r4m0n

[PART, 1.0.2] Anatid Robotics / MuMech - MechJeb - Autopilot - Historical thread

Recommended Posts

Pictures of the "Jool Core 2"

screenshot531.jpg

screenshot532.jpg

screenshot538.jpg

The last image shows the position of the last, large inertial stabilizer just above the small fuel tank where the nuc engines attach.

This whole bit goes into orbit, is mated with a payload, and then sent on its mission. What's not seen in these pictures are the stages needed to get it into orbit. It does go into orbit reasonably easy with MJ on. A single refueling mission of twin orange tanks tops it off.

Edited by AstroRick

Share this post


Link to post
Share on other sites

screenshot532.jpg

In this pic it looks like you have fuel lines leading in and out of the radial orange tanks. If so, you have created "loops" in the fuel flow logic, something that KSP does not handle well. More than likely this is causing one or more tanks to empty out sooner than the others which in turn leads to an unstable craft with an off center of gravity. The craft is the problem. not mechjeb. Remove the fuel lines leading out to the radial tanks. Edited by sojourner

Share this post


Link to post
Share on other sites

Whatever it is you're trying to do with fuel lines, forget it and get TAC fuel balancer instead. I learned this the hard way too!

Share this post


Link to post
Share on other sites

Try locking the gimbal on some of the engines. I have some rockets that just locking the single engine in the middle drastically improves stability. Try that or locking three of the outer engines.

Check your tank levels when it starts going out of control. If they're uneven you do have a pipe routing problem. Instead of having the pipes flowing out from the center to the outer engines, connect from each outer tank to one of the outer engines. That will keep a balanced draw on the outer tanks, same as if you had an engine mounted directly to each tank. The out and back pipes at the top keep the outer engines from running out of fuel before the center engine, but the outer tanks will (should) empty first.

Fuel is always drawn first from the tank furthest away from the engine(s). I'd bet the way you have the pipes at the bottom is futzing with how the game decides which tank is farthest away. Change the bottom pipes to feed from the outer tanks and it should drain equally.

If each engine and tank stack is identical then you don't need pipes to balance fuel flow, each equal stack that's lit at the same time will always burn fuel the same.

This SSTO is loaded with such pipe loops to keep all the engines burning. Launched in .21 with MechJeb 2.0.9 on its default ascent path to a 70KM orbit. End circularization with just a little bit of fuel left in the lowest center tank. I just kept adding fuel tanks, and changing to more and more powerful engines until it'd make it up there without having to drop the stage and finish with the lander's engines.

9913456815_d5d592a365_z.jpg

Then I removed the loop-back pipes, rearranged the other pipes and decoupler staging to convert it to asparagus and sent it to Mun. Here's the lander that's on top of that massive pile-o-rocket. It also has looped pipes so I can pop off the side engines when required, instead of when their tanks are empty.

10034354184_54e7ec3dd1_z.jpg

Second manned Mun landing.

What really screws up the balance is when you make the loop go around in a circle instead of in and out. I tried a ship to Mun or Minmus once with four orange tanks in a parallel square arrangement (thanks to ReStock's bigger, badder quad coupler). I put pipes both ways around the tanks, forming a bi-directional circular loop. What happened is it drained all the fuel first from two tanks on one side instead of all equally or from two diagonally opposite.

The odd thing was it did the final burn and circularization to LKO that way just fine, but after refueling and attempting to do the transfer burn is when it decided to do the uneven drain. It quickly got reverted and the pipes removed.

Still fiddling away time with various experiments while waiting on the Duna to Kerbin transfer window. The ship has a 64 and four FL-T800 tanks full of extra fuel it won't need for the optimum transfer. I can take that out and dock the lander to the main core that has four 64's and four skippers. Those alone can push it all back to Kerbin with fuel left over after getting into Kerbin orbit. (Of course with even more left over if not pushing the weight of five redundant tanks and engines.)

Would be nice to be able to have MechJeb calculate an earlier transfer start and orbital insertion based on burning nearly every drop of fuel available except for what's in the lander's X200-16 tank. It can land with that and its single poodle, no problem. Screw efficiency! Jeb and Bill (or Bob, whomever is always second) would like to start home soon as possible. ;-) Even if it used most of the lander's fuel too, I have fuel depots at various altitudes around Kerbin, Mun and Minmus.

