Jump to content

Simple question about F.A.R


Recommended Posts

Ok, i'm not an absolute noob in KSP but I'm far from being a veteran. I'm the perpetual student, always learning (and occaisionally even succeeding). So here's my question:

The FAR mod, yay or nay? I try to keep my install pure stock, but the idea of a more realistic aerodynamic model is compelling. Further, I'm running KSP 1.10.0 and will try to upgrade to 1.10.1 this wekend. Is FAR worth it, especially on these newest versions?

 

Link to post
Share on other sites

This is my personal opinion based on several hours I spent testing FAR and I am open to objections.

I think FAR is far from an "absolutely accurate" aerodynamic model. It does some things better than the stock game, but in other aspects it is actually worse. And here is what i mean by this.

1) Cambering. FAR takes zero account for cambering (the shape of the cross section of the wing), which means that unless you have a positive angle of attack or a positive incidence, your wings produce zero lift. This is unrealistic and the stock game does this actually better, producing lift even if your craft moves in a parallel direction to the plane of your wing surfaces.

2) Ground effect. In reality, the lift coefficient of your lifting surfaces is somewhat higher when you are close above the ground, which does actually help noticeably during take offs and landings by decreasing the stall speed. FAR, as far as I can tell from the tests I did with it, does not take ground effect into account.

3) Take off and landing speed. In reality, a typical airliner is capable of taking off and landing at speed somewhere between 70 m/s and 120 m/s. In FAR, I haven't managed to get into the air at speeds lower than 200 m/s. I performed an exhausting number of test flights - with my crafts, with stock crafts and even with the crafts that came with FAR, and with the exception of the FAR included fighter jet (I cannot recall its name), it was impossible to maintain altitude at speeds below cca 200 m/s, which makes a safe landing of your average SSTO pretty much impossible. The only reason why the FAR fighter jet had a better performance than the other aircraft was because it was equipped with an unrealistic amount of control surfaces. This is not how real aircraft should behave IMHO.

4) In flight stability. Again, all the aircraft I tested, including stock and FAR aircraft, displayed a severe and immersion breaking instability in pitch and yaw directions even close to trans-sonic speeds. Real aircraft do not wobble up and down, FAR aircraft did for me. Perhaps I was doing something wrong, but all the SAS on/off and FAR stability autopilot on/off combinations have been tried with zero effect.

5) Aerodynamic shadow. In reality, if you put the tail wings close to the main wings, the tail wings will be robbed of the stable airflow they need by the wings in front of them and will not function properly, if at all. None of this is to my knowledge simulated by the FAR model.

 

Having said all this, I think your best bet is to install FAR, play with it for a couple of hours, try different types of aircraft, watch their behaviour closely, and then decide whether you prefer the stock model or the FAR one. Neither of them is perfectly realistic, so do not believe people who say otherwise.

Edited by Aelipse
Link to post
Share on other sites
2 hours ago, Aelipse said:

This is my personal opinion based on several hours I spent testing FAR and I am open to objections.

You'll get them, because "several hours testing" isn't anywhere near enough to be proficient in designing aircraft for FAR, and any "testing" with aircraft not designed for FAR is meaningless.
 

2 hours ago, Aelipse said:

I think FAR is far from an "absolutely accurate" aerodynamic model.

It's not meant to be. It's a more accurate aero model, that's all.
You're missing the big selling point anyway, which is shape-based aerodynamics and accurate drag simulation - the two historical problems with KSP that FAR was created to solve.
See, one upon a time, drag in KSP had nothing at all to do with the shape of a craft, and worse still, it was based on the mass of a part rather than it's cross-section. Thus "pancake rockets" were the norm.
The new KSP aero model assigned "drag cubes" to individual parts based on their cross-section, while FAR bypassed the "bunch of parts" concept altogether by calculating the shape of the overall craft.
Both still make assumptions and rely on approximation, it's simply not practical to make an "absolutely accurate" aerodynamic model in the context of a game like KSP.
 

2 hours ago, Aelipse said:

Cambering. FAR takes zero account for cambering (the shape of the cross section of the wing), which means that unless you have a positive angle of attack or a positive incidence, your wings produce zero lift.

Not true. FAR assigns magic "wing values" to wing parts just like stock does, and those attempt to take into account cross-section just like stock does. The actual implementation (and therefore results) differ though, and any mod wings will indeed have the basic "flat plane" lift all parts get unless patched for compatibility.
 

2 hours ago, Aelipse said:

Ground effect.

