Jump to content

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

r4m0n

By r4m0n
in KSP1 Mod Releases

 Share

Recommended Posts

Jethro Bodine

Jethro Bodine

Posted
Posted
On MechJeb and FAR/RSS/RO/whatever

First, topheavy is GOOD...

Hehe, I get the part about top-heavy being good, CoM in front of CoL, and all that jazz. However, when your dart or paper airplane lacks adequate control surfaces... top-heavy = uncontrollable tumble/stall the moment you attempt a maneuver. :P

That was the issue I was experiencing, though MechJeb (bless its heart) was trying desperately to keep me on a proper trajectory. So instead of the expected tumble, it just appeared as though MJ was grossly over-steering and/or turning prematurely.

I suppose I should have done a better job emphasizing the lack of control was the real culprit, not mass distribution. My bad. :)

Link to comment
Share on other sites
Jethro Bodine

Jethro Bodine

Posted
Posted
This is strange. The previous version had toolbar support this one doesn't seem to have it. I can neither add the individual dialogues to the toolbar. I have the side button.

Are you trying to add buttons via MJ or the Toolbar menu itself? I used the Toolbar addon's menu and it worked fine.

Link to comment
Share on other sites
Oinker

Oinker

Posted
Posted
This is strange. The previous version had toolbar support this one doesn't seem to have it. I can neither add the individual dialogues to the toolbar. I have the side button.

Just tested the new dev version. The toolbar integration is perfect. Slider button is out of the way (once you change it in settings). All buttons showed up on toolbar, even when main MechJeb window was hidden.

Toolbar setup: I didn't test whether the buttons were available for "configure visible buttons" when the main MechJeb menu was hidden. So if they're not, just click open the side menu and then "configure visible buttons" on toolbar. Give yourself the main MechJeb menu button on the toolbar and any other windows you want available. Change the MechJeb settings window to disable the menu tab.

Link to comment
Share on other sites
Commissar

Commissar

Posted
Posted (edited)

I think I hit a bug with the docking AP: it says my target size is 474, and the safe distance is 880m.... neither of which are correct.... so, it moves out of range of the docking port targeting. i'm running the stable release version. the target is maybe a hundred meters on the largest axis. i left it go, and it used up 600 monoprop..... and didn't even get to the docking port.

4hHSzhS.png

Edited by Commissar
Link to comment
Share on other sites
sarbian

sarbian

Posted
Posted

I need to work on that. I can see how the current implementation could give you twice the longest axis, but 4 times ... :(

I'll add something to set it by hand tomorrow

Link to comment
Share on other sites
Subcidal

Subcidal

Posted
Posted (edited)
On MechJeb and FAR/RSS/RO/whatever

Appreciate the thorough explanation I understand how gravity turns work in reality, but the rest of the info you gave here is really good. I'd suggest putting it on your RO thread.

Edit :

Figured i'd take a few screenshots of what i'm experiencing.

http://imgur.com/a/gptfu

The last two pictures have the ascent guidance target on, forgot to turn it on. Just showing which direction the target is and notice the lack of pitch input. It's like MJ isn't even trying to stay on course, just off doing it's own thing. Anybody have any clue what might cause this to happen? The rocket launches great manually, though looses a bit of control once passing mach 1. Nothing that should be causing this though.

Also notice how it makes an extreme veer to the south east almost instantly. It's like mechjeb thinks vertical is somewhere to the south east at about a 78 degree pitch. It only takes 3 seconds of flight for the rocket to end up pointing south at a 75 degree pitch. And this is at 160 meters, my gravity turn is set to start at 1.3 KM... Something is wrong but I have no idea what.

Edit #2: Just found that this problem seems to be exclusive to the AIES RC3 Maxur pod. Same rocket with different command pods works great. Maxur seems to not understand which way is 90* vertical. Just incase anybody else runs across the same problem.