Edited by Galane

Share this post


Link to post
Share on other sites

um, I installed it, added the part (one of the parts looks like its still greyed out (even though I unlocked it in the R&D section, but i still put the one on, my question is that I do not see all the same settings listed like ascent autopilot and several others?? whats going on?

Share this post


Link to post
Share on other sites

its odd, in the science hierarchy, it shows like 3 or 4 different mechjeb moduals but i have researched them all, but only one actually is usable. the other one is greyed out. any thoughts?? I wanna use it lol. hopefully after the update from today doesn't throw it out of whack

Share this post


Link to post
Share on other sites
um, I installed it, added the part (one of the parts looks like its still greyed out (even though I unlocked it in the R&D section, but i still put the one on, my question is that I do not see all the same settings listed like ascent autopilot and several others?? whats going on?

you dont get full functionality all at once. it requires research of other RD nodes. The extra grayed out bits is because of extra files that need deleting. Read a few pages back for details

Share this post


Link to post
Share on other sites

All of the things required are researched, but what needs deleted? and with 419 pages, a link could be helpful, or a general idea as to were it was.

thanks

Share this post


Link to post
Share on other sites
Ok, something is strange.

For now : Edit your GameData/Mechjeb2/Parts/MechJeb2_AR202/part.cfg and replace this part :


MODULE
{
name = MechJebCore
MechJebLocalSettings {
MechJebModuleCustomWindowEditor { unlockTechs = flightControl }
MechJebModuleSmartASS { unlockTechs = flightControl }
MechJebModuleManeuverPlanner { unlockTechs = advFlightControl }
MechJebModuleNodeEditor { unlockTechs = advFlightControl }
MechJebModuleTranslatron { unlockTechs = advFlightControl }
MechJebModuleWarpHelper { unlockTechs = advFlightControl }
MechJebModuleAttitudeAdjustment { unlockTechs = advFlightControl }
MechJebModuleThrustWindow { unlockTechs = advFlightControl }
MechJebModuleRCSBalancerWindow { unlockTechs = advFlightControl }
MechJebModuleRoverWindow { unlockTechs = fieldScience }
MechJebModuleAscentGuidance { unlockTechs = unmannedTech }
MechJebModuleLandingGuidance { unlockTechs = unmannedTech }
MechJebModuleSpaceplaneGuidance { unlockTechs = unmannedTech }
MechJebModuleDockingGuidance { unlockTechs = advUnmanned }
MechJebModuleRendezvousAutopilotWindow { unlockTechs = advUnmanned }
MechJebModuleRendezvousGuidance { unlockTechs = advUnmanned }
}
}

by


MODULE
{
name = MechJebCore
}

This will change you AR202 to have all the module unlocked whatever your progression in the tech tree.

Then could you send me you /save/<nameofSaveGame>/tree.cfg and /save/<nameofSaveGame>/persistent.sfs (not sure of the name) so I can try to understand you bug.

Use dropbox or pastebin. Thanks

I think I am having the same issue, this is the Only mod i have installed. and its a fresh steam download from like 2 days ago.

I will try the above.

Share this post


Link to post
Share on other sites

What really screws up the balance is when you make the loop go around in a circle instead of in and out.

A loop is a circle. You should use a different term instead of "loop" to avoid confusion. Maybe use "stack" as that seems to be what you're doing with your ssto.

Share this post


Link to post
Share on other sites

I just pulled a patch by CodePoet in dev :

* Changed the fuelflow simulation to not stage if the new stage does not detach any parts and only adds new engines when there are already engine running.

This is to address a situation that is most obvious if you have an ion probe with seperatons attached to be used as breaking rockets on landing. In this situation the simulation assumes the seperatons should be fired as soon as the ion engine becomes available, and so the ion stage is assumed to have a high TWR, which results in mechjeb mis-timing maneuver nodes.

* Improved the landing predictions by better simulating parachute openings, using the predicted ASL of the landing site, better modeling the drag of parachute parts in various deployment states, deploying parachutes at a height above the landing site, rather than actual AGL, to prevent early opening when landing near mountains.

Edit : I had some hicup when testing this. Sometime the AP use the engine and land shorter than expected. BUT if you disable the engine after the deorbiting it lands with awesome precision (it fine tune when to open the chute). I am looking into why it does the bad engine correction.

Edited by sarbian

Share this post


Link to post
Share on other sites

I just wanted to say thank you for an awesome plugin. I love KSP but I am terrible at the more complex transfers and docking maneuvers. This plugin made it possible for me to enjoy designing and operate ships without constantly failing.

