FAR Fighter Challenge BD Armoury AI: 2!!

[Guest]

Right, I have one plane with small jet engine that beats K-191 most of the time. I submit it once I can be sure who the opponent will be.

I also have tons of almost final builds which I can base on in case I need a different profile (i.e. something different than insanely tight cornering). So if you ask me stock is not dead yet.

Still I am looking forward to having new parts to play with eventually.

RevanCorana, great stuff you are making! I always wanted to model some craft after a real plane, but in the end I seem to lack the patience. Your F22 and Rafale are really stunning. How many hours did you put into them?

Thanks!

I think I have over 2k hours of KSP or smth like that so yea.. with this much practice I can build faster now.

For planes it takes me around 1 hour or so for the rough model/structure but usually a couple more hours are needed polishing details.

It's a great process tho, very fun especially with FAR, which I'm still learning to use

How accurate is stalling by the way when it comes to super critical AoA ?

I can't reach high AoA often, but high lift seems much easier.

I'll have to see what the BD engines are like now: B9 used to have some mad vectored thrust engines which I've not tested in the current one ( engines are one of the things that'd most likely to be broken ). I do wonder if we actually need vectoring engines given we're pulling 30-40 degree AoA without them, not sure how the AI is going to handle full supermanoeverability

Blow itself apart eventually ^^

IRL aircraft needing thrust vectoring for maneuverability are usually flying bricks like the F-22

Stuff with low wingloading and streamlined airframe like the Rafale have more lift at lower AoA and less drag (and less inertia). Front canards also give the equivalent of nose control gained from TV

It seems where it's mostly useful is for low speed maneuvers like VTOLs etc, but overall it's a waste of both fuel and acceleration.

Edited November 1, 2015 by RevanCorana

[Van Disaster]

Ignoring the Harrier, it's most useful post-stall which is not something Western airforce doctrine usually gets near - hence no vectoring on the existing Typhoon ( or the Rafale ) for one, it doesn't need it for how it's meant to be flown. I note the Turks are after a vectoring EJ200 so there's a chance for a vectoring Typhoon demo at some point. Canards contribute positive lift during pitch-up which is why they're so good if you can stop them stalling at high AoA ( they also delay wing root stalling if you stick them close enough to the wings, iirc ).

FAR stalls deltas from the back of the wing at extreme AoA which is what's meant to happen, and it does it quite progressively. The worst part of this challenge is tuning the plane for the AI. This was taken with tint stall *on*, so nothing is stalled at a quite ridiculous AoA - I think I could maybe push it to 40 degrees but that'd be like dumping a parachute out.

As I thought, the AI is a bit overexcited about the BD vectoring engine, I'm really not sure what plane design could use this sensibly - plus they're over 2 tons, and while a real engine is also heavy it also doesn't concentrate all it's mass at the tailpipe. B9 pWings don't seem to crossfeed - I seem to remember a trick to get round that other than fuel lines but I can't remember what. We'd probably want this fuel pump mod if we were going to use them ( it's tiny & awesome ).

Edited November 1, 2015 by Van Disaster

[Guest]

I can't get over 20° because of lack of torque in the controls except when at very low speed, like climbing straight up then start spinning around with nearly no control/ no lift but very high AoA.

I'm guessing thrust vectoring becomes useful in these case as well as a way to trade speed for position.

I red front canards by lifting tend to create vortex behind in front of the delta wing at high AoA which increase lift of the delta itself or smth like that.

Is FAR is that accurate? noo

Typhoon is heavier and less nimble thrust vectoring could be nice.

For the crossfeed may be tweakableeverything

Edited November 1, 2015 by RevanCorana

[Van Disaster]

I think you're probably stalling, to be honest - the visualisation tools will tell you straight off because your wing will look like this:

If the plane is apparently bumping against something, it's probably a stall ( perhaps the control surfaces, especially with canards ). The Typhoon & Rafale get away with it because they've got way better flight control software than us and I think the Typhoon can happily turn itself into a stall if the FCS would let it - the problem with pulling high AoA manoevers is that you bleed energy like crazy even with vectoring. It's something we run into with these challenge craft.

