Scott Manley showcases everything I find wrong with FAR

[stevehead]

Most of the problems come from the existence of stock aero model in the first place. Everybody using FAR likes it. And even those new people who know how real planes look and how you have to fly them have no problem playing with FAR.

Most people saying FAR is bad are probably people who 1. have never flown a simulated plane before and are new to all the controls and behaviours that are involved AND 2. Probably started with stock Aero and got used to it, by now they know how stock behaves, what is possible and what not and how planes have to be flown in stock.

A lot here comes from the fear of having to go through the learning curve again, they have mastered one curve already (for stock) and don't want to lose all their Atmo capabilities and have to relearn all the things they want to do.

This right here. If a FAR-like (or at least a NEAR-like) aerodynamics model had been introduced much earlier, people would not be protesting against a more realistic model as they are now. KSP would have been just as popular as it is now. It's the sociological problem that people are generally resistant to change is what truly is driving this debate, not the differences in difficulty of the two aerodynamic models in question. The switch to FAR from stock is a perceived difficulty because it is different to what the person is used to, and that person uses this perceived difficulty as an excuse rather than acknowledging their resistance to the change. I don't fault these people because everyone does it throughout their entire lives. I've been playing with FAR for over a year now, and the other day I tried the game with stock aerodynamics and I struggled; I recognized that this was the perceived difficulty I had mentioned since I struggled just the same when I made the switch from stock to FAR a while back (at the time I initially blamed it on FAR being more difficult). But I also do acknowledge that there are a few people that legitimately find FAR to be more difficult than the current stock model, but the same can be said about the reverse of that statement. I guess my point is that you cannot debate the difficulty of the two models unless you have played both of them extensively. If you are still in the learning curve for FAR, of course it is going to be more difficult.

As for the new aerodynamics model that Squad is working on, I guess we'll just have to wait and see what they ultimately come up with. Lately there has been a lot of promise from what Squad has mentioned to the public, I really hope where Max said "the old 45@10k ascent will still likely work" from the recent Squadcast really means "the old 45@10k ascent will only work if you completely over-engineer your rocket with reaction wheels, etc.". The game is still in development, so backwards compatibility and the current working knowledge of the aerodynamics model should not even be a factor in the development of the new model.

[Motokid600]

You can 45@10km with FAR, but it's horribly inefficient. And that's how it should be with 1.0. Go up to 10km if you like before you pitch just know your burning over a 1000 dV fighting gravity.

[stevehead]

I know that you can do the typical 'stock gravity turn' in FAR if you over-engineer your rocket. My point was more aimed at aerodynamics instead of fuel efficiency. A rocket in the new stock model that is going typical ascent speeds and doesn't have excessive torque through reaction wheels that attempts the 45@10km manuever should either find it impossible to do if the rocket is stable (i.e., center of mass towards the front of the rocket and center of lift towards the back) or it should spin-out uncontrollably.

Edited February 15, 2015 by stevehead
Failed to mention speed as pointed out by ObsessedWithKSP

[r_rolo1]

@ the OP

Well, I know I already commented on your post, but i guess that what I said what partly unjust. Your ideas are not bad and I'm 100% with you on the need for a better flight UI. The issue is that you used Mr Manley video to try to showcase that, and wrongly attributed the piloting issues to FAR, instead of being issues with the particular design he was using or other unrelated KSP shortcomings ( say, the reason why planes are hard to stop in the game as soon as you put the air drag in more realistic settings is because: a) lack of thrust reversers; KSP surfaces are very slippery; c) no surface drag chutes ). It was a bad picked example and actually made people rant on you just because of that and miss your point at all.

Oh and the amount of people coming by and saying "haven't read but let me toss my love/hate for FAR" is staggering :/

Edited February 15, 2015 by r_rolo1

[BudgetHedgehog]

A rocket in the new stock model that doesn't have excessive torque through reaction wheels that attempts the 45@10km manuever should either find it impossible to do ... or it should spin-out uncontrollably.

