[RESOLVED] Flew FAR jets in stock 1.1 souposphere, not impressed.

[Pds314]

EDIT: THIS WAS PROBABLY CERTAINLY CAUSED BY A BUG IN MY INSTALL OF UNKNOWN ORIGIN! REINSTALLING HAS FIXED THIS.

I installed the 1.1 prerelease (which is amazing), and therefore was forced to use the stock atmosphere instead of FAR.

Things I liked about using the current stock souposhphere:
...

Things I didn't like:
#1. The FAR Firehound MS can fly at 26 m/s in level flight without stalling. It weighs 14 tonnes! By comparison, a 12-tonne F-16 will not fly effectively below 60 m/s in real life! Fighter jets with all sorts of compromised wing geometries can fly at such low speeds one would think they were trying to pull over a car on the highway, when in reality, those jets can't fly below 80 m/s!
#2. The FAR Firehound MS has a top level flight speed of 150 m/s dry, 200 m/s wet, even at altitude. Even diving from 10000 meters, I couldn't get it to go supersonic no matter what. Keep in mind that it supercruises in FAR. This means that it is moving slower than a real WW2 monoplane unless it is on full afterburner, in which case they're tied.
#3. The FAR E42 has a top level flight speed of 330 m/s wet. Again, another subsonic supercruiser.
#4. The FAR SkyEye has a top level flight speed of 112 m/s even though its an AWACS/Airliner hybrid. (A rather odd combination, Ferram4, but I digress). Normally, it is around 280 m/s. At least!
#5. The stock Dove, equipped with a RAPIER, cannot go beyond 310 m/s at sea level.
#6. The lift vector appears to be in line with the wing, not with the prograde vector. Why is this? Lift goes perpendicular to the prograde IRL.
#7. The atmosphere feels more forgiving than and just as oppressive as the pre-1.0 souposphere. For one, being able to ditch ANY plane by just hovering over the water at 30 m/s is ridiculous. At least in the past, stall speeds didn't usually go below crash survival speeds. Now, you can all but land an F-104 on a helipad.

Suggestion: make planes act like planes. Fiddle with the drag settings and such so that you get experimentally correct results. I.E. A realistic fighter jet has a stall speed of somewhere close to 60 to 80 m/s. An airliner moves at low transonic speeds, and can even do so near sealevel. Realistic fighters with engines as powerful as the Panther can essentially always go supersonic and usually can supercruise. Tiny KSP fighters with engines as powerful as the panther would DEFINITELY supercruise.

Fun fact: Real 5th gen fighters with TVC can fly at 70 knots or less, by definition, but only stalled at angles of attack over 70 degrees AOA. The Firehound MS was going 50 knots... at 15 degrees AOA...

Also: this was not a complaint on the fidelity of the stock model. I realize that Squad doesn't intend to make something that matches FAR or any other flight simulation technology in its predictive capability of real life results. What my main complaint was about was the systematic error of my install with respect to over-estimating lift and drag. i.e. An airliner should go airliner speeds. A fighter should go fighter speeds. Both should takeoff in the neighborhood of 70 m/s, not, for example, 30. If it were merely a poor simulation, and not systematically off, we would expect to see accurate results for the most commonly built types of aircraft. I.E. airliners moving at Mach 0.8 and  fighters moving at Mach 2+. This was apparently not part of 1.1 or the sotck model, but a problem with my particular installation.

Edited April 4, 2016 by Pds314

[RA3236]

All the FAR planes are not meant to work in stock. As for the atmosphere, SQUAD will get to it.

#6 is wrong. The blue marker is in fact the centre of lift, which means even in real life it won't be in line with prograde all the time.

Edited April 3, 2016 by RA3236

[steve_v]

Out of curiosity, this is with reference to the 1.0.x atmo, yes? I use FAR pretty much excusively, but last i checked (1.0.5) most FAR craft flew okay-ish in stock.

Guessing this just needs some minor tweaking, have you logged it as 'feedback' on the tracker?

[cantab]

FAR uses a whole-shape approach to determining drag, while stock uses a part-by-part approach. This means that optimising a plane in FAR often involves clipping parts together, especially when you get into area ruling. Those clipped parts produce excess drag in stock, especially if you have open attachment nodes.

I doubt we'll get anything significantly better in stock any time soon. After all, if Squad had wanted good aero why did they not approach ferram?

[klgraham1013]

3 hours ago, RA3236 said:

 As for the atmosphere, SQUAD will get to it.

They already did for 1.0.  I'd be surprised if we see any more drastic changes to the aero.

[Pds314]

7 hours ago, klgraham1013 said:

They already did for 1.0.  I'd be surprised if we see any more drastic changes to the aero.

