Jump to content

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


r4m0n

Recommended Posts

Did some some searching, couldn't find anything. Was wondering if there was a way to see the g force applied to a vessel through the mechjeb UI? I want to see actual numbers, not just the gauge beside the navball. Any help?

Edited by Dreadnaught426
Link to comment
Share on other sites

I noticed an issue with MJ Delta-V info.

I've a quite big rocket

- Lower stage is a radial asparagus

- Middle stage has droppable lateral tanks

- Third stage (lander) has it's fuel tank locked (so DV = 0m/s)

When in VAB, the DV si calculated at 8460m/s. When on pad, the DV si calculated at 17780m/s. Why such a difference ?. I think the correct value is 17700m/s, as I did a trim to Moho and came back with 5500m/s left (not counting the locked fuel tank for landing)

be18133e-3157-4ff8-b251-aa6250deeaec.jpg

On the pad

9c74b0a2-398c-485a-94c7-257220f2f8eb.jpg

Notice that KER gives mes 13500m/s both on VAB and PAD (which is not good either...).

Link to comment
Share on other sites

I can't see those image behind my evil work proxy are the info on the other stage changing too ? The new MJ code handle staging differently than the old MJ code and it could be that. But I don't understand why it would give such different numbers. One of the other explanation would be that when in flight in now display the dV for the current atmo even in the vacum column and in the Editor it properly display vacum/atmo as different values.

Link to comment
Share on other sites

here are the vaccum stats I get

Stage	MJ/VAB	MJ/Pad	KER/VAB	KER/Pad
0 0 0 0 0
1 0 0 0 0
2 1984 3097 0 0
3 0 3699 0 0
4 866 4245 0 0
5 0 0 6743 1949
6 0 1960 1960 1746
7 2459 2828 2828 2520
8 0 724 724 645
9 2017 92 92 497
10 1134 1134 1169 647
Total 8460 17779 13516 8004

I said that KER was 13500 each, but I was wrong. The correct value should be MJ/Pad, but even though the deltav for each stage doesn't seems to be exact. for example, it says that the stage 9 duration is only 2s. That is very wrong.

The rocket is built using 2 sub-assemblies (the orbital stage and the interplanetary stage)

I can't post the craft file now, but I'll try at home.

Link to comment
Share on other sites

@sarbian Do you have any thoughts abouthe what could be causing my control problems and/or how I might be able to get around them (suggested settings for attitude adjustment page)? I don't like to bug, but I'm currently unable to use mechjeb with a vehicle which doesn't include a strong reaction wheel due to the constant oscillation around the desired attitude when using RCS as the primary control mechanism.

Link to comment
Share on other sites

@sarbian Do you have any thoughts abouthe what could be causing my control problems and/or how I might be able to get around them (suggested settings for attitude adjustment page)? I don't like to bug, but I'm currently unable to use mechjeb with a vehicle which doesn't include a strong reaction wheel due to the constant oscillation around the desired attitude when using RCS as the primary control mechanism.

Have you tried using RCS balancer to balance out the RCS? A poorly balanced RCS will cause problems even for MJ.

I noticed an issue with MJ Delta-V info.

-snip-

Hm, I decided to check this out out of curiosity and checked with a simple setup of Probotodyne HECS, orange jumbo, and a mainsail. While I'm not seeing the discrepancies like you're seeing, I am seeing that both of them show a small increase in the atmospheric deltaV than it was shown in the VAB, not sure what's up with that.

Also, for Duna, with the same setup, KER shows a higher atmospheric deltaV than MJ does, not sure which one is supposed to be correct.

I'll look at some of my more complex craft.

Edited by smjjames
Link to comment
Share on other sites

Have you tried using RCS balancer to balance out the RCS? A poorly balanced RCS will cause problems even for MJ.

I build all of my craft with RCS build aid to make sure they are balanced out of the box. That being said, yes, I did try using the balancer, but it did not fix the problem.

EDIT: Something I want to add after some more playing around is that the primary problem seems to lie in roll control, though pitch and yaw do seem to exist as well. It seems that after a vehicle arrives at the desired attitude, it's the roll thrusters that are left constantly blasting while the pitch and yaw only seem to fire occasionally.

CP1KfHY.png

Edited by SpacedInvader
Link to comment
Share on other sites

OK I found what disturbed KER

I had 2 dropable tanks with fuel conduit going to a third tank. The two tanks where not part of the same stage. When I chained them, calculations where better for KER.

Then I had another issue. My "second stage" is able to work with remnant with the "first stage" (so I jettison the first stage after the "second stage" is activated. KER seems to assume that I drop the engines which are still attached (??)

But KER is not the subject here. MechJeb calculate everything correctly on the Pad, but not in the VAB.

Here is the corrected file (but still goes wrong in the VAB for MJ

http://n.lejeune.free.fr/ksp/moho2.zip

Edited by Warzouz
Link to comment
Share on other sites

What's up with the different atmospheric calculations for Duna though while in editor? It may be because I'm using FAR, not sure.

I build all of my craft with RCS build aid to make sure they are balanced out of the box. That being said, yes, I did try using the balancer, but it did not fix the problem.

EDIT: Something I want to add after some more playing around is that the primary problem seems to lie in roll control, though pitch and yaw do seem to exist as well. It seems that after a vehicle arrives at the desired attitude, it's the roll thrusters that are left constantly blasting while the pitch and yaw only seem to fire occasionally.

http://i.imgur.com/CP1KfHY.png

Do you have an SAS on that ship? I can't tell, though the command pod should have some anyhow.

What about toggling RCSbalancer to use or don't use RCS for rotation?