Ehh.. atmospheric density aside, it depends how fast you're going. If you're going slow enough (i.e. below say, 60 m/s), the airflow isn't enough to flip you or keep you stable, you're just going too slow for it to have any effect. That combined with the unnecessary gravity losses for 10 km are why it'll be inefficient. Hell, you can go up to 10km, slow down to literally hovering, pitch over 45 degrees using engine gimbal and then carry on with your ascent. It won't be efficient at all and is probably the second worst way of getting to orbit behind going straight up until you reach 80km and then turning right. Go slow enough and it'll probably work, as long as you have the fuel. Go fast enough and it won't. It's as simple as that.

[juanml82]

It comes down to basic design theory. If you have a poor design it is going to perform, poorly. If you have a good design, it will perform well. FAR just highlights many peoples bad designs and they can't handle it. It is not forgiving for "fail" designs, not in the slightest. If people cant get a grasp on that, then perhaps they shouldnt use FAR.

You just wrote a good reason to stay out of FAR if the player isn't into spaceplanes. You may write it as "Hey, I'm awesome and everyone who doesn't play KSP as good as I do sucks", but if the system is unforgiving, it doesn't have a proper friendly GUI to check the design before flying, it doesn't have a proper friendly GUI to inform the player they are about to loose control (or how much they can push it), it doesn't have a proper friendly GUI to set up and use flaps & airbreaks, etc, etc... well, to quote Ferram writing to Harvester "stop making it like a simulation and start making it like a game"

Eh... Then you may want to rethink your approach. You gotta start it WAY back. I'll kill the throttle several km from the runway. Keeping the prograde marker in the blue at that point will kill your speed enough to make it work. No airbrakes, spoilers, nothing.

Do airplanes land with their engines off?

One more thing about making HOTOL spaceplanes that look like atmospheric airplanes: No HOTOL has ever made it to orbit IRL yet. And the one design which might work doesn't look, at all, like a fighter jet. That's something to keep in mind before praising something for realism

[BudgetHedgehog]

One more thing about making HOTOL spaceplanes that look like atmospheric airplanes: No HOTOL has ever made it to orbit IRL yet. And the one design which might work doesn't look, at all, like a fighter jet. That's something to keep in mind before praising something for realism

To be fair, that's more due to the fact that Kerbin is stupidly small. For rocket to reach LKO, it takes about 3500 with FAR. For LEO, you need about 9500m/s. The ISS is travelling at around 7500m/s. You need a heck of a lot of fuel (or a super advanced engine) to be able to achieve that kind of ÃŽâ€V with one stage IRL. In KSP, comparatively speaking, it's laughably easy.

[Renegrade]

Do airplanes land with their engines off?

Half of my storebought ones did

No, scratch that, pretty much all of 'em. NiCads have amazing, violent power..and almost no runtime.

[Motokid600]

Do airplanes land with their engines off?

Nope, but I do. And it works really well. Holding an approach altitude with no power scrubs off all the speed you need to make the landing.

[Frafor]

Nope, but I do. And it works really well. Holding an approach altitude with no power scrubs off all the speed you need to make the landing.

You are right, gliding to the runway from far away at a really low descent rate makes it easier. Anyway, It think approach speed is to fast and the glide ratio is to good for the planes I tested.

[cpast]

Do airplanes land with their engines off?

Almost, actually. A standard landing for a jet happens at idle; while there is thrust, the thrust level is dictated by how long it'll take to spool up in a go-around (which is why the engine isn't *off*, and is at "flight idle" which is higher than "ground idle"). Exceptions include aircraft carrier operations (where you go to max power as soon as you touch down in case you miss the wires) and when you can't do your preferred descent, but you generally want to be as close as possible to no thrust to descend (if you need thrust, it means you're either wasting fuel or you wasted fuel previously in the descent). Prop planes may be different, but from what I can tell the preferred approach profile for jets is the one that lets you be at the lowest thrust setting that lets you spool up to go-around power in 6 seconds.