FAR doesn't model close coupled surfaces - it does do *something* because there's a quite nasty biplane effect if you overlap wings, and also if your tail is directly in line with your wing and you stall the wing, you might end up stalling the tail as well.

Rather use GPOPump than tweakable everything because it's focussed - contests needs focus.

[Guest]

Yea seems like it because my CoL is in front of the CoM so shouldnt have lack of maneuverability.

Anyway I do get the "minor stall" status but major stall never happens above certain speeds no matter how much % the elevon are set at.

FAR doesn't model close coupled surfaces - it does do *something* because there's a quite nasty biplane effect if you overlap wings, and also if your tail is directly in line with your wing and you stall the wing, you might end up stalling the tail as well.

Do you mean that 2 wing totally clipped into each other willl NOT behave like 1 wing? That sounds ....ed up!

However it's normal for a tail to stall if the delta in front of it stalls, because its like a big chunk of airflow blocked ahead of it that's also why front canard are better at stall recovery usually. lol

Typhoon and Rafale (Gripen too) also have a short burst of angular momentum when pitching a lot quickly because front canard design

From this source https://defenseissues.wordpress.com/2013/08/24/dassault-rafale-analysis/ it is said the Rafale max AoA is 100°!!! probably low speed / high G cobra stuff.

You think this could work in FAR?

Edited November 1, 2015 by RevanCorana

[Van Disaster]

Do you mean that 2 wing totally clipped into each other willl NOT behave like 1 wing? That sounds ....ed up!

Yeah don't do that - Ferram has told me off in the past for just clipping wing pieces into other pieces, let alone completely. Clipping control surfaces is a real nono also. I think the new wing code will probably change matters. As far as I know clipping lifting surfaces into non-lifting parts is ok, it'll just cheat the voxelization a little.

What I meant by biplane effects is that if you have two wings *not* clipped but stacked close to each other and overlapping if you look from the top, you'll probably get less lift than just having one of them. That should show up in the graphs.

Typhoon and Rafale (Gripen too) also have a short burst of angular momentum when pitching a lot quickly because front canard design

I would imagine they're pitching up & hitting flight envelope limits, and the FCS goes "woah there". You can do that in a rear tail craft too, you just get better low speed performance with a canard and a bunch of built in instability to go with it...

it is said the Fafale max AoA is 100°!!! probably low speed / high G cobra stuff. You think this could work in FAR?

With manual flying, sure. I've pitched right over & flown backwards for a bit, there was a bunch of discussion about supermanoeverable craft in the FAR thread in about May or so. For the BD AI, no - it'll keep trying to pull for it's setpoint & not recover. This is why we have Dynamic Deflection as a requisite mod, so you can reduce control authority before it stalls the craft. I spent quite a while today trying to build something with the BD vectoring engines that the AI could fly, and I'm pretty sure there isn't a solution that wouldn't be equally as good with just the stock gimballing engines.

Javascript is disabled. View full album

Edited November 1, 2015 by Van Disaster

[Van Disaster]

Right, well given it's no longer the weekend I guess I'll run the next fight, and depending on results the one after that.

[Guest]

Didnt know FAR still differentiate wing pieces from non wing pieces makes no sense since the voxel is used.

Ive tried the mod B9 procedural wings and it's absolutely awesome! The wave drag is so low and the part count too now.

In addition having a single wing part versus several wings struted together is a huge difference in handling.

The new rafale is better with 40 parts total lmao

[Van Disaster]

Didnt know FAR still differentiate wing pieces from non wing pieces makes no sense since the voxel is used.

It isn't - wings are currently the same as they've always been. It uses voxelisation of the wings for wave drag, nothing related to lift.

[Guest]

Ok so if I make a delta wing shape with many small solar panels it will have the exact same lift and wave drag properties than one made with 1 delta wing? Not sure, it's so much smoother since I use the procedural wing but it may be due to bugged voxelisation too.

But now I have a new problem the engines overheat for no apparent reason above 450m/s

Edited November 2, 2015 by RevanCorana

[Van Disaster]

Ok so if I make a delta wing shape with many small solar panels it will have the exact same lift and wave drag properties than one made with 1 delta wing? Not sure, it's so much smoother since I use the procedural wing but it may be due to bugged voxelisation too.

