[1.3.1] Ferram Aerospace Research: v0.15.9.1 "Liepmann" 4/2/18

[Gordon Dry]

Regarding:

Quote

Tryfix Trajectories creating phantom aerodynamic force arrows

 

[Insanitic]

58 minutes ago, Maeyanie said:

Well, I have to admit, I've questioned FAR's drag number recently myself. I built a plane for a suborbital tourist hop mission, using a pair of AJE Su-27 engines, and roughly the same weight as a Su-27 (entirely coincidentally, they look nothing alike). Despite being significantly less aerodynamic (built for high altitude, stability, and hauling big SRBs) it was able to easily supercruise even on the "full drag, strict area rule" hard-mode setting.

Ferram said it's probably legit, and he certainly knows a lot more about this stuff than me, but I was under the impression supercruise was a lot harder to do.

If Ferram says it's legit, it's legit in my books too then! 

I can't imagine why my plane is a magical levitating UFO compared to the real thing though. 

As for the B9 Pwings L/D, does anyone have any experience with B9 pwings + FAR? If so, how is the L/D of the mod? I'm finding they provide literally no lift and I'm using the latest version of the mod  

[Maeyanie]

14 minutes ago, Insanitic said:

As for the B9 Pwings L/D, does anyone have any experience with B9 pwings + FAR? If so, how is the L/D of the mod? I'm finding they provide literally no lift and I'm using the latest version of the mod  

I used the B9 procedural wings in that plane and they worked alright. I didn't compare them to stock so not sure if it's the same, but, they did work.

https://www.dropbox.com/s/dxuhrtv1jdc0r20/Kerbal Space Program 2017.06.23 - 23.21.21.61.png

[Insanitic]

21 minutes ago, Maeyanie said:

I used the B9 procedural wings in that plane and they worked alright. I didn't compare them to stock so not sure if it's the same, but, they did work.

https://www.dropbox.com/s/dxuhrtv1jdc0r20/Kerbal Space Program 2017.06.23 - 23.21.21.61.png

Weirdly enough I'm getting decent lift with B9 at supersonic speeds (similar to yours) but at takeoff, I get 1:1 ratio. What is that plane's L/D at liftoff with B9? I know it's not the same as my concorde replica, but I'm wondering if it's just my wing design that's causing a crappy L/D at low speeds with B9 pwings.

Edited July 4, 2017 by Insanitic

[Starlionblue]

6 hours ago, Insanitic said:

Thanks for the informative post. Yeah I got mixed up on the L/D ratios since I was getting conflicting numbers from Wikipedia and that concorde site. 

Yes my plane is lighter than the concorde but the thing is, I don't need to resort to a high AoA (only around 5-10 degrees) to maintain my extremely high L/D of 22 at Mach 0.35 while the concorde needed to at subsonic speeds.

---

When you say AoA, do you mean angle of attack or pitch angle? What is the actual AoA readout? You can get this from debug menu AeroGUI or Ferram itself.

Screenshots of the plane in flight?

[Insanitic]

3 hours ago, Starlionblue said:

When you say AoA, do you mean angle of attack or pitch angle? What is the actual AoA readout? You can get this from debug menu AeroGUI or Ferram itself.

Screenshots of the plane in flight?

Here's one at takeoff, max loading, 144 tons:

https://ibb.co/noXqSv

Here's one subsonic, climbing

https://ibb.co/dQWH7v

Supercruise

https://ibb.co/jVj4nv

Landing, fuel almost empty

https://ibb.co/b7aASv

Do they seem correct? To me, they seem slightly off but then again, my replica isn't really a replica at this point.

Also, can you inspect the same plane rebuilt with B9 Pwings? I rebuilt it with the exact same shape but the coefficients are drastically different, with B9 having a very high drag coefficient.

B9 Pwings

https://ibb.co/iz5ASv

Stock Tweakscaled

https://ibb.co/dcPFua

 

Edited July 5, 2017 by Insanitic

[Starlionblue]

Cool pics. Lovely aircraft.