[klesh]

People mod the aerodynamics to suit their needs before 1.0, and they will do the same after.

Non-issue.

[PB666]

I know that you can do the typical 'stock gravity turn' in FAR if you over-engineer your rocket. My point was more aimed at aerodynamics instead of fuel efficiency. A rocket in the new stock model that is going typical ascent speeds and doesn't have excessive torque through reaction wheels that attempts the 45@10km manuever should either find it impossible to do if the rocket is stable (i.e., center of mass towards the front of the rocket and center of lift towards the back) or it should spin-out uncontrollably.

A stock game 45'@10K is not efficient now, I did some testing a while back on efficiency and found that even a launch pad and turn to 88' is starting to lose efficiency. The best turn start around 5k and turn around 1' per 1000 feet up to about 20K where one can move to about 2->3' per 1000 feet up to 32000 feet.

- - - Updated - - -

Ehh.. atmospheric density aside, it depends how fast you're going. If you're going slow enough (i.e. below say, 60 m/s), the airflow isn't enough to flip you or keep you stable, you're just going too slow for it to have any effect. That combined with the unnecessary gravity losses for 10 km are why it'll be inefficient. Hell, you can go up to 10km, slow down to literally hovering, pitch over 45 degrees using engine gimbal and then carry on with your ascent. It won't be efficient at all and is probably the second worst way of getting to orbit behind going straight up until you reach 80km and then turning right. Go slow enough and it'll probably work, as long as you have the fuel. Go fast enough and it won't. It's as simple as that.

Right, and with jets and fuel efficiency, you can literally trudge up at 20m/s to its IAS equivalent at 30km

If you are talking about this why not trudge up to 30,000 feet with a jets and simply relaunch after the engines die down and vertical acceleration reaches zero, its like launching from an non-atmospheric body. With jets you have enough fuel, so . . . . .

Edited February 15, 2015 by PB666
Did not picture jpg

[Pseudo Gold]

How about you try NEAR instead? It's more realistic than stock but is still easily playable.

[DaMichel]

I believe planes in NEAR are just as hard to stop as in FAR. Even worse, NEAR neglects some supersonic drag effects which help slow you down on reentry.

Furthermore, i hear that parts In KSP have higher density than in RL. So say, you build a replica just considering the shape of the plane. You would get the L/D aka. glide ratio of the real thing. But due to the heavier parts it would have higher wing loading and therefore need to fly faster to achieve sufficient lift.

Personally, i never noticed excessively long glide distances. Normal planes have glide ratios of over 10:1. So for 10 km altitude you should manage to fly 100 km distance. This matches my experience in game quite well. Has anyone complaining here actually done the numbers?

Edited February 16, 2015 by DaMichel

[Tygroux]

Played a hardcore-modded install recently, and I think FAR is fine.

What WOULD be a huge improvement, both in playability and realism is:

-Bigger, better wings. (B9 does that)

-Better brakes. (RealChutes help with drogues, I'm almost sure someone made a mod that increase drastically the stock brakes strength. B9 gives pretty strong airbrakes. (seriously, tried 6 of the double ones -reversed-. It's radical.))

-A 2km long runway. Doesn't kerbinside add that? I removed it because memory, but now that I fixed that I need to take a look.

Increasing the aerodynamic failure threshold (that's somewhere in FAR options, the thing that decide when your wings rips off) also help a lot, and, as always, watching tutos on how to aerodynamic stability, sub/supersonic transition and orbital-speed re-entry helps a lot.

[Van Disaster]

Troll thread seems to be turning useful, so...

Kerbinside has a 4 or 5km runway in the desert. Bahamut0's adjustable landing gear seems to have some pretty fierce brakes ( the old TT gear has strong enough brakes that I've broken both the runway and the plane landing occasionally... ). B9's procedural wings can store fuel, that changes a *lot*.

