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

[Capt. Hunt]

Ok, I reverted from the dev build to 2.5.1 on a hunch, so far I've only done one launch but it seems to have gone ok. I don't recall if I had the AOA limiter set, but I'll try turning it off. I am noticing some reentry heating on launch despite having the terminal velocity limiter set though, so I'm not sure that's kicking in.

[smjjames]

It's known that terminal velocity doesn't work very well, do limit acceleration to 18m/s until around 20-25km instead. Tweak as you prefer.

[Capt. Hunt]

ok, I'll try that. This might explain why my rocket also appears to be burning more fuel then usual to get to orbit. This same design used to make orbit with half of the fuel left in it's second stage, and now it's burning the SM engine to get into orbit.

[MrDude]

Is there a solution to when MJ (latest Version) glitches during maneuver node executing?

It Aims at the marker just fine, but doesn't begin the auto warping, nor does it begin the burn. (yes Auto warp is on etc etc)

It happens with about 1/4 of my missions and not always from the very start, but once it happens it seems to persist with that vessel. (I read some where that manual maneuvering (spin and burn a few delta V's) snaps it back into shape... and I have had it do that occasionally but not over the last few days.

I've tried re building the nodes but lately once a ship glitches it stays glitched... grrr. Just starting my bigger missions.

[Higgs]

it would seem that what ever bug or glitch or error that was haunting me and my family during KSP 0.90 and MJ 2.4.2 has blossomed into something even uglier. My issues were unending warping, pre launch time clamp firing and random paths to orbit not called for. It seems now that some people are not even getting auto warp to function correctly as well as a few others. My fear has been confirmed, what ever the bug was that no one could find when i brought it up, has now gone thru a full transformation and has reared its ugly head. <runs off crying>

[mikegarrison]

Is there a solution to when MJ (latest Version) glitches during maneuver node executing?

It Aims at the marker just fine, but doesn't begin the auto warping, nor does it begin the burn. (yes Auto warp is on etc etc)

It happens with about 1/4 of my missions and not always from the very start, but once it happens it seems to persist with that vessel. (I read some where that manual maneuvering (spin and burn a few delta V's) snaps it back into shape... and I have had it do that occasionally but not over the last few days.

I've tried re building the nodes but lately once a ship glitches it stays glitched... grrr. Just starting my bigger missions.

That's exactly the same thing I was talking about. I flew that ship all the way from Minmus to Eve orbit to Gilly orbit, dropped Val to jetpack down and then back up to the orbiting ship, and then flew home. MechJeb never worked right the entire time. But only when I was going to separate the capsule for re-entry did I realize that the engines were not staged (due to having docked and then undocked). I wonder now if that is what causes this behavior.

[smjjames]

@those having problems with maneuver nodes: What other mods are you using? Maybe there's a common thread somewhere as I've never ran into that problem myself.

also, post output logs because sarbian can't help if you don't.

[DaniDE]

Hi there Sarbian,

I have small UI request. Could you change the Time Warp window to use buttons instead of a drop down menu? I use this feature extensively (as almost everyone else probably) and I find myself clicking a lot in there to switch between i.e. SoI and atmospheric entry alot. It is a bit tedious after a while Since there are not that many options in there anyway, I would like to propose to change it to a button like layout like this:

Excuse the crappy ASCII version, but I have no nifty graphical editor on this machine right now.

Edited June 2, 2015 by DaniDE

[Two]

From the changelog:

#464 (20.05.2015 23:23:47)
Work around for #591 (node burn problem) since I pulled a unfinished KER branch

So it seems like this is a bug in the current version and already fixed in a not yet released update.

[sarbian]

You mean the "not yet released update" that are linked in the first post ?

I'll stop replying here. I posted instructions on how to report bugs. If you don't want to follow them fine, just don't expect the bug to be looked at.

Bugs reports will now be handled exclusively thru github. Duplicates or bug without info will be closed without any remorse on my end.

[smjjames]

You mean the "not yet released update" that are linked in the first post ?

I'll stop replying here. I posted instructions on how to report bugs. If you don't want to follow them fine, just don't expect the bug to be looked at.

Bugs reports will now be handled exclusively thru github. Duplicates or bug without info will be closed without any remorse on my end.

Might want to link the instructions, not for me (you seem pretty satisfied with the way I do bug reports though), but for those who don't know where the instructions are.

And okay on github.

