Odd docking port behaviour...

[Cmdr. Arn1e]

32-bit install, only running MechJeb and Kerbal Alarm Clock...

So I've been messing about with huge docked craft for various tours and eventually a Jool-5 attempt, but on firing up the engines for the Duna test transfer, only two out of the four banks of engines have fired up... The info on the left hand side shows they are out of fuel... how can this be, I pondered, so went through all 18 docked ports to see if any of them weren't docked right... nope, they are all saying docked... All the Orange Rockomax tanks are full... And all have crossfeed enabled... (possibly a mistake on my part, as they seem to even suck the fuel out of the tug attached to the bottom...)

Some pictures may help illustrate the problem:

It's just so confusing why 2 of the fuel lines seem to be working, and 2 don't...

Luckily, it's opposite engines that are affected, at least in this instance, which means my testing can continue... and I will be trying to repeat the problem again in sandbox and then see if it happens in career too...

And this isn't high priority, just mildly annoying that I have to endure such long burn times lol!

[Red Iron Crown]

Sorry if these are obvious questions, but is there a chance the two fuel lines for those engines are reversed? Did you by some chance tweak the thrust down to zero on the non-working engines?

[Padishar]

Sorry if these are obvious questions, but is there a chance the two fuel lines for those engines are reversed?

Or perhaps the fuel lines are hitting a strut (or some other part) rather than the intended one...?

[Cmdr. Arn1e]

Sorry if these are obvious questions, but is there a chance the two fuel lines for those engines are reversed? Did you by some chance tweak the thrust down to zero on the non-working engines?

4-way symmetry on the fuel lines, so they should all be leading from the FL-T100 tanks to the smallest Rockomax tanks with the engines mounted on them... The engine worked fine while they got the small amount of fuel in the tanks they were attached to and the FL-T100's:

Just going to do a close-up inspection of the fuel lines, something might have gone wrong with them...

Edit: I can see nothing wrong with the fuel lines... no struts intersecting, connected to the right tanks, arows pointing the right way... As I said, I will be trying to recreate the problem, but once I have finished testing the current vessel...

Edited October 20, 2014 by Cmdr. Arn1e

[Red Iron Crown]

Curious. Are you able to link to a craft file of the core section?

[Cmdr. Arn1e]

Sure thing, hope you don't mind mediafire...

http://www./view/dx3g7kazp5xxh9d/HMS_Invincible_Mk_II_Overview.craft

Feel free to break it into sub assemblies for easier launch, or just hyperedit it into orbit... if you hyperedit, you may need to undock then re-dock the 4x3 external fuel tanks, as i'm not sure the docking port Jnr.'s will be docked as-is...

[Red Iron Crown]

I'll probably just test it on the pad, though it won't be for a few hours.

[Padishar]

A bit later (once I get home from work) I'll have a go at turning off all but one of the problem engines and then running a "Log Sim" operation on it using KER 0.6. This should dump considerable details of the fuel flow that is happening and should allow the problem to be found quite quickly...

[Cmdr. Arn1e]

I'll probably just test it on the pad, though it won't be for a few hours.

That should be fine - just be sure it's a save with the building damage off, or kiss your launchpad goodbye! hehe! And it will get round the problem of the tug being docked and having all it's fuel sucked up as well...

A bit later (once I get home from work) I'll have a go at turning off all but one of the problem engines and then running a "Log Sim" operation on it using KER 0.6. This should dump considerable details of the fuel flow that is happening and should allow the problem to be found quite quickly...

Thanks to you both... I get the feeling it could be just too many parts for my PC to handle, or something like that, but I think it's worthy of checking out... I have seen others using similar crafts, and no word of a similar bug/glitch/problem... at 891 parts (!!!) it's a bit of a CPU melter... in fact, I haven't had to turn the heating yet because my PC is nice and toasty heh >.<

[Padishar]

I've never seen a problem anything like this because a vessel has too many parts for the machine to handle. The only things that happen are the game slows down or crashes completely.

Well, after a bit of messing about (installing MJ, adding launch clamps because it collapses on the pad under its own weight, reorganising the staging etc.) I've managed to run a few tests on your craft and it certainly behaves a bit strangely though not in the way you describe the docked together version behaving. The small docking ports that attach the bottom tug to the bottom side fuel tanks aren't docked to the side tanks though this shouldn't be affecting anything. All of the puller engines seem to be correctly drawing fuel from their respective side tank stacks but only so far as the middle Jumbo-64. These drain first, followed by the top Jumbo-64s, then the FT-T100s and then the X200-8s. From looking at the vessel's tree structure dumped by KER it appears that the docking ports at the top of the bottom fuel tanks aren't connected properly to the fuel tanks (the fuel tanks are children of the docking ports but the attach nodes aren't set up correctly). Detaching and reattaching these fixes that issue, KER shows increased deltaV due to more fuel being available and the fuel is drawn from the bottom tanks on the pad.

However, the above doesn't do anything to explain your issue as all of the engines are burning correctly. Do you have a save game with the vessel in orbit with the engines misbehaving as you describe? I suspect the problem has been caused by the vessel tree rearrangement that happens when you dock vessels together but I would need to look at the in-flight save to see what the vessel tree looks like. If this is the case then it may be possible to work around it by setting the type of the top part of the vessel (with the cupola) to a station when it has nothing else docked to it (preferably at launch time) and make sure that none of the other vessels docked to it are set to station. This should ensure that the cupola remains as the root part of the overall ship.

Edited October 20, 2014 by Padishar

[Red Iron Crown]

I also could not reproduce the described behavior and observed the same drain behavior as Padishar. Fuel lines appear to be working properly, but the bottom orange tanks don't drain at all.

[Cmdr. Arn1e]

Thanks for that guys; I had noticed similar with my quick tests yesterday!

As you both found, it is a very complex ship, and this has possibly occured due to the way I made it (about 8 or 9 sub-assemblies, launched seperately then docked together in LKO - If the bottom tanks weren't set up right, then this goes part-way to explaining it...)

It reminds me of the other sub-assembly related glitch I found but didn't bother saying about before because I had forgot all about it (and found out how to correct it) - I had saved a LV in sub-assemblies, but glitched on save, wo whenever I clicked it in the sub-asembly menu it would instantly appear attached... where I had clicked it, meaning it was wayyyy off-centre, with one of the liquid-boosters embedded in the main ship I was trying to attach it to... Deleting the sub-assembly, going back to the ship it was originally on and savig it again solved that one

Anyway, I have gone back to the drawing board on the whole thing, in an attempt to make a much more interesting ship for Ziv to judge when I submit my next attempt

[Padishar]

Did you use SelectRoot when creating the sub-assemblies? The really strange part is that the side stacks appear to be each made of three identical craft (a tank and probe core with a port on each end) but only the bottom ones have the missing AttachNode between the top port and the tank. I would be very interested to see the sub-assembly file for the tanks because there should be no way of attaching one part to another without the attach node being correctly set up...