Neither stock aero nor FAR simulate ground effect in any way. It's been suggested repeatedly, but the usual answer is that it wouldn't add enough to gameplay to justify the effort.

 

2 hours ago, Aelipse said:

Take off and landing speed. In reality, a typical airliner is capable of taking off and landing at speed somewhere between 70 m/s and 120 m/s. In FAR, I haven't managed to get into the air at speeds lower than 200 m/s.

Take off and landing speeds are higher in FAR, but nowhere near as bad as you make them out to be. This is largely because parts in KSP are unrealistically dense, so your aircraft are unrealistically heavy. Garbage in, garbage out.
I regularly take off and land at anywhere from 70-120m/s in FAR, with very normal looking aircraft. My standard issue "orange tank to LKO" FAR spaceplanes land at 80-100m/s.
That's not to say a 100m/s landing speed isn't problematic, the runway is short and KSP's janky landing gear are prone to random misbehaviour at anything much over 80. Drag 'chutes are the answer, just as they were on shuttle.

2 hours ago, Aelipse said:

a safe landing of your average SSTO pretty much impossible.

Rubbish. I land SSTOs in FAR all the time (sometimes even on the runway :P), my late-game is almost exclusively spaceplanes, and I haven't played without FAR since 2014.
 

2 hours ago, Aelipse said:

In flight stability. Again, all the aircraft I tested, including stock and FAR aircraft, displayed a severe and immersion breaking instability in pitch and yaw directions even close to trans-sonic speeds.

Aircraft that are stable according to FAR's design and simulation tools are stable in flight, and I regularly fly smooth and level with only trim. Did you run an analysis on those "severe and immersion breaking instabilities"?
If you're using SAS, don't. It's tuned for stock and will almost certainly induce pitch oscillations with FAR. Real aircraft don't have SAS or reaction wheels anyway.

 

1 hour ago, lucho said:

lots to mull over.

Unfortunately the biggest difference (IMO) hasn't even been covered. Namely FAR's voxel system, which throws out the "just a bunch of parts" idea and actually analyses the shape of the craft.

In FAR, cargobays and fairings aren't special magical parts with a "shielding zone". You can make a hollow structure from any old parts and have it shield stuff inside.
Non-wing parts get (simple) lift simulation too, so lifting bodies work even when not made from mk2 parts and planes made from structural plates will (mostly) fly.
If you clip things into the fuselage to make a smooth profile, it has the drag of a smooth profile. If you clip wings into the fuselage, the clipped sections don't provide lift. No more "hidden wings".
Various weirdness like open-nodes having ridiculous drag just go away, because nodes have nothing to do with drag in FAR. Parts offset 1mm too far outside a cargobay don't suddenly get huge drag values either.
Mach 1 isn't a magic "multiply drag by X" zone in FAR, if you area-rule your craft like a real supersonic design, it has the wave-drag of a real design too. You even get a bonus for realistic intake-engine geometries, as if the parts between were actually hollow. The key to supersonic in stock is "moar engines", in FAR it's "moar slippery shape".
Supersonic biplanes and other nonsensical wing geometries suck in FAR almost as much as they do IRL. Stock aero only really cares about how much wing you have, FAR adds where and what shape to the equation.
That which looks like a RL aircraft generally flies like a RL aircraft. That which looks "so kerbal" with wings stuck on at random will likely go into a flat spin or catastrophic stall 3' from the ground.

There are however some caveats:
Lift is generally lower in FAR, because it doesn't try as hard to balance against the nonsense stock density values. This increases takeoff/landing speeds and exacerbates wheel-jank.
Many (most in fact) stock craft will be unstable in FAR, because stock aero is super forgiving. You will almost certainly need a bigger vertical stabiliser for a start, for some reason yaw instability isn't a thing in stock.
Stalls in FAR can be and often are unrecoverable, just like IRL. If you want supermaneuverability, you need to carefully design for it.
You will spend a lot longer designing and tuning your aircraft in general. Getting some aerodynamic theory under your belt is advised, because you will want to be able to read the information the design tool gives you.
Water is death. Ferram tried several times to get water to behave properly with real-ish fluid simulation, and I suspect he kinda gave up in the end. Seaplanes are still possible, but they're hard.
Many other mods such as autopilots, MechJeb etc. have a hard time, because they're tuned for stock aero.

 

Also,

 

Edited by steve_v
Link to post
Share on other sites
2 hours ago, Aelipse said:

Cambering. FAR takes zero account for cambering (the shape of the cross section of the wing), which means that unless you have a positive angle of attack or a positive incidence, your wings produce zero lift. This is unrealistic and the stock game does this actually better, producing lift even if your craft moves in a parallel direction to the plane of your wing surfaces. 

