[1.12.x] Anatid Robotics / MuMech - MechJeb - Autopilot - [2.14.3] [4th March 2023]

[MikeO89]

This is one of the reasons  I'm staying with 1.31 for now. Mechjeb worked perfect in those. Last time picking landing coordinates in Mechjeb which just happened to be a base ship I made on the MUN. Well the ship land directly on top of it. Did a Falcon 9 landing on that sucker.

I've spend so much time playing 1.31 with all the mods I'm using. I like what I like, it,s working fine. Personally I'm still not convinced yet that moving to 1.4.2 is going to be a good thing. I'm still waiting and watching to see what's up. If I loose even one mod I have (or if it doesn't work right), that would be reason enough for me to stay where I am.

Edited March 30, 2018 by MikeO89

[sarbian]

1 hour ago, MikeO89 said:

This is one of the reasons  I'm staying with 1.31 for now. Mechjeb worked perfect in those. Last time picking landing coordinates in Mechjeb which just happened a base ship I made on the MUN. Well the ship land directly on top of it. DId a falcon landing on that sucker I'ved spend so much time playing 1.31.with all the mods I'm using.I like what I like, it,s working fine. Personslly i'm not convencents moving to 1.4.2 was going to be a good thing. Im still waiting and watching to see what's up.

Nothing changed in KSP that impacts the landing precision since 0.90. What does change is MechJeb code.

[Kerbital]

14 hours ago, steve_v said:

Eh? I just added an exception without problems, on FF 58.0.1. Is there more "security" stupidity in 59.x?

Yeah, maybe they indeed"improved" security in 59

[El Sancho]

I just visited the MJ2 Jenkins site and DL'ed v794 without problems

https://ksp.sarbian.com/jenkins/job/MechJeb2-Dev/

I had the same problems as everyone else yesterday, but it seems to be fixed now.

[viperwolf]

8 hours ago, El Sancho said:

I just visited the MJ2 Jenkins site and DL'ed v794 without problems

https://ksp.sarbian.com/jenkins/job/MechJeb2-Dev/

I had the same problems as everyone else yesterday, but it seems to be fixed now.

everything seems back to normal

[sarbian]

On 3/30/2018 at 12:19 AM, BARCLONE said:

Sarbian,

See if these links get those files correctly. First time I've used this service (DataFileHost).

KSP.log

player.log

Screenshot

Screenshot

No idea. The log is filled with error from your outdated SmokeScreen version but I did not find anything bad related to MJ.

[sarbian]

2.7.2

• Fix for the engine plate autostagging
• Fix for the scripting module SAS controls
• Fix the MOI code to handle "control from here" properly. Should help with Docking

Download 2.7.2 for KSP 1.4.1

[sarbian]

And it's a double release!

2.7.3

• Fix the fuels sim in the editor with surface attached decouplers...

Download 2.7.3 for KSP 1.4.1

[Gargamel]

One of my favorite mods, I appreciate all the work put into it, so a big thank you for all the years of fun I've had.

Couple things, and you might not want to bother with one of them, since it's for an older release. 

Running 2.6.0 in 1.3.1: 

While using the docking autopilot, it has been having a tendency to over shoot the docking axis for alignment before moving in to dock.    Normally when the DA goes sideways, it's due to poor ship design, but this has been frequently happening with identical ships (different flights).  It doesn't matter what the target is.  But the kicker is, it happens only sometimes.   One flight will dock perfectly under DA, the next 2 will need me to take over (and is easily done so, I just use DA to skip the drudgery), and the next will dock perfectly, seems to be random.   The ship flies past the docking axis, trying to slow down and overshoots, and then overshoots again trying to correct.    The farther off axis the ship is when DA is enabled, the more likely this is to happen.  I'm trying to get log files for you, but the last 3 dockings have gone smoothly on DA.  I'll post up when I can.

 

Second item is a request.  I'm not sure if you can even do this, given it's an interaction with another mod.   If MJ is controlling the time warp speed (ie warping to the next node), and Kerbal Alarm Clock fires off stopping the warp, MJ will fight KAC until it resumes the previous warp value.  If possible, I'd like MJ to respect the 'drop out of warp' when KAC triggers an alarm.  I wouldn't ask for most mods, but KAC is one of the 'standard' mods people use. 

 

Thanks. 

[Loren Pechtel]

2 minutes ago, Gargamel said:

While using the docking autopilot, it has been having a tendency to over shoot the docking axis for alignment before moving in to dock.    Normally when the DA goes sideways, it's due to poor ship design, but this has been frequently happening with identical ships (different flights).  It doesn't matter what the target is.  But the kicker is, it happens only sometimes.   One flight will dock perfectly under DA, the next 2 will need me to take over (and is easily done so, I just use DA to skip the drudgery), and the next will dock perfectly, seems to be random.   The ship flies past the docking axis, trying to slow down and overshoots, and then overshoots again trying to correct.    The farther off axis the ship is when DA is enabled, the more likely this is to happen.  I'm trying to get log files for you, but the last 3 dockings have gone smoothly on DA.  I'll post up when I can.

As far as I can tell this sort of problem is due to a failure to adequately compute when to start stopping motion.  While this particular instance sounds like it's MJ's fault I've seen it more than once in the core game.  Any time it takes too long to stop motion things go badly.

[sarbian]

It is most likely related to the bug I fixed in the last release. Build for older KSP are harder to do with the engine change...

[Starwaster]

post deleted... didn't realize imgur album linking was still so damned broken. Maybe next year we'll get working imgur albums again. Or not. Screw it.