Edited June 2, 2015 by smjjames

[Climberfx]

The basic sense of Mj Smart SAS work like Stock ones on timewarp are still wrong on Mj (to hold rotation). It won't keep the desired angle when time warp, it free the rotation, so that way any Mj function that put a node direction and time warp won't work as it should. Ascendence guidance, maneuver planer, landing, etc...

This will be fixed? or i will need to keep changing from stock to mj sas manualy overtime?

Edited June 2, 2015 by Climberfx

[Foxster]

You mean the "not yet released update" that are linked in the first post ?

I'll stop replying here. I posted instructions on how to report bugs. If you don't want to follow them fine, just don't expect the bug to be looked at.

Bugs reports will now be handled exclusively thru github. Duplicates or bug without info will be closed without any remorse on my end.

Good idea. There have been some rude and demanding people posting here recently with really vaguely defined problems and/or demands.

Don't lose heart though - most of us out here are just incredibly grateful for all you hard work!

[mikegarrison]

@those having problems with maneuver nodes: What other mods are you using? Maybe there's a common thread somewhere as I've never ran into that problem myself.

also, post output logs because sarbian can't help if you don't.

1) MechJeb, SpaceY, NearFuture, KIS/KAS, Planetshine

I had the problem happen to me most recently when I detached a ship after attaching it to a base using KAS.

[smjjames]

How do I get rendezvous to target to work from a location other than Kerbin? I'm not complaining or making a bug report, just asking for some help.

To set the situation, my target is in an orbit of approx 110.6-109.9km (it's fluctuating a little due to a number of reasons, not massively though) and in an inclination of 15.623 while the ship I'm launching is at a latitude of 9 degrees (how do you do a degree symbol on a keyboard?) 2' South. Editwhile typing: The rendezvous planner now says 110x111km, but I guess it's rounding a bit.

I know there are posts saying about how to rendezvous to target, but they're all assuming that you're at KSC, so I have to do a quicksave and revert to that.

What I've tried are:

1. Launching into the targets orbital height, doing circularization, reverting (F9, not revert to launch), clicking launch into plane of target and trying to do that.

2. Clicking launch into plane of target to set the inclination, fly that to 110km and circularize, revert, do rendezvous to target.

3. Launching to the target height, circularizing, reverting, rendezvous to target using that orbital inclination (it went to -15 rather than 15 though).

4. Same as 3, just an inclination of 0, which is the latest one.

My latest attempt:

This is actually the closest approach that it has gotten so far.

[sarbian]

Dev build #467

- Redo the terminal verlocity. This one should be a bit more stable

- Fix variable mass modules. Total mass was not proper with fairrings

- Bring back an updated version of the Mechjeb native fuel sim. I updated the old MJ fuel sim for 1.0, added some goodies and optimization.

Hopefully it will fix all the node execution problem. The only test ship that did not work was the SSTO someone sent me that still reports 0 dv while the engine still works. I'll have an other look later.

[Guest]

Sarbian, thanks for all your work and dedication. I just launched a Tantares Soyuz craft to a Salyut 1 station. Rendezvous was spot on, within 10km of target. Docking went amazingly well, with only a marginal monopropellant overuse due to my being a bit sloppy with the RCS pod placement. Your work makes my KSP experience pleasurable since I'm a much better engineer than I am a pilot. Please know that you are appreciated.

[mikegarrison]

How do I get rendezvous to target to work from a location other than Kerbin?

I think your basic problem is that what you are trying to do is almost impossible.

First, you can never directly launch to an orbit that has a lower inclination than your starting point. So if you were trying to hit an orbit of 9 degrees from a launch site of 15, you would be SOL. (You can use a dog-leg launch, where you launch toward the plane you want, get your apoapsis onto it, and then circularize at your apoapsis but in the target plane rather than your orbital plane. Costs fuel, but not as much fuel as if you circularized before changing planes.)

Given that you are trying to do the opposite (go from 9 to 15), it is possible to put your ship into that orbital plane. You have to launch when your ground site is in that plane. So you are constrained in timing. But in order to launch directly to a rendezvous, you can only launch when the target is at the right spot. That ALSO constrains your timing.

The chances that the target will be in the right spot at the same time as your launch site is in the orbital plane are very, very low. I think that's why MechJeb can't do it for you automatically.

Best thing usually is to just launch to the target's orbital plane (if you can), establish an orbit either higher or lower than your target, and then do an in-plane rendezvous at the appropriate later time.