It makes sense that AoA is higher just after rotation, at low speeds. What also makes sense is that L/D is better with a lower AoA at high speed compared to a very high AoA just after rotation. The high AoA gives a lot of lift, but it also has a very detrimental effect on drag.

 Can you post a screenshot of the Ferram lift graph thingy for, say AoA 3 degrees?

 

[Maeyanie]

19 hours ago, Insanitic said:

What is that plane's L/D at liftoff with B9? I know it's not the same as my concorde replica, but I'm wondering if it's just my wing design that's causing a crappy L/D at low speeds with B9 pwings.

I took a few shots of another flight with a slightly modified plane; I wanted to try KSPWheels gear to see if they were less prone to flipping than the stock ones.

So remember when I said it had poor transonic design? I may have understated that slightly.

Just after takeoff, I levelled out and took a shot at low speed and altitude. The L/D was around 15 at an AoA of 6.2 degrees.

Second shot was at cruising altitude, high subsonic, with a L/D of 8.5 at AoA 13 degrees.

I then pushed through the sound barrier without needing afterburners, topping out at around a mach 1.15 supercruise, with L/D 5.3 and AoA 9.7. Didn't realize it was that steep until just now looking at the screenshot, I must've been cruising far too high for that design.

After dumping the SRBs and heading home, the lighter weight gave me a supercruise of mach 1.84 -- and still accelerating by the time I had to descend. Despite having the engines idle and following a 3 degree slope, I still was gliding at transonic speeds by the time I reached the KSC, and despite using the (admittedly small) airbrakes, had to go around.

No springy main gear flipping the plane over on landing though.

Edit: Just in case anyone is interested, here's the craft file. It does need quite a number of mods though. Not sure which ones exactly.

Edit 2: Looks like it needs B9, B9 PWings, Bluedog, Vanguard, Kerbal Foundries, Mk2Expansion, QuizTechAeroContinued, RealChute, RLA Stockalike, and SpaceY for parts. Also AJE, RealFuels, and (of course) FAR to work properly.

Edited July 5, 2017 by Maeyanie

[Insanitic]

12 hours ago, Starlionblue said:

Cool pics. Lovely aircraft.

It makes sense that AoA is higher just after rotation, at low speeds. What also makes sense is that L/D is better with a lower AoA at high speed compared to a very high AoA just after rotation. The high AoA gives a lot of lift, but it also has a very detrimental effect on drag.

 Can you post a screenshot of the Ferram lift graph thingy for, say AoA 3 degrees?

 

By the Ferram lift graph, do you mean the tab with the data+stability derivatives? I'm not sure how to change the AoA setting in that tab, it stays at 14 degrees.

Here it is

https://ibb.co/jsUVCv

And here;s the transonic design tab

https://ibb.co/ewqyza

[Starlionblue]

1 hour ago, Insanitic said:

By the Ferram lift graph, do you mean the tab with the data+stability derivatives? I'm not sure how to change the AoA setting in that tab, it stays at 14 degrees.

Here it is

https://ibb.co/jsUVCv

And here;s the transonic design tab

https://ibb.co/ewqyza

 

I mean the Transonic Design tab graph. Doesn't look too bad.

10 hours ago, Maeyanie said:

I took a few shots of another flight with a slightly modified plane; I wanted to try KSPWheels gear to see if they were less prone to flipping than the stock ones.

So remember when I said it had poor transonic design? I may have understated that slightly.

Just after takeoff, I levelled out and took a shot at low speed and altitude. The L/D was around 15 at an AoA of 6.2 degrees.

Second shot was at cruising altitude, high subsonic, with a L/D of 8.5 at AoA 13 degrees.

I then pushed through the sound barrier without needing afterburners, topping out at around a mach 1.15 supercruise, with L/D 5.3 and AoA 9.7. Didn't realize it was that steep until just now looking at the screenshot, I must've been cruising far too high for that design.