Most of your other points were already responded to, but I wanted to address this one specifically. 

Stock does NOT simulate the effect of wing camber in any way.  I have tested this many times and looked through the .cfg files to boot.  Lift in stock is exclusively AoA based.

Link to post
Share on other sites
36 minutes ago, Lt_Duckweed said:

Stock does NOT simulate the effect of wing camber in any way.  I have tested this many times and looked through the .cfg files to boot.  Lift in stock is exclusively AoA based.

Quoting for truth. Not sure how anyone could mistake this, as it governs the entire simulation. In stock, wings at zero incidence provide ZERO lift, period.

 

1 hour ago, steve_v said:

for some reason yaw instability isn't a thing in stock.

It is, and it is possible to experience flat spins and severe sideslip with flawed designs. Under FAR it's much more of a factor though, this is true.

Link to post
Share on other sites
4 hours ago, steve_v said:

Real aircraft don't have SAS anyway.

Real aircraft do, on the other hand, have fly-by-wire avionics.

And KSP aircraft kinda do require SAS to stay level over long distances because setting trim in the SPH to keep the plane level during subsonic flight causes nose-up at supersonic speeds and vice versa, optimizing for level flight at supersonic causes nose-down while still subsonic.

Link to post
Share on other sites

OP i was in your position back when i had gotten into KSP enough to start mucking about with planez a year or two ago.

FAR was the first mod for KSP i ever downloaded.

That being said; my experience was that it's a massive trap. 

Why? Well you're going to need more lifting area than stock to get planes in the air with "Reasonable" speeds, and thus you'll eventually be getting a mod for procedural wings. Then you'll realize that without the KSP stock fudge factor that makes all wings generate double or quadruple the lift (Something somewhere is squared, i forget what it is). That your mass ratios for larger planes are all outta wack, and thus will likely end up getting SMURFF. And then you'll realize stock engines aren't enough, nor are stock plane parts. So you'll download a variety of those part mods, then realize it would be nice to be able to power all of this stuff. Then you'll have Near Future Electric, and perhaps at that point would've gotten frustrated with the EXPLETIVE landing gear. So you'll go grab KSP wheel, roll some Stock Patches for the stock gears. Then you might find yourself needing some unusual sized bits, so now you have Tweakscale...

Before you know it; you now have a 20-30 mod install just to feed the monster that is this incredible mod.

I personally have no regrets, and while some people will say KSP "Aero modeling" has gotten better. The fact is that KSP doesn't attempt to even approximate how aerodynamics even works for aircraft, the implementation is mostly intended for rockets and was backported for aircraft only after people had expressed some interest in them. This leads to rather silly situations where people can take aircraft from low supersonic to hypersonic by just wacking nosecones into engines and intakes, because all KSP is looking at are integer values for Drag and Lift....

FAR isn't perfect, and it can be frustrating at times. But you don't need a degree in aeronautical engineering to use it, just patience. And once you do, you'll find the rewards far greater than anything stock Aero could ever provide.

Link to post
Share on other sites
11 hours ago, swjr-swis said:

It is, and it is possible to experience flat spins and severe sideslip with flawed designs. Under FAR it's much more of a factor though, this is true.

I find I have to try to make something unstable in stock aero, and just sticking a single tail fin on at random seems to be enough for "stability".
IMO, planes in stock are so forgiving and controllable it feels like cheating.
 

8 hours ago, Fraktal said:

Real aircraft do, on the other hand, have fly-by-wire avionics.

Modern ones do, but aircraft were flying level with manual trim and good design long before it was a thing.

8 hours ago, Fraktal said:

And KSP aircraft kinda do require SAS to stay level over long distances because setting trim in the SPH to keep the plane level during subsonic flight causes nose-up at supersonic speeds and vice versa

So trim in flight, again like real aircraft do.

There are flight assist mods and autopilots which work well with FAR anyway, stock SAS just isn't one of them.