Exactly. I'm not familiar with 1.0.5 aero, but I feel like it might have gotten worse (more soupy/floaty) since then.

Eh, maybe I'll edit the aero myself to match FAR performance?

10 hours ago, steve_v said:

Out of curiosity, this is with reference to the 1.0.x atmo, yes? I use FAR pretty much excusively, but last i checked (1.0.5) most FAR craft flew okay-ish in stock.

Guessing this just needs some minor tweaking, have you logged it as 'feedback' on the tracker?

My impression is that this minor tweaking should involve reducing drag by maybe a factor of 10, and reducing lift by maybe a factor of 5. Guess I'll try messing with the aero settings to match that...

Basically, in order to be unable to exceed 200 m/s with the force of 2 full-afterburning panthers, a fighter would need a Cd of 1 and a subsonic drag area of 12 square meters!

Edited April 4, 2016 by Pds314

[regex]

Yeah, ferram4 upped the stakes with nuFAR and voxelization.  nuStock (1.x) really just took an approach similar to oldFAR AFAICT.  I find stock craft fly much more like oldFAR since wing shape is assumed to be the best possible whereas using nuFAR wing shape is crucial.  Stacking wings isn't always the best thing in oldFAR or nuStock because wings add drag and you want to cap attach nodes since drag is computed on those...  nuFAR just cares about the shape of the voxels comprising the craft.

nuStock is just fine for KSP (infinitely superior to the old force field drag model).  If you want more you install FAR.

Edited April 3, 2016 by regex

[Nich]

2 hours ago, Pds314 said:

Exactly. I'm not familiar with 1.0.5 aero, but I feel like it might have gotten worse (more soupy/floaty) since then.

Eh, maybe I'll edit the aero myself to match FAR performance?

My impression is that this minor tweaking should involve reducing drag by maybe a factor of 10, and reducing lift by maybe a factor of 5. Guess I'll try messing with the aero settings to match that...

So SSTOs are capable of going interplanetary on airbreathing alone or does FAR reduce thrust too? 10x less drag means about 3x higher top speed

Edited April 3, 2016 by Nich

[pincushionman]

3 hours ago, Pds314 said:

…Eh, maybe I'll edit the aero myself to match FAR performance…

Not really possible. You might get something oldFAR-ish, but the nuStock and nuFAR approaches are about as different as can be. The only way to get nuFAR-like performance is…

…install FAR. Whodathunk?

[Pds314]

1 hour ago, Nich said:

So SSTOs are capable of going interplanetary on airbreathing alone or does FAR reduce thrust too? 10x less drag means about 3x higher top speed

No, not really. most planes, like IRL, are temperature-limited or exhaust-speed limited. A reasonable 2x panther aerodynamically sane craft can reach 800 m/s with no issues. This is inline with real top speeds of modern fighter jets.

Drag-limited jet flight ended by the late 1950s or early 60s. Most planes fly at temperature (Fighters) or exhaust speed (Airliners) limits. The only jet I can think of today that can't even get close to its exhuast speed or temperature limit is the A-10 (Even so, 440 mph vs. 580 or so isn't that big a difference), plus possibly some small private jets (which often are limited by other factors like cruise altitude) or model planes (which have almost fighter-like exhuast velocities but not supersonic intakes or parts meant for ram-compressed air) or something.

54 minutes ago, pincushionman said:

Not really possible. You might get something oldFAR-ish, but the nuStock and nuFAR approaches are about as different as can be. The only way to get nuFAR-like performance is…

…install FAR. Whodathunk?

Yeah, but you could still match experimental drag values, lift values, stall speeds, etc with nuFAR, real life, or even oldFAR and get something better than stock.

Edited April 4, 2016 by Pds314

[cantab]

Stock aero could be tuned to make simple, straightforwardly-built aircraft fly more realistically, but it will then continue to produce excess drag in clipping-heavy aircraft. Such clipping is likely when making replicas as well as when optimising craft for FAR, because you need to clip stuff in order to sculpt more complex shapes.

If you reduce the drag so those clippy craft behave themselves, then the simple craft will be too slippery.

[Pds314]

3 minutes ago, cantab said:

Stock aero could be tuned to make simple, straightforwardly-built aircraft fly more realistically, but it will then continue to produce excess drag in clipping-heavy aircraft. Such clipping is likely when making replicas as well as when optimising craft for FAR, because you need to clip stuff in order to sculpt more complex shapes.

If you reduce the drag so those clippy craft behave themselves, then the simple craft will be too slippery.

I don't think it will. Even simple craft have unrealistically terrible performance. Some stock jets don't even outperform first-generation jet fighters. Even if we use stock craft as the baseline, the drag should be many times lower.

