Service bay reentry

[swjr-swis]

On 12/16/2018 at 2:55 AM, JeyJeyKing said:

New player here, having the exact same issue.

Did you look at @Tyko's suggestion (and my video, which amounts to the same thing)? The answer hasn't changed since this question was asked - it works without fail.

Welcome to the forum btw.

Edited December 17, 2018 by swjr-swis
dutchisms

[Fraktal]

10 hours ago, JeyJeyKing said:

New player here, having the exact same issue.

As far as I can tell, it primarily seems to be a drag issue: the service bay has very high drag, which is what's causing the craft to flip by creating torque the moment the drag vector is not going exactly through the CoM (ie. the moment you are not pointed perfectly retrograde, assuming a perfectly symmetrical weight distribution in your craft). The higher the AoA becomes (ie. the further away from perfectly retrograde you are pointing), the further away the vector is pointing from the CoM, which means the greater the torque. The moment this happens, the low-drag front and the high-drag back are now both pointed into the airstream but since the drag distribution is asymmetric (one side has more drag than the other), the high-drag back (ie. the service bay) creates more torque than the low-drag front (ie. the command pod), resulting in the craft orienting itself to point the high-drag side backwards and the low-drag side forwards (ie. prograde) to reach dynamic equilibrium (ie. where any deviation in heading causes the service bay's drag to counteract and force the craft back to pointing perfectly forward), unless something like SAS or aerodynamic stability via fins/winglets is actively holding against it. Someone likened the relevant physics to a badminton ball: no matter what orientation you throw it in, it will immediately reorient itself into pointy end forward and finned end backwards because that's when drag is the lowest.

Attaching a heat shield helps because the heat shield has lower drag than the service bay, but it's not a perfect solution, it merely decreases the torque to the point where SAS can handle it at low AoA, but it will still flip without SAS actively holding against it because even at 69 km altitude and even if your deviation from prograde is less than 1°, the torque is still non-zero. Attaching more weight to the bottom helps by moving the CoM downwards and further away from the CoL, with the resulting aerodynamic stabilization partially or entirely counteracting the drag torque. Of course, this makes the top stage heavier, which in turn reduces dV at launch.

A different solution would be to deliberately create drag at the top of the command pod to stabilize it. Of course, this runs into the issues of 1) those parts burning off and 2) the front of the rocket having higher drag means the drag will now want to flip your rocket at takeoff instead, unless you have enough fins. I ran into this second problem while building a munar lander: by mounting four FL-T100 tanks (with nosecones on both ends) to the transfer stage radially instead of centered in order to reduce the rocket's height for less flexing when steering, those tanks create enough drag that flying more than 10° off prograde causes the rocket to flip as soon as the SRBs are dropped, despite being aerodynamically stable (ie. CoM in front of CoL). Doesn't appear to be related at first, but is actually the exact same problem: aerodynamic drag creating too much torque for SAS to compensate for.

Edited December 16, 2018 by Fraktal

[swjr-swis]

On 12/16/2018 at 1:43 PM, Fraktal said:

aerodynamic drag creating too much torque for SAS to compensate for.

Your explanation contains a lot of truth and insight... but have you looked at the video I posted? I explicitly sent the pod tumbling into reentry with SAS off, to show that when placed correctly (ie. with CoM as low as possible towards the bottom of the service bay) the craft will reorientate itself correctly, without needing SAS.

[Fraktal]

On 12/17/2018 at 5:17 PM, swjr-swis said:

Your explanation contains a lot of truth and insight... but have you looked at the video I posted?

I tried it, about an hour ago. I put everything on the bottom of the bay, did the offset-into-bottom-of-command-pod thing. Reenters fine with a heat shield, flips without a heat shield (and I was coming in at a very shallow angle). So I'm keeping the shield on. It's a highly inefficient suborbital rocket, so it's not like dropping one small fuel tank to not go over 30 parts is a critical shortcoming because it doesn't have the dV to stay in orbit either way (and if I bring the fuel tank instead, I have to lose some fuel to stay within 18 tons anyway).

Edited December 21, 2018 by Fraktal

[NerbalTheKerbal]

I was also having this issue. Turns out I had my heat shield under my decoupler, so of course it wasn't doing anything. I swapped their positions and reentry was once again a cinch.

[Fraktal]

Wouldn't it save ablator to have the decoupler stay on and burn off from the heat?

[Streetwind]

Sometimes the spacecraft gets torn apart instantly and entirely if a stack-mounted part at the bottom goes boom. Not always, but sometimes. I would consider any such sacrificial part a possible loss-of-mission risk, and avoid it like the plague.

Besides, I never bring ablator anyway, so there's nothing to save. I've found that any flight profile inside Kerbin's SoI, or even standard Hohmann transfers from Eve or Duna, can do direct reentries just fine with just the heatshield's own passive heat tolerance, which is extremely high. And for those rare few interplanetary returns that do get too hot for a direct reentry, a one-pass aerobraking to bring your apopasis inside Kerbin's SoI and looping around for reentry will do the trick.

(Your mileage with system rescale mods will certainly vary...)