aft - is what I'm seeing there just the rockets sticking further aft of the jet, is that what's causing that?

[Wanderfound]

I think that's the first time I've seen anyone manage wave drag area higher than the total cross-section area; I didn't realise that was even possible...

It's apparently doable to crack Mach 2 on basic jets alone, but it requires a lot of aerodynamic polish to achieve. However, as you've discovered, with enough rocket thrust you can punch through by brute force.

Bad Mw is usually caused by too much lift up front and not enough at the back. Either shift some wing rearwards, shrink the canards, expand the stabilators or pull the CoM forwards.

Given that it isn't intended to go to space, vacuum isp isn't of huge importance. So, I'd probably replace the LV-909s with Thuds; more thrust, more gimbal, and the radial attachment of them would let you slim the ship down to a single-fuselage layout. Something vaguely similar to this one:

Javascript is disabled. View full album

As well as substantially reducing your overall cross section (and saving a fair bit of part count), the Thuds are very useful for fine-tuning wave drag. Shifting them forwards/back/in/out by tiny amounts can have a huge impact.

[capi3101]

I'll give the Mk-55s a shot. Any suggestions for how to counter the added mass aft? The CoM is already further back than I'd like for it to be.

I guess if I switch to Thuds that'll free up four parts (the two side tanks and the two nose cones) - I could just add an additional FL-T400 up in the front or something, lengthen the fuselage and maybe give myself some more pitch authority in the process. I should mention I've got those Tail Fin canards set as static wing surfaces (i.e. an attempt to thicken the fuselage up there by the front of the wing). I might not need them at all if I switch over to a single fuselage design.

I thought that having the wave drag area higher than the cross-section area was unusual...

[Hodo]

The CoM can be moved forward by shifting where your fuel is actually stored on the craft. There are structural parts just for that reason.

I know on my Akula I built last version and it still works in this version it still works in NuFAR and is still capable of achieving orbit and returning even though it is quite a small aircraft.

I need to take new pictures of it.

The trick is using structure parts or intercoolers to shift where the CoM is actually on the craft.