Phantom (?) Yaw forces showing up on overlay, causing flat spins.

[SavageSS27]

Hi all,

As the title says, I'm having some issues with Yaw forces that I don't think should really be there (I could, of course, be wrong). I did some searching and didn't really seem to come up with anything. I'm using B9 Proc-wing, OPT, FAR and AJE (Many others as well, but these seem to be the relevant ones).

I've placed the engines individually after placing their respective intakes and I don't seem to be getting any unequal thrust. I haven't even gotten high enough to see if the engines flame-out evenly or not.

Please see the picture attached below and if you need any more info, let me know and I'll do my best to gather it.

It's probably worth mentioning that I haven't spent much time with aircraft in KSP and this is my first solid attempt at an SSTO. I know it's lacking sufficient orbital engines on the design right now, but I wanted to try to get the high altitude/speed flight mechanics down before trying to work on the orbital bits.

Aero Overlay

(Not sure if this album will show up properly. I also put in some bonus pics of the same phenomenon occurring on an SR-71 model from the MK2 Expansion mod)

Thanks in advance for your help, guys. This community has be a HUGE source of great info and mods. Keep being awesome!

Edited July 5, 2015 by SavageSS27

[capi3101]

Your picture's not showing up, just FYI.

You're using FAR? What are your stability derivatives doing as you ascend? Check 0k/Mach 0.35, 5k/Mach 0.85, 10k/Mach 1, 20k/Mach 3 and 30k/Mach 4. If any of the numbers turn red, we need to check some things. I'd wager good money that your N_ß parameter goes red at some point. Yaw issues usually warrant a bigger fin; shoot for one that's about 10-12% your total wing + tailplane area, with 25% of that area dedicated to rudder.

There's also been issues between B9 Procedural Wings and FAR; are you using the latest development version of FAR? If not, you should get a hold of it; I know that among other things the latest release version of FAR doesn't account for the bottom half of Mk2 Cargo Bays. A new version of B9 Procedural Wings came out just today, incidentally.

[Warzouz]

I recently experienced that with stock spaceplane. There was a yaw force at 15km. The plan has 2 rapier and one central TurboJet. I noticed that shuting down the TJ solved the problem.

This problem don't show on each ascent. It's quite rare. I didn't figure what created it.

[imaspacestation]

Intakes placed in symmetry mode.

There's apparently a bug that the mirrored part actually gets less airflow, thus less thrust for the corresponding engine, thus yaw...

[SAI Peregrinus]

A fixed album embed:

Javascript is disabled. View full album

Just FYI, you get this via:

[​imgur]t7jIh[​/imgur]

But I've inserted some zero-width spaces to make the forum print the actual square brackets, so you can't copy/paste it directly.

[Guest]

Intakes placed in symmetry mode.

I just tested your hypothesis with six different aircraft, all built using symmetry for intakes and engines. There was no differential airflow or thrust on any of them.

[capi3101]

There is a known bug where the first engine you place on your craft will hog any air from intakes that are on the craft at that point; Intake Build Aid is designed to fix that bug at the push of a button.

Not relevant to the OP's question - since he mentions that he's not getting up high enough to see flameouts - but that might be something for all y'all to look at if it's something with which you're experiencing problems.

A fixed album embed:

http://imgur.com/a/t7jIh

Just FYI, you get this via:

[​imgur]t7jIh[​/imgur]

But I've inserted some zero-width spaces to make the forum print the actual square brackets, so you can't copy/paste it directly.

I'll have to take a closer look at these when I get an opportunity; busy right now. At a glance all the derivatives look okay...makes me wonder what's going on. What happens with a stability analysis in the simulator?

Do you need instructions on how to make the simulator tab work? (Not being facetious here - I know what that damn thing does).

[ajburges]

There is a known bug where the first engine you place on your craft will hog any air from intakes that are on the craft at that point; Intake Build Aid is designed to fix that bug at the push of a button.

Not relevant to the OP's question - since he mentions that he's not getting up high enough to see flameouts - but that might be something for all y'all to look at if it's something with which you're experiencing problems.