Edited by smjjames
Link to comment
Share on other sites

What's up with the different atmospheric calculations for Duna though while in editor? It may be because I'm using FAR, not sure.

Do you have an SAS on that ship? I can't tell, though the command pod should have some anyhow.

What about toggling RCSbalancer to use or don't use RCS for rotation?

It does have SAS, but the torque has been disabled for testing purposes. That said, I'm using SETI so most of the reaction wheels have been either disabled or significantly scaled down, leading to the need for good control through RCS.

EDIT: Turning off the ability to use RCS for rotation disables all RCS control except for translation...

Edited by SpacedInvader
Link to comment
Share on other sites

It does have SAS, but the torque has been disabled for testing purposes. That said, I'm using SETI so most of the reaction wheels have been either disabled or significantly scaled down, leading to the need for good control through RCS.

EDIT: Turning off the ability to use RCS for rotation disables all RCS control except for translation...

I did a test with my orbital tug and I'm not seeing any problems, but then again, I don't have realfuels, so maybe look into that?

Link to comment
Share on other sites

What about toggling RCSbalancer to use or don't use RCS for rotation?

Please no. As I said more than once the balancer is not perfect. If you have an already balanced craft then DO NOT use it.

The problem here lies elsewhere but I can't work on it for now.

Link to comment
Share on other sites

So now I'm faced with a new and perplexing problem. Whenever I try to have MechJeb execute a maneuver node, the autopilot will work just fine right up until the last 0.2m/s at which time, instead of the engine shutting off, the maneuver node slides up to sit perpendicular to the horizon, with MJ following it up there, and then will not complete ever, with MJ constantly thrusting at minimum power until for at least several minutes, or even until all fuel is burned and without ever shutting off automatically. This is a new effect which I have never seen before, and for which I can't think of any cause except my attempts to tweak the attitude adjustment.

Unfortunately, though, I can't get it to go away at all now. I've deleted and reinstalled MechJeb several times now, deleted and rebuilt several ships to ensure there weren't any old MJ modules sitting somewhere in the craft files. None of this has corrected this new problem at all. I'm unsure if it even relates to MJ directly or if I'm getting the issue from somewhere else, but the only variable I've changed recently was Mechjeb settings, so if there is any other possibility, I'm at a loss to know what it could be. Any help with this new issue would be greatly appreciated as I rely too heavily on MJ to simply go without at this stage.

Link to comment
Share on other sites

Thanks. I even have DMP working. Why doesn't it work when in Program Files?

I did not look into it but I guess windows is wary of unsigned DLL that were added after the original install.

SpacedInvader : for large ship you need to lower the node tolerance, like under 0.2 or 0.3.

Edited by sarbian
Link to comment
Share on other sites

I did not look into it but I guess windows is wary of unsigned DLL that were added after the original install.

SpacedInvader : for large ship you need to lower the node tolerance, like under 0.2 or 0.3.

This is the case even for a 4 part ship consisting only of an alcor pod, a reaction wheel, fuel tank, and engine. I should also note that ships that performed flawlessly for me in the past are now exhibiting this behavior so it is something more general than node tolerance.

EDIT: It also appears to be something limited to a specific install of the game and not really the fault of MJ in particular as a test install I put together to try and narrow down the cause of my RCS issue does not display the same behavior. I guess I will have to figure out where this conflict comes from before continuing.

EDIT2: I can also confirm that the original RCS issue is directly related to MechJeb as using RCS only to make attitude changes through MJ results in the same overuse of monopropellant as I was experiencing in a fully modded build, though at a much lower rate. I think that the core issue lies with MJ, but then it is drastically exacerbated by the changes that Real Fuels brings to RCS thrusters / fuel.

Edited by SpacedInvader
Link to comment
Share on other sites

I don't know how you did your re-install but have a look at the cfg in gameData/Mechjeb/plugin/plugindata/Mechjeb (or something like that). Those are your MJ settings, so move/delete them to reset everything.

Also as I just said in the RCSFX thread your screenshot show no available roll torque which is quite suspicious since your RCS should clearly provide some. MJ fails to detect it. But I won't have time to look into it until at least later next week...

Link to comment
Share on other sites

I don't know how you did your re-install but have a look at the cfg in gameData/Mechjeb/plugin/plugindata/Mechjeb (or something like that). Those are your MJ settings, so move/delete them to reset everything.

Also as I just said in the RCSFX thread your screenshot show no available roll torque which is quite suspicious since your RCS should clearly provide some. MJ fails to detect it. But I won't have time to look into it until at least later next week...

The reinstall was a completely fresh install of the game from the 0.90 zip provided by Squad. Also the MechJeb, Real Fuels, and ModuleRCSFX installs were all fresh downloads.

Link to comment
Share on other sites

Dear sarbian and other MechJeb Contributors,

When KSP 1.0 launches next Tuesday, please make MJ2 compatibility with KSP 1.0 your life's priority. Set aside your family, your friends, your girl friend, your job, even your health to ensure this happens quickly. If you must stay up all night awaiting the Steam download and making the fixes once you get it, please do so. I'll supply the coffee, Jolt cola, and pizza.

Your fans will appreciate it.

thank you.

Link to comment
Share on other sites

Dear Apollo13,

Sorry but Avengers Age of Ultron is out tomorrow and I'll spend the next week going to the movies every night.

Lol Apollo.

Shouldn't take more than a day or two. Adapting to the new stock aero just means plugging in the numbers, right?

*gets the sarcasm detector*

ANALYSIS INCONCLUSIVE

*go the movies*

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
×
×
  • Create New...