Share this post


Link to post
Share on other sites

Dev #128 does some strange burn on landing AP but I can't look into for now. Try it if you want to test the parachute landing precision, but be warned that you'll need to disable the engine before 10km to get the precision.

Share this post


Link to post
Share on other sites

Stupid Question; Has something drastic changed for the Transfer Autopilot? I used to be able to pick what planet I wanted and Mechjeb would try its hardest to get there... and when I try the Rendevous Planner OR autopilot to select something not in kerbins SOI, it tells me "Target not in SOI".... am I missing something simple or has the transfer orbit autopilot been removed?

On a side note; I'm noticing the bug with the AR202 case (the fully functional one) being grayed out. Amusingly, it is doing it to me even AFTER I tweaked the settings to allow me to use the damn thing. Even more amusingly, I WAS able to use the functional AR202.... until the next time I started up KSP. Luckily I left the functional AR202 in Sub-assemblies, but still... whats going on?

Share this post


Link to post
Share on other sites

rendezvous planner/autopilot is for meeting up with other craft in the same SOI, not for getting to other moons or planets. Use maneuver planner instead.

Share this post


Link to post
Share on other sites

Did a fresh install of the latest version. The mechjeb info tab only shows up for the mechjeb pod v2, not for the usual stock probe and crew parts. Any ideas for a fix?

Share this post


Link to post
Share on other sites

Guys, tell me please...

I cannon land on Duna. When I press Land at target, some actions may happens (sometimes not), but, anyway, I've got "Reentry simulation timed out" in autopilot status. The vessels stays on orbit, nothing is happening.

I use last dev build (128) and RT2 (but vessel is manned, and controllable, with "Local control").

Edited by MaxP

Share this post


Link to post
Share on other sites

Ihave a slight problem with MJ, I have KSP 0.22, MJ2.1 And for some reason MJ is not able to align the ship. Whenever i use SMARTASS, or maneuver planner to point any ship in particular direction, Ship starts to violently spin and turn and othervise chaotically move like crazy. so i have to do everithing manually.

If someone had same experience with MJ and found out the cause, or the solution, please share.

Share this post


Link to post
Share on other sites

KOCOUR this may be related to a bug in MJ when RT2 is active. Can you try with the dev build ?

MaxP you're sure the engine are on ?

Share this post


Link to post
Share on other sites
I just pulled a patch by CodePoet in dev :

* Changed the fuelflow simulation to not stage if the new stage does not detach any parts and only adds new engines when there are already engine running.

This is to address a situation that is most obvious if you have an ion probe with seperatons attached to be used as breaking rockets on landing. In this situation the simulation assumes the seperatons should be fired as soon as the ion engine becomes available, and so the ion stage is assumed to have a high TWR, which results in mechjeb mis-timing maneuver nodes.

This is the long awaited fix to prevent it from firing engines placed in a higher stage at the same time a lower stage is firing?

For example, start with an OKTO2, stick a nose cone on top of it, an RT-10 beneath it. Stick some sepratrons around the RT-10. Put a decoupler and another RT-10 beneath the first RT-10.

Arrange the staging so the sepratrons are stage 0, the decoupler and upper RT-10 are stage 1 and the lower RT-10 is stage 2. What will happen is the first RT-10 will burn, then the second RT-10 will light, followed a fraction of a second later by the sepratrons even though they're not in stage 1. (But mount the sepratrons to decouplers or have a decoupler in their stage and MechJeb runs the staging properly.)

Is that issue now fixed?

Share this post


Link to post
Share on other sites
I just pulled a patch by CodePoet in dev :

...

Edit : I had some hicup when testing this. Sometime the AP use the engine and land shorter than expected. BUT if you disable the engine after the deorbiting it lands with awesome precision (it fine tune when to open the chute). I am looking into why it does the bad engine correction.

I think it would be best to create a new "final decent with parachutes open" landing step, in order to get it to not try to to course correct once the chutes are open as it is imposible to control attitude from there on in. I am not sure if I will get time to do more work in this soon, but let me know if you are on it so we do not duplicate effort.

Share this post


Link to post
Share on other sites

MaxP you're sure the engine are on ?

Of course :D

One more thing on Duna: the blue marker jumps along the path, so the engines starts to jittering as the distance to target changes very quickly.

Edited by MaxP

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.