Wave drag, probably. Lift, not - wings do NOT use voxelisation to work out lift at the moment.

But now I have a new problem the engines overheat for no apparent reason above 450m/s

That I can't help you with - and for the purposes of entering the contest as it stands it's not really relevant given we limit to 400m/s for now. There's good reasons for that, the AI tends to get out of range too easily above that. I might try running a few fights at 450m/s to see if it makes them a bit more vertical, but most of our current planes would fall apart given they're getting pretty optimised for slow speed.

[Guest]

Thrust limited to 80% fixed it, also this brings the power closer to real rafale which is always good.

Wave drag also influence maneuverability at high AoA right? Not sure what it represent other than the reduced drag at transonic speeds.

This is what I got so far so good but seem to have lost a bit of lift in the process...

[Van Disaster]

Swallow F2 vs K-191 fights are filmed & composited; rendered output is 685MB so that will be uploading for quite a while.

[Van Disaster]

And the fights: filmed for fullscreen viewing & uploaded in 1080p

Grats EzBro for your first championship spot.

As promised, my basic engine entrant: Hummingbird F2

[E805BzRo]

Yey!

Now I have to put some truth to my boasting about planes that are even better... (*download Hummingbird F2...)

Cheers,

EzBro

[Van Disaster]

Could do with a couple more entries too... Hodo? Dwerto? TE? SPA? Wanderfound?

[SpaceplaneAddict]

Sure, I'll start building some randome entries ASAP.

-SPAddict

Sorry I didn't record this weekend, I'm trying to circumnavigate Tylo as fast as possible.

- - - Updated - - -

Oh wait, the queue is finished? Time to suggest mods to vote on later! If you want a mod to be included in this challenge, suggest it for later voting! I'll pick 2 of the best looking mods to vote on.

-SPAddict

[Van Disaster]

It's not finished - two more fights in it plus whatever anyone else throws in. Still some mileage left in the parts rules we have too.

For further down the road I am testing b9 pWings, and I think I'll test proc parts. I tried the BD engines but I don't think they work well with the AI, unfortunately. If you lock the gimbal they're not bad aside from the extreme mass... I'll give the B9 vectoring engines a shot too but I think that'll end up with the same result.

[SpaceplaneAddict]

https://drive.google.com/file/d/0B588M8xKv0QKSl84TDg2emZVSmM/view?usp=sharing

There you go, the Meme Fighter. I believe you should run it, it's a bit cheeky for me because its my own.

-SPAddict

- - - Updated - - -

nvm don't DL that.

It's still illegal.

- - - Updated - - -

https://drive.google.com/file/d/0B588M8xKv0QKTzZxODNXYnZUTkE/view?usp=sharing

This is legal now.

[E805BzRo]

I do have stuff in the pipeline, but I don't want to hold everybody up by uploading challengers that end up fighting my own plane. My next submission will be after K-191 is beat - or after we start a new contest with new mods.

The latter would mean that my plane stays on the throne indefinitely, right?... can't decide if that would make me ecstatic or embarrassed...

EzBro

[Van Disaster]

I'll run Magnavox' craft vs K-191 tomorrow, then it's my craft again and after we'll see who SPA's is up against. I have a much better Swallow around which is not terribly ( if at all ) better than my current entry, and another airframe ready but not really tuned yet; after that I'm barren right now but let's see what happens.

If you want something a bit different to try, let's install Better Bouyancy & proc parts, and have a seaplane class

Edited November 2, 2015 by Van Disaster

[SpaceplaneAddict]

oooh, sounds like a something I'd try.

I'll consider.

-SPAddict

[Van Disaster]

Might want to try doing what EzBro is & finding some more FPS temporarily - your plane is rather violently exploding at 1400m pulling over from the climb

[SpaceplaneAddict]

FPS? wha?

[TrainEngie]

(insert random lack of sleep or low grades comment)

I'm not sure how to design a proper fighter jet. I keep coming back to this challenge, only to find out the next 100 new TE designs don't work, and that clones of the F-15 don't work either.

(excuse the narcissism)