this rocket/plane flips with no apparent reason

[king of nowhere]

I am trying to build a mobile base that doubles as ascent vehicle for an eve-like planet. basically, it's a big plane with a large living space, it then drops the living space, flies up with propellers above the atmosphere, drops the propellers and wings and ascends as a rocket.

unfortunately, even though the shape looks fine, it keep trying to flip.

this is the full plane with the base part. not really relevant here. this thing flies passably, has some issues, but i'll improve it after I can get the orbiter to work.

 

to start orbiting, I drop the living space and this part flies up on propellers. it flies well, it is very stable. However, if I steer too much away from prograde, this thing flips and cannot be recovered. it appears to be aerodinamically unstable.

I would blame the wings being that much forward, but the center of lift is in the right place. doesn't look right at all.

Anyway, if I don't do any sudden turn, I can fly up nicely. The flat part behind looks ugly and drags heavily, but it doesn't create any real issues. The really bad part starts when i ditch the wings

 

This is the rocket that ensues. I removed the fairing to see it better. Well, this thing flips immediately. The flaps in the back are a latest addition to try and prevent that, but they do no good.

Ok, I can see that the center of mass is shifted in the back of the rocket, which can create problems. on the other hand, I have a large flat surface, which should make drag. the front is very aerodinamic. and I have just added a bunch of flaps. I also have a powerful engine (it needs to be, the planet has 2.2 g) with a lot of gimbaling.

I suppose I could try to fix the center of mass, but it's chancy. The upper stages are meant to be discarded from the top, so I can keep the engine (a necessity to get the high deltaV and thrust required to leave Derbin); but they are used in vacuum. The first stage, with the vector, is dropped while still within a significant atmosphere; I'd really not have to tilt the rocket and discard a tank from the front in that condition - plus, I'd lose the aerodinamic tip. Aside from that, I have done what I can to minimize the center of mass issue. No good; I have launched a lot of horrible payloads without fairings, and this looks to be the most unstable of all.

 

At some point I discovered that the fairing wasn't working properly and the things inside were still causing drag. I fixed the issue - the screenshot below shows, I am selecting a bunch of stuff inside the fairing to check that their drag is 0. still no good.

that rockomax tank in the back has an abnormally high drag; it's the same drag of the backwards fairing, which has an open node in the back. so it's like that tank has an open back. which it hasn't. I can't understand what's causing this, but it's drag behind the center of mass, so it should actually help with not flipping.

this final shot shows the moment I release the wings. the rocket flips upwards immediately. resistance is still 0 inside the fairing, staging didn't mess with it. yet there is one big red line starting inside the fairing, as if there was something actually messing up that could be responsible. however, I selected each and every part inside that fairing, and couldn't find anything with any drag.

I suppose I could try to move the CoM forward, though - as I explained - it would cause a lot of other problems. I'd rather not have those problems. This rocket really should not flip

[Lt_Duckweed]

Fairings apply all their body lift and drag at the base of the fairing, so that front fairing facing backwards is applying most of your fuselage lift and drag far forwards of the CoM.

As for the high drag of the fuel tank generating high drag.  You mention that the tank doesn't have an open back, but the thing is, it does.  Attaching things to nodes reduces area of the associated drag cube face based on the size of the drag cube of the part you attached.  In the debug data in your screenshot, the drag vector is (0.1 , 1 , 0) which means YP is the frontal face.  This has a OccA of 0, and a WDrg of 0, so all good there.  But your rear face, the YN face, has a OccA of 3.73 and a WDrg of 0.92.  Comparing the OccA of 3.73 to the WArea of 4.85 means that the attached vector reduced the OccA by only 1.12 (4.85 - 3.73).  On top of that, a WDrg of 0.92 is really bad.  That's pretty much indistinguishable from a flat plate.  Thankfully this is a rear face, so by the time you are approaching the speed of sound it will reduce quite a bit (rear face drag falls off substantially above ~Mach 0.9)

 

EDIT: to add on to this, I would highly recommend installing the mods "CorrectCoL" (take all parts into account for CoL indicator, not just wings, and also gives stability graphs) and "RCSBuildAid" (see separate indicator for dry CoM) as they provide invaluable help when designing planes.

Edited May 9, 2023 by Lt_Duckweed

[king of nowhere]

30 minutes ago, Lt_Duckweed said:

Fairings apply all their body lift and drag at the base of the fairing, so that front fairing facing backwards is applying most of your fuselage lift and drag far forwards of the CoM.

oooh! once more, the responsible is a glitch in how the game handles aerodinamics.

I tried changing place on that fairing. indeed, the situation improved significantly, though the rocket still flipped. now that I understand the problem, I can go back improving on my own.

Quote

As for the high drag of the fuel tank generating high drag.  You mention that the tank doesn't have an open back, but the thing is, it does.

I do not doubt your analysis, but the thing is, there is an aerodinamic fairing that should cover the back there, and a vector engine attached on the back node, and I checked that the vector also has 0 drag while the back fairing is there.

[Lt_Duckweed]

7 hours ago, king of nowhere said:

I do not doubt your analysis, but the thing is, there is an aerodinamic fairing that should cover the back there, and a vector engine attached on the back node, and I checked that the vector also has 0 drag while the back fairing is there.

The vector will indeed be shielded by the fairing.  But the fuel tank is not "in" the fairing, nor is it node attached to the fairing.  Thus, from the perspective of the fuel tank, the fairing might as well not exist.

As for the vector being node attached, keep in mind it only reduces the exposed area commensurate with the size of the vector.  Since the vector is about 1/2 the radius of the fuel tank, it reduces the exposed area of the fuel tank rear face by about 25%

[king of nowhere]

6 hours ago, Lt_Duckweed said:

The vector will indeed be shielded by the fairing.  But the fuel tank is not "in" the fairing, nor is it node attached to the fairing.  Thus, from the perspective of the fuel tank, the fairing might as well not exist.

 

I think I'm going to redesign the thing entirely, without any kind of fairing.