Edited April 4, 2016 by Pds314

[omelaw]

that'd make Kerbin's atmosphere more like Duna's at subsonic range, and I don't think low-tech rockets (like, mk.1 command pod and a Mk16 Parachute) can land safely with that low drag.

Edited April 4, 2016 by omelaw

[Pds314]

Accurate-weight, thrust, and size stock F-16 "replica" with no body clipping and minimal wing clipping. Fails to exceed Mach 1 even in a dive from 10,000 meters. Top speed regardless of altitude is 245 m/s on full afterburner. This is all with a clean configuration. No drop tanks or similar.



I remember a lot of nonsense about the F-35 having slower transonic acceleration at 10000 meters than some (very impractical) F-16 configurations awhile ago. Well, you can rest assured that this F-16 doesn't have ANY transonic acceleration because it can't reach transonic speeds.

It can fly at speeds as low as 30 m/s, and pull sudden 25-gee turns, but it can never reach mach 1. The real thing goes mach 2 at altitude and mach 1.2 at sea level.

1 hour ago, omelaw said:

that'd make Kerbin's atmosphere more like Duna's at subsonic range, and I don't think low-tech rockets (like, mk.1 command pod and a Mk16 Parachute) can land safely with that low drag.

So make the Mk16 have a bigger drag area? Like a real 'chute? Doesn't seem particularly problematic.

Edited April 4, 2016 by Pds314

[regex]

19 hours ago, Pds314 said:

#2. The FAR Firehound MS has a top level flight speed of 150 m/s dry, 200 m/s wet, even at altitude. Even diving from 10000 meters, I couldn't get it to go supersonic no matter what. Keep in mind that it supercruises in FAR. This means that it is moving slower than a real WW2 monoplane unless it is on full afterburner, in which case they're tied.

After I removed some of the excessive drag producers (forward facing control surfaces and those extraneous tanks on the side) I managed to get it supersonic.  It could probably go a lot faster with some additional tweaking that doesn't necessarily involve removing parts.

This should hopefully illustrate why you download and install FAR if you want a realistic flight model as those extra controls aren't really doing anything for you except slowing you down.  IIRC wing shape is assumed to be ideal at all times in nuStock so I don't believe there is any need for them (although they might help for the very slow speed flight?  Never tried.)

29 minutes ago, Pds314 said:

Accurate-weight, thrust, and size stock F-16 "replica" with no body clipping and minimal wing clipping.

And yet still has extraneous control surfaces on the front of the wing, big ones at that.

Edited April 4, 2016 by regex

[Pds314]

F-16 effective drag area at sea-level is like 4 m^2! By comparison, a real F-4 phantom has a subsonic drag area of 1.1 m^2, and it weighs 50% more than an F-16, and was known for its combination of bulk and drag, being called an "aerodynamic brick" on multiple occasions.

A real F-16 has about ~0.5 m^2 drag area. That's not zero-lift either. That's in subsonic cruise.

Edit: Apparently, there's an AeroGUI that tells me this in the menus. Stock now has a flight data menu. Thanks squad.

Edited April 4, 2016 by Pds314

[Pds314]

Made a reduced-surface version with each wing as 1 part and the vertical stabilizer as 3 parts. It gained an astounding 4 m/s and still has 8x the drag area it should.

Edited April 4, 2016 by Pds314

[Nich]

What are the L/D ratios in FAR?  I have been wanting to make a glider with 60:1 or 100:1 but sadly wings in stock only get up to 20:1 and increasing AR doesn't help

[regex]

8 minutes ago, Pds314 said:

Made a reduced-surface version with each wing as 1 part and the vertical stabilizer as 3 parts. It gained an astounding 4 m/s and still has 8x the drag area it should.

Huh.  I made this thing to test for another error (thinking I'd "kill two birds with one stone") and it goes way too fast (I can get it over 850m/s)...

All stock parts and configs, uses the afterburning engine.

Edited April 4, 2016 by regex

[kujuman]

My F-16 has no problems in stock :-) It's almost like I designed something based upon the conditions in which it'd operate. Needs better FBW though.

@regex your F-16 is one nice looking bird :-)

Spoiler

 

Used Input Pen to Scribe Post on Keyboard, Not Impressed.

I was going to post earlier, but I had some trouble--I wanted to respond with my tablet, but it's low on battery, so I have to use my desktop, and I'm forced to use my input pen on a physical keyboard.