After dumping the SRBs and heading home, the lighter weight gave me a supercruise of mach 1.84 -- and still accelerating by the time I had to descend. Despite having the engines idle and following a 3 degree slope, I still was gliding at transonic speeds by the time I reached the KSC, and despite using the (admittedly small) airbrakes, had to go around.

No springy main gear flipping the plane over on landing though.

Edit: Just in case anyone is interested, here's the craft file. It does need quite a number of mods though. Not sure which ones exactly.

Edit 2: Looks like it needs B9, B9 PWings, Bluedog, Vanguard, Kerbal Foundries, Mk2Expansion, QuizTechAeroContinued, RealChute, RLA Stockalike, and SpaceY for parts. Also AJE, RealFuels, and (of course) FAR to work properly.

 

What I'm noticing is that your pitch angle and AoA are equal. This indicates that you are deriving all your lift from AoA and nothing from the shape of the wing. Think paper airplane. Is there any way you could camber the wings and/or give them an angle of incidence? Most real world designs have angles of incidence around 2 degrees.

I don't know how much improved area ruling would decrease wave drag, but it might be worth a shot to optimize this.

[Insanitic]

10 hours ago, Maeyanie said:

I took a few shots of another flight with a slightly modified plane; I wanted to try KSPWheels gear to see if they were less prone to flipping than the stock ones.

 

Can you post a screenshot of your stability derivatives? I'm not convinced that the B9 wings are generating propery lift in any aircraft since I too require a high angle of attack and my pitch angle become equal with my angle of attack when I build my concorde wings out of B9 pwings.

[Maeyanie]

3 hours ago, Starlionblue said:

What I'm noticing is that your pitch angle and AoA are equal. This indicates that you are deriving all your lift from AoA and nothing from the shape of the wing. Think paper airplane. Is there any way you could camber the wings and/or give them an angle of incidence? Most real world designs have angles of incidence around 2 degrees.

I don't know how much improved area ruling would decrease wave drag, but it might be worth a shot to optimize this.

Yes, it's a flat wing. The Mk2 body is fairly flat too, so I'm kind of just using the whole body as a wing and relying on AoA. Plus I've had bad experiences in the past with AoI and wingtip SRBs.

Not claiming that's the best way to do things, I just kind of threw it together for a contract. It's not one of my better planes and I freely admit this. And yet, despite being terrible, it can still trivially supercruise, when a real-world plane, with the same engines and a similar weight, but optimized by a professional aircraft designer, can't...

 

2 hours ago, Insanitic said:

Can you post a screenshot of your stability derivatives? I'm not convinced that the B9 wings are generating propery lift in any aircraft since I too require a high angle of attack and my pitch angle become equal with my angle of attack when I build my concorde wings out of B9 pwings.

Sure, here it is (at the default mach 0.35, which is slower than it can actually go) with gear down, and with gear up. I just noticed those are identical. I'll have to check the debug voxels next time I'm there, the gear might be broken, but, that's not related to the current issues.    

To test your theory, I slapped together a test plane with stock boards for the left wing, and a B9 PWing for the right wing. I tried to make them as identical as possible, and, it didn't look good in the VAB; the CoL was offset to the left, so the stock wings were likely producing more lift. Should be close enough to be flyable, though, just have to keep some left stick.

It took off with some mild skidding around on the runway, and despite the CoL indicator, I actually had to hold RIGHT stick to fly straight. Here's another shot with the data window and aero indicators.

Edited July 6, 2017 by Maeyanie

[Insanitic]

19 hours ago, Maeyanie said:

Yes, it's a flat wing. The Mk2 body is fairly flat too, so I'm kind of just using the whole body as a wing and relying on AoA. Plus I've had bad experiences in the past with AoI and wingtip SRBs.

Not claiming that's the best way to do things, I just kind of threw it together for a contract. It's not one of my better planes and I freely admit this. And yet, despite being terrible, it can still trivially supercruise, when a real-world plane, with the same engines and a similar weight, but optimized by a professional aircraft designer, can't...

 

Sure, here it is (at the default mach 0.35, which is slower than it can actually go) with gear down, and with gear up. I just noticed those are identical. I'll have to check the debug voxels next time I'm there, the gear might be broken, but, that's not related to the current issues.    