Link to post
Share on other sites

wow, lots of info. In stock I don't have trouble taking off at speeds of 80-120 m/s, my problem is keeping them on the runway - they always want to veer off to the left or right. It's also not difficult to fly them so fast  they turn into a ball of fire. it seems FAR could teach me something about real aerodynamics ;)

 

Link to post
Share on other sites
Just now, lucho said:

wow, lots of info. In stock I don't have trouble taking off at speeds of 80-120 m/s, my problem is keeping them on the runway - they always want to veer off to the left or right. It's also not difficult to fly them so fast  they turn into a ball of fire. it seems FAR could teach me something about real aerodynamics ;)

Yeah that has more to do with KSP's stock wheel implementation and the lack of tools to get them nice and perpendicular with the ground, but again there's mod(s) for thattm

Biggest thing in FAR i found personally wasn't even takeoff, it was landing. The gears in stock have abysmally low collision tolerance, and will break at speeds >80-120m/s upon attempting a landing. I just used the same MM patch i implemented the KSPWheel stuff into and whacked in a section to increase their tolerance. IRL there's much heavier gear (though often at the cost of extra mass and volume, just look at the difference between carrier-bound and land aircraft for an extreme example).

As for the teaching, absolutely. You'll likely find yourself actually looking at things like "What makes a classically stable aircraft" (One that attempts to return to level flight when the nose is pitched down, therefore resisting changes in pitch) or exploring things like how supersonic flow differs from subsonic and transonic. And how to shape an aircraft to get the best performance wherever you most want it.

I don't think any less of anyone for not using FAR, i might cringe at Matt Lowne when he's desperately whacking nose cones into engines or intakes. But without a doubt i'd consider him a better player than me.

I came from flight simulators, and therefore the egregious disparity between KSP stock aero and IRL aero bothered me significantly. But Stock Aero does the job fine for just lobbing rockets, so for the vast majority it's fine. If anything, I'm just glad the option is there in the form of mods for weirdos like me that want somewhat accurate behavior of planes in their space game.

Link to post
Share on other sites

Yep, landing has always been my weak spot, with or without FAR

 

Ok I downloaded and installed FAR. I have to get back to work, so I was only able to give a quick test

First I took my Lowrider (shameless plug: Give it a testdrive ). First impressions: Oh. My. God. Horizontal speed has doubled,  and the climb rate is INSANE . I  let my attention lapse for a moment, and the Lowrider was at 10km altitude :o

I also tested another couple of VTOLs and a generic SSTO that I whipped up the other day. In general the controls are a bit more sensitive but performance seems broadly similar (my inerpretation: I need more testing on more designs).

I'm going to have fun testing FAR when I get some time later :cool:

 

Link to post
Share on other sites
12 minutes ago, lucho said:

First I took my Lowrider (shameless plug: Give it a testdrive ). First impressions: Oh. My. God. Horizontal speed has doubled,  and the climb rate is INSANE .

I have never actually used those new BG parts in FAR (TBH I think they're a bit kludgy and OP in general and stuck with modded single-part prop engines), I'm mildly curious as to how they work out... I suspect the answer will be "horribly unbalanced  and/or janky" since they're designed exclusively around stock, but I'm open to pleasant surprises.

Edited by steve_v
Link to post
Share on other sites

To answer my own question: No, apparently FAR does not properly support BG rotors/robotics, because Squad didn't expose the events needed.

And,

Aside,

56 minutes ago, lucho said:

In general the controls are a bit more sensitive

FAR's DPCR (Dynamic Pressure Control Reduction) flight-assist is pretty much an always-on for me, it'll (mostly) keep those sensitive controls from ripping your wings off at supersonic speeds.

Edited by steve_v
Link to post
Share on other sites
  • 3 weeks later...
  • 2 weeks later...

For those constructing ssto spaceplanes, especially for larger planets like RSS Earth, the biggest problem with FAR is the hypersonic L/D ration. Apparently FAR lacks proper compression lift modelling, as a result, hypersonic lift to drag ratio is very small, compared to stock aero, or real life(at more ideal, better opptimised circumstances), even if you use extreme amount of part clipping to get most of your body drag away. This forces you to use higher twr propulsion methods, such as chemical rockets, intead of low thrust high isp propulsion methods such as lvn for the closed cycle phase of the ascend, and also eat away your jet top speed, or forces you to use more jets, hence more dry mass. While stock aero isn't realistic, interms of the shape-aero correlation, it clearly will result in better performence of the vehicle suppose enough optimisation is put it.

For wing camber, any "reasonable" aero model must treat wings symmetrcially, since when mirror image of parts in editor is applied, it's actually rotation of the part rather than the true reflection. So if there is a built in camber recognition of the wings based on it's geometry, then what happens is that your craft will start rolling since one side of your lift is pointing up, while the other side is pointing down, as soon as you use mirro symmtry to buid wings. 

 

Edited by moar ssto
Link to post
Share on other sites
This thread is quite old. Please consider starting a new thread rather than reviving this one.

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...