Things I like about my stock keyboard:

---

Things I don't like about my stock keyboard:

#1. My writing speed has slowed to a crawl--I could get 80+ WPM on my pen using shorthand, but can only get ~5 WPM on this keyboard...it's so an unrealistic representation of my skill as a writer!
#2. I drank a lot of caffeine, and I can't respond to posts as fast as they are coming in. Even turning off my WiFi for 30 minutes, I couldn't keep up no matter what. Keep in ming that my input pen can do 80+ WPM on my tablet. This means that I'm writing slower than carving my own blocks to use in a printing press, unless I drink two Red Bulls, in which case I'm tied.
#3. I just tried to log-in to my bank, but the log-in page timed out due to how long my password took to enter. Again, no problem on my tablet!

 

TL;DR Use the tools presented to you the way they are meant to be used. For FAR, think in voxels. For stock, think in parts. It seems like you're claiming stock KSP is wrong because something you made to work with FAR isn't compatible...should one complain to Ferram that planes designed for stock don't work in FAR?

[Pds314]

4 minutes ago, regex said:

Huh.  I made this thing to test for another error (thinking I'd "kill two birds with one stone") and it goes way too fast (I can get it over 850m/s)...

All stock parts and configs, uses the afterburning engine.

How did you go that fast? Can it do that at all altitudes? That thing is nearly identical to mine and somehow gets reasonable speeds.

Gonna try wiping my aero configs and reinstalling KSP if it can do that near the ground.

Edited April 4, 2016 by Pds314

[regex]

3 minutes ago, kujuman said:

@regex your F-16 is one nice looking bird :-)

I copied OP's, but thanks.

3 minutes ago, Pds314 said:

How did you go that fast? Can it do that at all altitudes?

Here it is "on the deck", top speed:

Craft file: https://www.dropbox.com/s/m6d8mwqe96dmckn/Thing.craft?dl=0

Like @kujuman says, they are two different aerodynamic simulations and you have to build to them.  One is very precise, deals with the entire craft in a novel way, and was built by an honest-to-Johnny aerodynamicist.  The other is made for a light-hearted, goofy space simulation that focuses more on building wacky craft that overly enthusiastic LGM from the 1950's might build.  One makes you worry about area ruling and those additional control surfaces every plane needs while the other focuses more on getting you in the air simply and easily for playability.  One is for enthusiasts, the other is designed to be "good enough".

I think nuStock aero is pretty damn awesome for stock Kerbin and you can make some pretty cool things with it.  I also very much enjoy FAR for making my rockets fly like they should when I play on Earth.

[Pds314]

36 minutes ago, kujuman said:

My F-16 has no problems in stock :-) It's almost like I designed something based upon the conditions in which it'd operate. Needs better FBW though.

@regex your F-16 is one nice looking bird :-)

  Hide contents

 

Used Input Pen to Scribe Post on Keyboard, Not Impressed.

I was going to post earlier, but I had some trouble--I wanted to respond with my tablet, but it's low on battery, so I have to use my desktop, and I'm forced to use my input pen on a physical keyboard.

Things I like about my stock keyboard:

---

Things I don't like about my stock keyboard:

#1. My writing speed has slowed to a crawl--I could get 80+ WPM on my pen using shorthand, but can only get ~5 WPM on this keyboard...it's so an unrealistic representation of my skill as a writer!
#2. I drank a lot of caffeine, and I can't respond to posts as fast as they are coming in. Even turning off my WiFi for 30 minutes, I couldn't keep up no matter what. Keep in ming that my input pen can do 80+ WPM on my tablet. This means that I'm writing slower than carving my own blocks to use in a printing press, unless I drink two Red Bulls, in which case I'm tied.
#3. I just tried to log-in to my bank, but the log-in page timed out due to how long my password took to enter. Again, no problem on my tablet!

 

TL;DR Use the tools presented to you the way they are meant to be used. For FAR, think in voxels. For stock, think in parts. It seems like you're claiming stock KSP is wrong because something you made to work with FAR isn't compatible...should one complain to Ferram that planes designed for stock don't work in FAR?

The difference is that planes designed in real life DO work in FAR, whilst according to my tests, NO planes seem to work appropriately in stock unless there's something wrong with my configs.

I'm gonna see if I can replicate Regex's plane and its flight conditions. If not, I'm gonna assume this is a bug in my install.

EDIT: 

Definitely our aero is not identical. My first thought was that it had to do with the build #, but I can see that's not the case. And since Kujuman is using the Windows player as well, that's not the problem either. I'm gonna try reinstalling...

Edited April 4, 2016 by Pds314

[frizzank]

I don't understand. Is your complaint that the atmosphere is not realistic, or you just think its too thick.

[Pds314]

1 minute ago, frizzank said:

I don't understand. Is your complaint that the atmosphere is not realistic, or you just think its too thick.

That it is far, far too soupy and thick. I get that there are limitations of the stock model, but at least on my install, which is being reinstalled as we speak, it is so dramatically thicker than realistic that the only planes that can go supersonic in level flight are tiny RAPIER planes.