[Frafor]

Furthermore, i hear that parts In KSP have higher density than in RL. So say, you build a replica just considering the shape of the plane. You would get the L/D aka. glide ratio of the real thing. But due to the heavier parts it would have higher wing loading and therefore need to fly faster to achieve sufficient lift.

Personally, i never noticed excessively long glide distances. Normal planes have glide ratios of over 10:1. So for 10 km altitude you should manage to fly 100 km distance. This matches my experience in game quite well. Has anyone complaining here actually done the numbers?

I tested a small plane that weights 1.1t and has a wing surface of 15.54m2. Comparing it with a Piper Tomahawk (757kg, 11.59m2 wing surface) it flies pretty similar but it glides a lot more, with a ratio of about 20:1 compared to the 7.5:1 of the real one. The other thing I noticed is that while the Piper stalls at 49 knots, the ksp plane stalls at 85 knots. Also, the ksp plane can float over the whole 2.5 km runway after crossing the threshold at about 110 knots (25kt above stall speed) some meters above it, while the Tomahawk in the same situation (100kt, 51kt above the stall speed) will float for about 700 meters before stalling and landing. Apart from this, both handle pretty similar in normal flying, with the diference that the real one cruises at 95kt and the ksp one can do it at 150/200kt.

What i think from this is that the FAR aero model is right, but both lift and drag are too low. Apart from landing and with slow planes, everything seems right.

[PB666]

Played a hardcore-modded install recently, and I think FAR is fine.

What WOULD be a huge improvement, both in playability and realism is:

-Bigger, better wings. (B9 does that)

yep, definitely

-A 2km long runway. Doesn't kerbinside add that? I removed it because memory, but now that I fixed that I need to take a look.

X68 is 3 nautical miles long, most commercial heavy AC runways are 2 nautical miles+.

Maybe they could make that a bonus upgrade. Call it the Manley upgrade package. (For those who don't need thrust-reversers or brakes)

When they come out with retractable flaps package, I will take the flight game seriously.

[Vindaloo]

Personally, i never noticed excessively long glide distances. Normal planes have glide ratios of over 10:1. So for 10 km altitude you should manage to fly 100 km distance. This matches my experience in game quite well. Has anyone complaining here actually done the numbers?

My first flight capable craft using FAR had massive wings and I was unable to land on the KSC runway. Part of that is my low level flying skills, but also the fact that the huge wing surface gave a gliding slope somewhere upwards of 20:1; any attempts to "flare up" to bleed speed would have the craft immediately gain altitude, so landing was very frustrating. I never tested it but I am sure that I could have glided for several hundred kilometres before finally stalling out. During these test flights I got so mad that I eventually gave up planes altogether and just used rockets to get from biome to biome in Kerbin.

However, I eventually went back to the SPH and built a heavier craft with less wing span, and now I can finally land that thing pretty much where I want, at roughly 75 m/s; really it lands like a dream (high speed stability is still far from perfect, but it is a work in progress). I understand the OP's, and other's, frustrations with FAR. But really, just give it time and eventually you will reach the point of no turning back. And you most definitely can fly ridiculous looking craft in FAR, especially into orbit. Keep the TWR low, accelerate slowly & straight up until 35,000 metres, keep the speed below 275 m/s; using that strategy you can get the Titanic into orbit.

[cpast]

yep, definitely

X68 is 3 nautical miles long, most commercial heavy AC runways are 2 nautical miles+.

Maybe they could make that a bonus upgrade. Call it the Manley upgrade package. (For those who don't need thrust-reversers or brakes)

When they come out with retractable flaps package, I will take the flight game seriously.

The Shuttle Landing Facility (which is now TTS instead of X68) isn't 3 nm, it's under 2.5 (it's 15,000 feet), which is 4.5 km. Many runways for commercial aircraft are around 10,000 feet (particularly at sea level), which is only 1.6 nm (and 3 km).