To test your theory, I slapped together a test plane with stock boards for the left wing, and a B9 PWing for the right wing. I tried to make them as identical as possible, and, it didn't look good in the VAB; the CoL was offset to the left, so the stock wings were likely producing more lift. Should be close enough to be flyable, though, just have to keep some left stick.

It took off with some mild skidding around on the runway, and despite the CoL indicator, I actually had to hold RIGHT stick to fly straight. Here's another shot with the data window and aero indicators.

So your plane seems like it has the same problem with my concorde when I use B9 pwings. The coefficient of drag is enormous, even greater than the coefficient of lift, which makes no sense since you have wings that are fairly large. Shouldn't the Cd be lower than the Cl?  Someone correct me if I'm wrong.

I think that's why your plane requires such a high AoA to maintain stable flight. Could you try rebuilding your SRB equipped plane with stock wings? even tweakscaled if need be? If you're willing that is. I'm curious to see how the Cd and Cl change wrt your plane in particular if you use stock wings.

I might be completely out of the loop with this assumption, but I think B9 pwings is still not fully compatible with FAR.

Oh and thanks for the aerodynamics help Starlionblue. Good to know my planes aren't just flying metal bricks in the sky!

Edited July 7, 2017 by Insanitic

[Maeyanie]

18 hours ago, Insanitic said:

Could you try rebuilding your SRB equipped plane with stock wings? even tweakscaled if need be? If you're willing that is. I'm curious to see how the Cd and Cl change wrt your plane in particular if you use stock wings.

I might be completely out of the loop with this assumption, but I think B9 pwings is still not fully compatible with FAR.

Well, I didn't feel motivated to rebuild that one (since it wasn't very good anyhow), but the next plane I built, I made two versions, one with B9 and one with normal wings.

In a post-takeoff level(ish) flight comparison, they had similar lift, but the B9 one had 172 kN drag, where the stock one had 41 kN drag. The planes and conditions weren't quite identical, sure, but that's a pretty dramatic difference.

[Drew Kerman]

I have links to a few references to check out but I keep forgetting what all the derivatives stand for. I've seen them explained before but still have not enough practice to remember. I do recall tho that most derivatives had general solutions that were good for most situations - such as "add more dihedral" or "increase vertical stabilizer size". Has anyone put together a chart or something highlighting derivative problems and their common solutions? Am I completely off my rocker in suggesting this?

[Insanitic]

4 hours ago, Maeyanie said:

Well, I didn't feel motivated to rebuild that one (since it wasn't very good anyhow), but the next plane I built, I made two versions, one with B9 and one with normal wings.

In a post-takeoff level(ish) flight comparison, they had similar lift, but the B9 one had 172 kN drag, where the stock one had 41 kN drag. The planes and conditions weren't quite identical, sure, but that's a pretty dramatic difference.

Alright, that's fine. 

Yeah this confirms my suspicion. I really think interactions of B9 Pwings with FAR are still not fully ironed out. I'll post the issue on github for both B9 Pwings and FAR.

Unless this is a known issue already? I've looked around the B9 and FAR forums and haven't found anyone addressing this excessive drag issue with B9 Pwings

 

Interestingly enough, the lift force on stock is slightly lower than the Pwings, but it isn't as dramatic of a difference when comparing the drag forces.

Edited July 7, 2017 by Insanitic

[Rodhern]

On 7/7/2017 at 10:19 PM, Drew Kerman said:

I have links to a few references to check out but I keep forgetting what all the derivatives stand for. I've seen them explained before but still have not enough practice to remember. I do recall tho that most derivatives had general solutions that were good for most situations - such as "add more dihedral" or "increase vertical stabilizer size". Has anyone put together a chart or something highlighting derivative problems and their common solutions? Am I completely off my rocker in suggesting this?

Hi Drew.

You are definitely not off your rocker with this suggestion. In my opinion one of the greatest difficulties in using the derivative numbers is that they are not all that intuitive in their current implementation / interpretation.