Edited by Subcidal
Link to comment
Share on other sites
Commissar

Commissar

Posted
Posted (edited)
I need to work on that. I can see how the current implementation could give you twice the longest axis, but 4 times ... :(

I'll add something to set it by hand tomorrow

well, now, it gives me the correct size after I turned off the build that was going with EL, however, it now says the "safe distance" is 440 meters... I would consider this a pretty major bug, myself. I find manual docking to be a real PITA, i've never even done it completely manually...

however, it only seems to happen when the craft is on the other side of the docking port.... it took some manual and smartass finagling, but i got it to work...

Edited by Commissar
Link to comment
Share on other sites
sojourner

sojourner

Posted
Posted

Docking using build 216, couple of issues I've seen:

Upon completing the docking, the RCS thrusters on the docked craft will continue to fire. sometimes. Closing the docking AP is the only way to shut them off.

Sometimes the Docking AP will shut off just before the dock is made, this is made worse if you left some command active in Smartass, which immediately starts to re-orient the ship, thus ruining any chance of the magnets completing the dock.

Activating "Smart Translation" in the RCS Balancer causes the RCS thrusters to "lock up" at random during docking ap use. I've had them lock while firing causing the craft to overshoot the target and accelerate in whatever direction docking ap was attempting to move. I've had them lock and not fire when docking ap attempts to causing missed maneuvers. In all cases deactivating Smart Translation has allowed Docking ap to regain control.

Link to comment
Share on other sites
cheech420now

cheech420now

Posted
Posted

I cant get the dock ap to work at all. When i engage the ap it just backs up untill it loses target. I even got it aligned perfect and it still just backs up and loses the target. The ap from dev build 198 worked perfect. Just revert the docking ap to that build untill you can get the new one fixed.

Link to comment
Share on other sites
JenBurdoo

JenBurdoo

Posted
Posted

Updated my MechJeb to 23.5 and while it is working, I am unable to view Delta-V stats at all. The function is available but invariably reads 0 d/v. Is this a bug, will it go away if I reupload, or am I not using it right? I see there is now a function to change to different planets, though I've set it on Kerbin and still nothing happens.

Link to comment
Share on other sites
daawgees

daawgees

Posted
Posted
Updated my MechJeb to 23.5 and while it is working, I am unable to view Delta-V stats at all. The function is available but invariably reads 0 d/v. Is this a bug, will it go away if I reupload, or am I not using it right? I see there is now a function to change to different planets, though I've set it on Kerbin and still nothing happens.

You need to update to the latest dev build of MJ along with the provided .dll. Make sure it's a clean install of MJ as well.

Link to comment
Share on other sites
OrbitalDebris

OrbitalDebris

Posted
Posted

I just tried build 216 and it's just fantastic. Everything I tried worked flawlessly, thanks for the hard work Sarbian. I also found out that my RPM problems were related to a version check, updating to the latest dev build fixes those problems too. So I'm back to happily strapping little green men to dangerously overpowered rockets, and thanks to MJ putting them down safely within a couple meters of a flag planted in a planet millions of Km away :D

Link to comment
Share on other sites
Subcidal

Subcidal

Posted
Posted

No problem, And to be fair it wasn't something I looked up until I became interested via KSP.

This little rocket I was using has a TWR of 2.5 It's a bit extreme I know but I stopped messing with my design once i realized I had a mechjeb problem, Which was caused by AIES's Maxur pod for some reason that I still can't figure out.

BTW thanks for the link to the RSS wiki, Missed that somehow.

Link to comment
Share on other sites
Einarr

Einarr

Posted
Posted

I seem to be having issue with the Launch To Rendezvous option in the Ascent Guidance. I've had this issue since .23. It's not that MJ will not launch as I enable to AP, then click on LTR. The problem is that MJ is launching into the entirely wrong plane, and the resulting inclination is exatly opposite of my target. For example, my station is inclined at ~15 degrees. MJ tends to launch my ships at ~-15 deg. This results in crossed orbital paths that look like an X when viewed from the side. and not an oddly shaped X, either (aside from it being wide from the fairly shallow inclinations involved.

I've tried clearing out the phase angle textbox before launch, then launching into plane of my target and at it's altitude, then reloading a quicksave from just before launch to hopefully get the right phase angle info...it still doesn't work... I could use a more in depth guide to getting better results from LTR.