I'll have to take a closer look at these when I get an opportunity; busy right now. At a glance all the derivatives look okay...makes me wonder what's going on. What happens with a stability analysis in the simulator?

Do you need instructions on how to make the simulator tab work? (Not being facetious here - I know what that damn thing does).

Looks like an issue with center of pressure (CoP) being ahead of CoM making the craft dynamically unstable.

CoP is center point drag forces act on. Keeping this behind CoM allows passive flight stability. Since CoP is not directly shared in game, most people use CoL as a proxy. It works for most cases, but since it doesn't track drag it doesn't give a complete picture.

You have intakes mounted fairly far forward. This will push CoP forward. At low speeds, your control features can compensate, but as velocity increases drag does as well. Eventually, the torque from drag forces overcomes the control authority. This imparts great rotational momentum to your craft and simultaneously kills speed as you stall. Enter the flat spin!

Moral of the story: robustly test craft with a lot of drag ahead of the CoM for control authority and/or passive stability or avoid the problem entirely by keeping intakes in the rear â…“ of your craft. Keeping CoP too far back does create overly stable craft though, but holding prograde with higher dynamic pressure is better that diverging from it.

Edit: wait, that album has two distinctly different craft between flight and SPH. Oh well without more evidence, CoP location is my best explanation for the observed effect.

Edited June 9, 2015 by ajburges

[SavageSS27]

SAI Peregrinus

A fixed album embed:

Thanks SAI, I'll have to remember that formatting for future use.

MailletC

Originally Posted by

imaspacestation

Intakes placed in symmetry mode.

I just tested your hypothesis with six different aircraft, all built using symmetry for intakes and engines. There was no differential airflow or thrust on any of them.

Spacestation/MailletC, I placed the intakes individually to avoid the known asymmetric flame-out bug, but I'll add the build-aid to my mod list so I can be sure everything is paired up correctly.

capi3101

Your picture's not showing up, just FYI.

You're using FAR? What are your stability derivatives doing as you ascend? Check 0k/Mach 0.35, 5k/Mach 0.85, 10k/Mach 1, 20k/Mach 3 and 30k/Mach 4. If any of the numbers turn red, we need to check some things. I'd wager good money that your N

ß

parameter goes red at some point. Yaw issues usually warrant a bigger fin; shoot for one that's about 10-12% your total wing + tailplane area, with 25% of that area dedicated to rudder.

There's also been issues between B9 Procedural Wings and FAR; are you using the latest development version of FAR? If not, you should get a hold of it; I know that among other things the latest release version of FAR doesn't account for the bottom half of Mk2 Cargo Bays. A new version of B9 Procedural Wings came out just today, incidentally.

Capi, I have the newest version of FAR (0.15.2 "Ferri"), but I'll go download the new B9 Proc wings now and recheck. I know the derivatives at 0k/0.35 are red in some cases, but I usually don't rotate until ~120m/s. That most likely means there's not enough horizontal wing surface for the weight of the craft, so that's probably something I'll have to address in addition to the vertical stab/rudder area.

As for the simulator tab, any help/advice is always welcome. I'm operating on the "green is good" principle, but a better understanding of WHY is never a bad thing.

ajburges

You have intakes mounted fairly far forward. This will push CoP forward. At low speeds, your control features can compensate, but as velocity increases drag does as well. Eventually, the torque from drag forces overcomes the control authority. This imparts great rotational momentum to your craft and simultaneously kills speed as you stall. Enter the flat spin!

Moral of the story: robustly test craft with a lot of drag ahead of the CoM for control authority and/or passive stability or avoid the problem entirely by keeping intakes in the rear â…“ of your craft. Keeping CoP too far back does create overly stable craft though, but holding prograde with higher dynamic pressure is better that diverging from it.

Edit: wait, that album has two distinctly different craft between flight and SPH. Oh well without more evidence, CoP location is my best explanation for the observed effect.

aj, I'll look into moving the intakes back as well. Not something I had considered before, but it certainly makes sense.

As far as the craft discrepancy goes, the last few pictures in the album are of a different craft, but illustrating the same issue. The first half of the album is the craft that I built.