[cantab]

KSC's runway, on the other hand, is 2.5 km. That's not extreme, but it is kind of short.

[DaMichel]

I tested a small plane that weights 1.1t and has a wing surface of 15.54m2. Comparing it with a Piper Tomahawk (757kg, 11.59m2 wing surface) it flies pretty similar but it glides a lot more, with a ratio of about 20:1 compared to the 7.5:1 of the real one. The other thing I noticed is that while the Piper stalls at 49 knots, the ksp plane stalls at 85 knots. Also, the ksp plane can float over the whole 2.5 km runway after crossing the threshold at about 110 knots (25kt above stall speed) some meters above it, while the Tomahawk in the same situation (100kt, 51kt above the stall speed) will float for about 700 meters before stalling and landing. Apart from this, both handle pretty similar in normal flying, with the diference that the real one cruises at 95kt and the ksp one can do it at 150/200kt.

What i think from this is that the FAR aero model is right, but both lift and drag are too low. Apart from landing and with slow planes, everything seems right.

Thanks. This is interesting. Could be that FAR doesn't get the lift / drag coefficients quite right. (Not complaining. It is cool that FAR works as well as it does!) The stall speed is odd though. Your plane has just a slightly higher wing loading according to your numbers. But one thing i have learned is that the number from the FAR gui is misleading because it includes the area of all wings including the vertical tail.

Here my only attempt at a real "replica". I designed it so that it has the same L/D ratio and wing loading as the archetype because KSP won't let me build any lighter. You see it floating here at 100 km/h which is actually the optimal speed for the real plane and just a bit over the optimum for the replica.

It floats really good. I put replica in quotes because there are some differences. The usual tricks to loose speed apply here, too: pulling g's, flying circles over KSC until low & slow enough, flaps, spoilers, sideslip (don't try without joystick). I did this a while ago and didn't think of recording numbers from actual flight testing. Perhaps i should give another attempt a go ...

Edited February 16, 2015 by DaMichel

[r_rolo1]

KSC's runway, on the other hand, is 2.5 km. That's not extreme, but it is kind of short.

True enough. KSC runway is of a perfectly serviceable size for a RL airport, but clearly short for a test facility. Say, compare it with Groom Lake ( aka the famous Area 51 ), that has a paved runway in excess of 7 km or the Edwards Base ones, that are even bigger , even if unpaved .

OFC that is peanuts if we don't consider that in game there aren't ( AFAIK ) any thrust reversers and the surfaces in KSP are in general ridiculously slippery. Without thrust reversers and in slippery conditions in RL, 2,5 km would probably be in the limit of what a decently sized passenger jet ( say , a Airbus 320 ) would find safe ( shorter runways, like São Paulo Congonhas, have see their fair share of accidents in wet condtions ) ... not mentioning that it would be unfit for any plane test facility :/

[Frafor]

Thanks. This is interesting. Could be that FAR doesn't get the lift / drag coefficients quite right. (Not complaining. It is cool that FAR works as well as it does!) The stall speed is odd though. Your plane has just a slightly higher wing loading according to your numbers. But one thing i have learned is that the number from the FAR gui is misleading because it includes the area of all wings including the vertical tail.

I took the number by measuring the wings with the SPH info tab, 1.3m wide and 11m from tip to tip. And I think FAR is getting the L/D coefficients right but at speeds higher than it should. I mean, the plane handles at 150kt like it should do at 90kt. For me, it feels like both lift and drag are nerfed so you need to go twice as fast to get the same results. If lift had twice the force and drag was 3x stronger I think it would feel better.

Here my only attempt at a real "replica".

Well, mine was not an exact replica. I sacrifaced everything to get a weight, wing surface and wing aspect ratio as similar as the real one as possible. I will keep testing to see if it's just me doing things wrong.