Help me troubleshoot this lander! (fuel flow does not respect lines)

[Colonel_Panic]

so this is a strange issue... I have fuel lines set up (in symmetry) basically like this:

radial tanks below the clamp-o-tron -> main tank below the clamp-o-tron -> radial tanks with engines below the clamp-o-tron -> clamp-o-tron -> upper stage main tank -> upper stage radial tanks with engines.

However, when I actually test this, the following happens:

radial tanks below the clamp-o-tron and ONE of the two radial tanks with engines below the clamp-o-tron begin draining immediately, the first ones at a faster rate (this should not ever happen. why is this happening?).

after the first tanks empty, both of the radial tanks with engines below the clamp-o-tron, and the main tank below the clamp-o-tron AND the upper stage main tank begin draining all at the same time, but one of the tanks with an engine drains faster.

after the tank with an engine runs out, the other tank with an engine, the main tank below the clamp-o-tron, AND the main tank above the clamp-o-tron drain, but the empty tank engine keeps firing, and the upper stage main tank empties faster than the rest

with the upper stage main tank empty, the upper stage radial tanks with engines begin draining too

If I disable fuel crossfeed in the clamp-o-tron, the lower stage weirdness mostly goes away, but the upper stage drains before the lower.

The way I need it to behave:

I need all of the lower stage tanks to drain before the upper stage ones begin draining.

I need the lower stage engines to fire until the lower stage tanks are all empty, and then stop firing.

I need the upper stage engines to always fire until everything is empty.

other requirement: there can be NO direct fuel linkages between the upper and lower stage (aka, must be able to assemble it in orbit, so nothing that needs to be added after the docking procedure to make it work.

other other requirements: why the bloody hell does this thing want to flip over so badly? it seems to try and turn the wrong way whenever I make course corrections while hovering. MJ can't handle it either. Could be related to the above problem with the upper tanks emptying first, putting the COM below the COL, but still, I don't think the engines are THAT high up....

.craft attached (stock).

http://herpnderp.info/Omni-Lander%20I%20s.craft

[Nao]

The problem: even without the fuel lines, there is already a fuel linkage between upper and lower part, through the docking ports that is, creating a fuel loop.

After playing puzzle fuel line connector for about half an hour, i found the "disable crossfeed" on the docking port senior. Welp that solves the problem in the craft you posted without changing anything

As for flipping, it wasnt that bad for me ... but i'm not sure about the gimbaling on LV-N's as they are close to CoM, thus provide almost no gimbal power (if they are above CoM they actually work wrong and can contribute to flipping) i would advise creating a control group to toggle the gimbal on them.

Cheers!

[Colonel_Panic]

see my above post: if I disable crossfeed, the upper stage drains before the lower. This should not happen.

The LV-Ns are the only engines with gimballing, so without them I have no control.

[stupid_chris]

He's right. The first problem is the dockingport. If you disable the fuel crossfeed on it, everything drains fine with your current fuel lines. Then its the gimbaling nuclear engines that set it off course because the CoM goes under the CoT. It's a known bug. There's really nothing else you can do but disable the gimbal and go with torque or change your design. I tested it with MechJeb and it went off course. Nothing goes off course with MechJeb excepted for CoT above CoM

[Nao]

see my above post: if I disable crossfeed, the upper stage drains before the lower. This should not happen.

The LV-Ns are the only engines with gimballing, so without them I have no control.

Oh ok i think i know now what are you trying to do. I jumped to fast on the epiphany of the crossfeed change.

I got some improvement but still not perfect:

1) put a pair of decouplers between LV-T30 and fuel tanks.

2) move the fuel lines from the main "32" tank to engines instead of tanks

3) create new lines from the side tanks (LV-T30 ones) to main "32" tank.

4) put the decouplers staging somewhere else so it doesn't get used.

Unfortunately after the side tanks get empty the fuel starts draining from both gray 2m tanks, (blocking the flow of fuel/oxy on the top one solves the issue)

With the current model of fuel flow it will be reaaaly hard to get it right without altering the design even harder.

EDIT: IT WORKS!!!!, just add:

5) another decoupler between main "32" tank and docking prt

6) fuel line between the 32 tank and decoupler

7) fuel line between decoupler and docking port. not needed actually YAY

I'll work on it a little more

As for gimbaling, you aren't getting any gimbal force from LV-N's anyway and if your CoM happens to be Below them they will work in reverse (struggling with lander can's momentum control), after you detached the bottom stage you can always toggle the gimbal on.

Edited May 31, 2013 by Nao

[Colonel_Panic]

Oh ok i think i know now what are you trying to do. I jumped to fast on the epiphany of the crossfeed change.

I got some improvement but still not perfect:

1) put a pair of decouplers between LV-T30 and fuel tanks.

2) move the fuel lines from the main "32" tank to engines instead of tanks

3) create new lines from the side tanks (LV-T30 ones) to main "32" tank.

4) put the decouplers staging somewhere else so it doesn't get used.

Unfortunately after the side tanks get empty the fuel starts draining from both gray 2m tanks, (blocking the flow of fuel/oxy on the top one solves the issue)

With the current model of fuel flow it will be reaaaly hard to get it right without altering the design even harder.

EDIT: IT WORKS!!!!, just add:

5) another decoupler between main "32" tank and docking prt

6) fuel line between the 32 tank and decoupler

7) fuel line between decoupler and docking port. not needed actually YAY

I'll work on it a little more

As for gimbaling, you aren't getting any gimbal force from LV-N's anyway and if your CoM happens to be Below them they will work in reverse (struggling with lander can's momentum control), after you detached the bottom stage you can always toggle the gimbal on.

Thanks! I'll rework these and post my results once I get back to the lander. I was really hoping to avoid the extra weight of decouplers to solve the fuel flow problem, as a heavy lander like this is crunched for weight as it is. I might see if there's anything else I can do to slim it down.

Well the thing is the cot -shouldn't- be above the com in this ship, unless every fuel tank except for the bottom 3 are empty (and it's well below it when fully fueled). I really don't understand how this gets flipped around when the lower tanks are (mostly) draining first... so I don't know why the gimballing would be broken on those engines. I can try disabling it, but I really don't think that lander can has anywhere near enough torque to keep the lander under control without some form of thrust vectoring.

[Colonel_Panic]

He's right. The first problem is the dockingport. If you disable the fuel crossfeed on it, everything drains fine with your current fuel lines. Then its the gimbaling nuclear engines that set it off course because the CoM goes under the CoT. It's a known bug. There's really nothing else you can do but disable the gimbal and go with torque or change your design. I tested it with MechJeb and it went off course. Nothing goes off course with MechJeb excepted for CoT above CoM

In all my experience MechJeb frequently decides he needs to do a backflip. just because. I never trust him for handling landers. :x