Thanks again to everyone for your input thus far. I'll go make some tweaks and report back.

[capi3101]

Edit: wait, that album has two distinctly different craft between flight and SPH. Oh well without more evidence, CoP location is my best explanation for the observed effect.

No...FAR would pick that up a CoP issue pretty easily; most of the longitudinal statistics would turn red if that were the case. It does look like the OP needs to switch over to the dev version of FAR, though.

Capi, I have the newest version of FAR (0.15.2 "Ferri"), but I'll go download the new B9 Proc wings now and recheck. I know the derivatives at 0k/0.35 are red in some cases, but I usually don't rotate until ~120m/s. That most likely means there's not enough horizontal wing surface for the weight of the craft, so that's probably something I'll have to address in addition to the vertical stab/rudder area.

As for the simulator tab, any help/advice is always welcome. I'm operating on the "green is good" principle, but a better understanding of WHY is never a bad thing.

Make sure you're using the latest dev version; Ferri's got the aforementioned bug when it comes to cargo bays, and I'm seeing it in your cross-section curve (the green one).

Now for the simulator tab - I came across this page regarding aircraft stability while researching a phenomenon called Dutch Roll. About a third a way down the page there are some sine wave graphs - the section header is "Dynamic Stability". Dynamic stability happens to be what the simulator tab will show you - it shows you the tendencies of your plane after the initial, static conditions. So what you do is pick an axis you want to test (you're curious about your yaw, so use "r") and put in an initial condition - I usually go with "5" - for that axis. Mash go and watch FAR generate the dynamic stability graph; the particulars of it aren't as important as the graph's shape. If the amplitude of the oscillations decreases from left to right across the graph, your plane will have a tendency to stabilize along that axis. If the amplitude stays the same, you've got neutral stability. If the oscillations increase in amplitude with time, you've got negative stability on that axis. And if things just fly off from the get-go, you have negative stability. With negative stability conditions, you want to re-design your plane.

I have come across situations lately where the static stability was all cool and froody, but the plane wound up having negative dynamic stability. You can try to correct for it with SAS modules, but I wouldn't rely on SAS to be able to fix the problem.

Edited June 9, 2015 by capi3101

[SavageSS27]

Make sure you're using the latest dev version; Ferri's got the aforementioned bug when it comes to cargo bays, and I'm seeing it in your cross-section curve (the green one).

Just downloaded the dev release as recommended. Also updated B9 Proc wing. Haven't retested yet, as the cross-section curve sill looks like this: http://imgur.com/HkmmkE4

I think it might have something to do with the issue being discussed here: http://forum.kerbalspaceprogram.com/threads/97525-1-0-2-WIP-OPT-Space-Plane-v1-6-9/page65

I'll paste in the fix that they're suggesting and recheck the curve.

On a related note, do you know if the stock physics.cfg has any bearing of FAR's calculations? I remember patching the 1.0.2 aero back to 1.0.0 after seeing several complaints here about the increased drag and whatnot. Just wondering if I should be reverting that .cfg back to its' 1.0.2 specs.

Also, sorry about my link formatting. It's been a while since I was a forum user and the intricacies of text formatting have escaped me.

[Crzyrndm]

Just downloaded the dev release as recommended. Also updated B9 Proc wing. Haven't retested yet, as the cross-section curve sill looks like this: http://imgur.com/HkmmkE4

I think it might have something to do with the issue being discussed here: http://forum.kerbalspaceprogram.com/threads/97525-1-0-2-WIP-OPT-Space-Plane-v1-6-9/page65

I'll paste in the fix that they're suggesting and recheck the curve.

On a related note, do you know if the stock physics.cfg has any bearing of FAR's calculations? I remember patching the 1.0.2 aero back to 1.0.0 after seeing several complaints here about the increased drag and whatnot. Just wondering if I should be reverting that .cfg back to its' 1.0.2 specs.

Also, sorry about my link formatting. It's been a while since I was a forum user and the intricacies of text formatting have escaped me.