I haven't really played that much Kerbal for quite a while, but I did 'update' the derivative interpretation in a private fork (Github); I haven't added a fix for the annoying change in derivative results at recalculation though. The update will not make it to the official FAR builds (Ferram has looked at my formulas and they seem to contradict the formula interpretation conventions of historical real-world aerospace literature).

If you are interested, please let me know if you want to give the 'alternative' derivatives a try. And if they make sense to you, then at least we will be two, and we could maybe try to create a cheat sheet, chart or something.

[Drew Kerman]

6 hours ago, Rodhern said:

If you are interested, please let me know if you want to give the 'alternative' derivatives a try. And if they make sense to you, then at least we will be two, and we could maybe try to create a cheat sheet, chart or something.

not really, I'm going to stick with what FAR gives me. Further consideration on this issue is that it serves me better to figure this stuff out as I go, since the KSA/C7 is not supposed to be experienced in aircraft construction. Maybe I'll be the one to whip up a chart some day...

[Rodhern]

Hi @kbios and @Virindi .

I have a new idea how we can maybe debug the wildly changing stability derivatives (#177). I have written a snippet that can be used to output FAR Wing Interaction Influences data to the log file (just some text lines).

In my scenario what is happening is that after the first load I have a good number of "nearby wing modules". After another few (two in my case) loads there are no "nearby wing modules" left in the lists. If this happens in your games too, then Ferram can try to see if it still does not happen in his games.

A compiled version of the script/mod (compiled against KSP 1.3 but seem to work for KSP 1.2.2 as well) is here: FerramDebug ver 1 .

License is all rights reserved. Source code is here, and source for the F-sharp core is on Github.

When in the SPH with a relevant aircraft loaded press F1 (I think F1 is also the screenshot key, but that didn't pose a problem to me) and a line for each wing interaction influence value is written to KSP.log (and trace listeners).

[evileye.x]

Hello @ferram4!

Is there anywhere a guideline how to add configs to control surfaces? In specific, I want to add config to ven's stock revamp basic canard.

For now, I just copied from your stock AV-R8 Winglet, but ven's basic canard is way smaller and I want to ask what is FAR analogue for relative surface area?

Thank you in advance.

[Gordon Dry]

@evileye.x:

This is what I found: https://github.com/ferram4/Ferram-Aerospace-Research/wiki/Deriving-FAR-values-for-a-wing-using-Blender-2.7

[evileye.x]

3 hours ago, Gordon Dry said:

@evileye.x:

This is what I found: https://github.com/ferram4/Ferram-Aerospace-Research/wiki/Deriving-FAR-values-for-a-wing-using-Blender-2.7

Oh, it's bit too complicated to learn it for 1 small part... I guess I'll just make basic canard with B9 proc control surface

[Mr. Sandman]

I'm having troubles with FAR and as a guy suggested me, I'll post the link here (because I don't want to fill an entire page).Disclaimer: It's probably just my fault and FAR is just fine.Thanks for the solutions.

Edit:

Managed to get the CoT down by lowering the thurst in the Solid motors for parts jettison.Going to test it now.

Edited July 20, 2017 by Mr. Sandman

[Psycho_zs]

A couple of observations with dev version on KSP 1.3:

Mk12-R drogue is very overpowered, 2xMk12-R slow down Mk1 capsule better than 1xMk16.

1.25m Heat shield handles like a piece of paper: it falls down at the same speed as Mk1 capsule under fully deployed Mk16 cute and starts zipping around in loops if tilted over.

[capi3101]

On 7/24/2017 at 2:01 AM, Psycho_zs said:

Mk12-R drogue is very overpowered, 2xMk12-R slow down Mk1 capsule better than 1xMk16.

Seeing the same thing here:



Not sure if this is really a bug or not, though - just corroborating an observation. I might try removing the Mk-16 entirely on the same craft setup to see just how much of an effect the drogues are having on their own.

Edited July 25, 2017 by capi3101
had a thought for a diagnostic test.