Link to comment
Share on other sites
Subcidal

Subcidal

Posted
Posted (edited)

Alright I dug a little deeper and my problems seems to be related to specific pods. I don't think it's mechjeb itself but rather some .cfg.

AIES probes all work fine except for Maxur, B9 capsules work fine, FASA works fine except for Gemini pod, The bumper nosecone from RftS pack works fine. Stock Airplane cockpits work fine as well.

All stock probes and pods (Spare MK1-2-3 cockpits) are affected and mechjeb has pretty much 0 control. They all do the same thing, launch, instantly veer to the east, south and occasionally north not following ascent path. Same exact rocket with a working pod (AIES, FASA Mercury, MK1 cockpit etc.) and the ascent is as smooth as butter. (Photo album of the problem http://imgur.com/a/gptfu )

Has anybody ran into anything like this before? Like I said I don't think this problem is mechjeb itself, but whatever is causing it somehow is affecting mechjeb, so i guess this is semi-related to the thread.

Edited by Subcidal
Link to comment
Share on other sites
Subcidal

Subcidal

Posted
Posted

So after a long day of searching i've narrowed this problem down to some .cfgs in realism overhaul. Pods_Squad/AIES Probes_Squad/AIES/AIES_New to be precise. Without those .cfgs any NEW ship i make performs correctly. If i load up a ship i made with those .cfgs enabled it retains the old behavior of not steering.

I have no idea why these configs would mess with mechjeb though. Nothing in them really pops out at me.

Link to comment
Share on other sites
Castun

Castun

Posted
Posted
So after a long day of searching i've narrowed this problem down to some .cfgs in realism overhaul. Pods_Squad/AIES Probes_Squad/AIES/AIES_New to be precise. Without those .cfgs any NEW ship i make performs correctly. If i load up a ship i made with those .cfgs enabled it retains the old behavior of not steering.

I have no idea why these configs would mess with mechjeb though. Nothing in them really pops out at me.

Are you sure they don't remove the reaction wheel torque? I know BTSM did this on all the stock pods and probes.

Link to comment
Share on other sites
JenBurdoo

JenBurdoo

Posted
Posted

They also don't show up on some of the ARM additions. That was my problem earlier. The two-engine launcher does not list D/V. The four-engine one does UNLESS a second stage is added above it, even if that second stage on its own appears in the deltaV indicator.

Link to comment
Share on other sites
BARCLONE

BARCLONE

Posted
Posted (edited)

Dev216, KSP 23.5...

Sorry to say, the LGAP is still wonky. Just had two Pathfinders slam into Moho at 500 m/s while using "Land Somewhere". MJ started the final burn much too late to be effective. When using "Land At Target", the ship starts oscillating +NML to -NML (around the RAD axis) when the delta-V remaining in the de-orbit burn reaches 0.1 m/s, burning in brief bursts as the ship crosses retro-grade, finally giving up and going into a spin around the RAD axis which it never seems to come out of. The ship just locks up with that 0.1 m/s remaining.

Interestingly, this same lander design functions perfectly with MJ at Mun and Minmus. It also appears to work properly at Duna.

The thought has occurred to me, is MJ being affected by FAR and/or thermal extremes here at Moho?

This error message is repeating itself throughout the logfile ad nauseum :

[ERR 08:58:42.712] Body::setMass: mass is 0.000000, should be positive!

The first instance I see comes within 200 ms of the launch, then proceeds to fill the remainder of the log with only a few entries from other items.

Spotted this error at multiple points in the flight (post-separation from carrier, pre-impact):

[ERR 11:01:18.745] MechJeb module MechJebModuleMenu threw an exception in OnDestroy: System.NullReferenceException: Object reference not set to an instance of an object

at MuMech.MechJebModuleMenu.OnDestroy () [0x00000] in <filename unknown>:0

at MuMech.MechJebCore.OnDestroy () [0x00000] in <filename unknown>:0

[ERR 11:01:18.747] Can't remove Part (Script) because PartBuoyancy (Script) depends on it

Significance?

Edited by BARCLONE
Link to comment
Share on other sites
Guest
This topic is now closed to further replies.
 Share
Go to topic listing
×
×
  • Create New...