1) FAR completely removes all stock aero effects, so you can safetly ignore that. Temperatures may have an effect, but it should be relatively minor

2) The original yaw issue was probably due to this issue with B9 and the FAR release version where the port facing wings would be seen as notably larger than starboard facing ones. It is fixed for anyone using the FAR dev build

PS

Little forum tip: Link posts rather than pages. Which page a post is on depends on how many posts per page you have the forum set to show so your link doesn't actually take me the same place it takes you

Edited June 10, 2015 by Crzyrndm

[sal_vager]

Someone on the #kspofficial IRC channel had this, and traced it to FAR flight assist, so it's worth checking that.

[SavageSS27]

Crzyrndm,

1) Ok, that's what I had suspected, but good to have the reassurance on it.

2) I didn't notice any asymmetrical voxel distribution on the B9 wings with 15.2, but I'm updated to the dev build now, so all should be kosher.

I hadn't considered the posts per page issue, so I'll link to specific posts from now on, thanks for the tip.

Sal,

I'm not using FAR's flight assist on this. It seems to cause too many oscillations. Same can be said for MJ and standard SAS. It's all "manual/mechanical" trim for me!

After adding the drag cube code as referenced in my last post, the cross-section curve still looks messed up, so I replaced it with a regular fuel tank, the results are below. Looks like the cross-section issue is isolated to OPT?

Javascript is disabled. View full album

[capi3101]

I would say it's likely - K.Yeon mentions in the initial OPT thread that "I have not implemented the drag cubes for cargo bays because not much information was released about them, i have yet to figure it out. "

What's the part you've got immediately aft of that cargo bay/fuel tank? I'm looking at the image with the fuel tank and I'm trying to figure out why it shows your cross-section area going down there.

In the meantime, if you need a cargo bay there, you might try a mid-sized Mk3 bay; hopefully it wouldn't look like dreck...

[SavageSS27]

I would say it's likely - K.Yeon mentions in the initial OPT thread that "I have not implemented the drag cubes for cargo bays because not much information was released about them, i have yet to figure it out. "

What's the part you've got immediately aft of that cargo bay/fuel tank? I'm looking at the image with the fuel tank and I'm trying to figure out why it shows your cross-section area going down there.

In the meantime, if you need a cargo bay there, you might try a mid-sized Mk3 bay; hopefully it wouldn't look like dreck...

Yeah, that's why i tried implementing the drag cube myself as per this post in the OPT thread. It's possible I didn't paste the code in properly, so I'll go recheck that.

Immediately aft of the cargo bay is another fuel tank. From fore to aft it goes capsule, adapter, science lab, cargo bay, fuel tank (empty), engine mount.

On a side note, I finally remembered how to embed URLs in text! Go me!

- - - Updated - - -

Upon further scrutiny, it does seem like i had the drag cube code pasted in after the final bracket in the .cfg. However, moving it to the right place doesn't seem to have helped the cross-section at all.

[Crzyrndm]

FAR doesn't use drag cubes, that's a stock aero concern. If it isn't voxelising properly, ferram and/or the OPT author may need to make some changes (Bac9's PWings doesn't use enough colliders so FAR has to use the visual appearance with this option. I don't know if that's what's required here, but it may help)

Edited June 10, 2015 by Crzyrndm

[SavageSS27]

FAR doesn't use drag cubes, that's a stock aero concern. If it isn't voxelising properly, ferram and/or the OPT author may need to make some changes (Bac9's PWings doesn't use enough colliders so FAR has to use the visual appearance with this option. I don't know if that's what's required here, but it may help)

Ah, that would explain why the addition of the drag cube didn't do anything, lol.

Is that .cfg a MM patch? Do I just dump it in my GameData folder, or do I have to cut and paste that code somewhere?

[Crzyrndm]

Standard MM patch, you do need to change the part name though and I don't know if it'll help or if it'll cause even more issues. It would be better to have the authors sort it out properly

[SavageSS27]

Looks like OPT updated to include FAR compatibility recently. Hopefully that will fix the issues I've been having with it. Thanks again for everyone's help and suggestions!