 

Edited April 1, 2018 by Starwaster

[BARCLONE]

On 3/31/2018 at 4:33 AM, sarbian said:

No idea. The log is filled with error from your outdated SmokeScreen version but I did not find anything bad related to MJ.

DL'd your 2.8.0 this morning, will try it out later...

As for the LGAP issue, my working work-around is to set a (Waypoint Manager) waypoint first, then just use the Lat/Lon coordinates to set the landing target. That way I have something to aim for. Gets very close, about the same as picking the target through LGAP normally.

[Jacke]

@sarbian , you said the recent MechJeb releases are for KSP 1.4.1.   Is the change going to KSP 1.4.2 likely to affect MechJeb?

[sarbian]

10 minutes ago, Jacke said:

@sarbian , you said the recent MechJeb releases are for KSP 1.4.1.   Is the change going to KSP 1.4.2 likely to affect MechJeb?

No, the patch does not seems the break anything.

[El Sancho]

Just another example of how incredibly accurate and precise the Landing Guidance feature of MJ2 can be when everything goes right.

 

This was a Mun rescue mission. The Kerbal you see on EVA was to be rescued. I input the exact coords to Landing Guidance, and the rescue capsule with Elizer on board set it down this close, first try. It just doesn't come any more precise than that. Excellent work, Sarbian, excellent work indeed.

 

Win 10 pro 64, NVidia GTX 680

KSP 1.4.2 64-bit,  MJ 794

Edited April 1, 2018 by El Sancho
to make language more precise

[LameLefty]

38 minutes ago, El Sancho said:

Just another example of how incredibly accurate and precise the Landing Guidance feature of MJ2 can be when everything goes right.

 

This was a Mun rescue mission. The Kerbal you see on EVA was to be rescued. I input the exact coords to Landing Guidance, and the rescue capsule with Elizer on board set it down this close, first try. It just doesn't come any more precise than that. Excellent work, Sarbian, excellent work indeed.

 

Win 10 pro 64, NVidia GTX 680

KSP 1.4.2 64-bit,  MJ 794

Yep, in a vacuum it's damn near perfect. Atmosphere landings are still somewhat problematic, even from a dead-circular equatorial orbit aiming for an equatorial landing spot. One single-stage lander design on Laythe might undershoot the landing target by 1 - 3 km; another single-stage lander design from the same orbit and target might overshoot by just as much. Not sure if it's because drag calculations are screwed up in KSP for certain parts or what, but back around 0.26 or so, another player/coder put a crap-ton of effort into helping get the landing code straightened out and it used to be possible to land, even with parachute assist, within a 10 - 20 meters of an equatorial target easily from a reference starting orbit of about 100km equatorial. 

[MassiveD]

It looks like MechJeb for all link is down  can someone give a working one

[Jim DiGriz]

23 hours ago, LameLefty said:

Yep, in a vacuum it's damn near perfect. Atmosphere landings are still somewhat problematic, even from a dead-circular equatorial orbit aiming for an equatorial landing spot. One single-stage lander design on Laythe might undershoot the landing target by 1 - 3 km; another single-stage lander design from the same orbit and target might overshoot by just as much. Not sure if it's because drag calculations are screwed up in KSP for certain parts or what, but back around 0.26 or so, another player/coder put a crap-ton of effort into helping get the landing code straightened out and it used to be possible to land, even with parachute assist, within a 10 - 20 meters of an equatorial target easily from a reference starting orbit of about 100km equatorial. 

One of these days after I've solved the ascent and node executor issues adequately, I plan on giving the landing guidance module a complete overhaul with a decent trajectory optimization algorithm and closed-loop guidance which should be able to hoverslam a rocket pretty near perfectly in any atmosphere.  There's months and months between here and there though.

[LameLefty]

I think that would be grand. I’d be happy if I could just get a mostly-parachute single stage lander to drop within 100m or so (save that precious prop for returning to orbit!). It used to be possible but the game atmosphere has changed and, to top it off, apparently now there’s an issue with farings producing too much drag. That’s a big deal because, for instance, it’s generally necessary to encapsulate  large rovers or complicated habitat hub modules to avoid tearing/burning off fiddly bits. 

Anyway, we non-coders do appreciate Sarbian’s efforts, as well as those of everyone else who has or will contribute to the codebase. 

[HaydenTheKing]

how does aircraft autopilot work? I engaged autopilot and it didn't do any of my settings? All other mech jeb features work.

[IntoSpaceAgain]

The autopilot is fairly straightforward. Select which parameters you want the autopilot to manage using the checkboxes, set the parameters and click 'set', then click 'engage autopilot'. I usually set up altitude and heading before takeoff, but only engage the autopilot once I've taken off manually.

If the autopilot's still not working after doing that, then Sarbian will need to see your logs in order to identify any problems.

[Gargamel]

Has anybody ever really used that goofy/scary looking command pod?  Ya know, that one that looks like the evil older Cousin to Glados (Portal)?  I've been playing with MJ for half a decade now, and I still haven't built anything with it.   

[LameLefty]

I’ve been playing since 0.19 (so yeah, 5 years!) and never used it. 

[El Sancho]

6 minutes ago, LameLefty said:

I’ve been playing since 0.19 (so yeah, 5 years!) and never used it. 

Which command pod are we talking about? I played Portal and Portal2, and I've been playing KSP since 2012, but Gargamel's description left me confused...

Frankly, the new spherical command pods are so very nice that I seldom use anything else now. Just the fact that they have built-in separator and ablative shield makes them so much more weight-efficient, but you have to look closely at the numbers to see that.

Edited April 3, 2018 by El Sancho