[futrtrubl]

How do I get rendezvous to target to work from a location other than Kerbin? I'm not complaining or making a bug report, just asking for some help.

To set the situation, my target is in an orbit of approx 110.6-109.9km (it's fluctuating a little due to a number of reasons, not massively though) and in an inclination of 15.623 while the ship I'm launching is at a latitude of 9 degrees (how do you do a degree symbol on a keyboard?) 2' South. Editwhile typing: The rendezvous planner now says 110x111km, but I guess it's rounding a bit.

I know there are posts saying about how to rendezvous to target, but they're all assuming that you're at KSC, so I have to do a quicksave and revert to that.

What I've tried are:

1. Launching into the targets orbital height, doing circularization, reverting (F9, not revert to launch), clicking launch into plane of target and trying to do that.

2. Clicking launch into plane of target to set the inclination, fly that to 110km and circularize, revert, do rendezvous to target.

3. Launching to the target height, circularizing, reverting, rendezvous to target using that orbital inclination (it went to -15 rather than 15 though).

4. Same as 3, just an inclination of 0, which is the latest one.

My latest attempt:

http://img.photobucket.com/albums/v214/smjjames/screenshot1_zpsq8mzguwz.png

This is actually the closest approach that it has gotten so far.

http://img.photobucket.com/albums/v214/smjjames/screenshot2_zpsmzlj0e71.png

Launching to rendezvous with an inclined target will almost never work, can almost never work. You have to launch at a certain time in order to match the plane, but you also have to launch at a certain time to get the right launch phase angle. These times will almost never coincide.

Best option is to into plane of target at a significantly different altitude and then do "intercept with Hohmann transfer".

Edit: If you want to train the launch to rendezvous for the phase angle while on a different body you will need to use the "mark" command at the start of the launch at the same time you launch. Mechjeb only writes the phase angles for the first launch of a ship, this is so that launches from other bodies or when your ship has changed (staging, fuel use, docking, etc) doesn't mess up your initial launches later with the same type of ship.

Edited June 2, 2015 by futrtrubl

[smjjames]

Launching to rendezvous with an inclined target will almost never work, can almost never work. You have to launch at a certain time in order to match the plane, but you also have to launch at a certain time to get the right launch phase angle. These times will almost never coincide.

Best option is to into plane of target at a significantly different altitude and then do "intercept with Hohmann transfer".

Edit: If you want to train the launch to rendezvous for the phase angle while on a different body you will need to use the "mark" command at the start of the launch at the same time you launch. Mechjeb only writes the phase angles for the first launch of a ship, this is so that launches from other bodies or when your ship has changed (staging, fuel use, docking, etc) doesn't mess up your initial launches later with the same type of ship.

I understand then.

Out of curiosity and for future reference, what's the inclination range that rendezvous with target will definetly work? I realize that it would vary from planet to planet, or moon to moon.

15 degree inclination is really steep for a body the size of Ike anyways. Might work okay at Kerbin with a 15 degree inclination, haven't tried.

I think your basic problem is that what you are trying to do is almost impossible.

-snip explaination-

Okay. I wasn't even sure if what I was trying to do even was possible (well, statistically possible) in the first place, and wasn't real clear on the whole mechanism behind the launch to rendezvous.

I can just do 'launch into plane of orbit' at a lower altitude and then hohmann maneuver to the target. Used that method while constructing the space station in my save.

Edit: I ran into a case of MJ ascent AP not waiting for the player to hit spacebar when autostaging is on. I'll put that up on github.

Edited June 2, 2015 by smjjames

[mikegarrison]

Edit: I ran into a case of MJ ascent AP not waiting for the player to hit spacebar when autostaging is on. I'll put that up on github.

Oh yeah, I have had that happen to me a lot. It's not the ascent AP that causes the problem for me. What happens is that when I have my ship sitting on the pad, if I turn on autostaging in the Utilities window, the ship launches immediately (no spacebar).

[smjjames]

Oh yeah, I have had that happen to me a lot. It's not the ascent AP that causes the problem for me. What happens is that when I have my ship sitting on the pad, if I turn on autostaging in the Utilities window, the ship launches immediately (no spacebar).

Feel free to add to the github issue that I've posted. I've noticed as well that it's something with having autostaging on and the ascent AP is usually where I toggle it at launch.

It's actually a problem that used to happen a lot more (back in 0.90 and in like January), but with some dev build updates, it was mostly gone and by the time 1.0 came out, it was almost totally gone. I rarely see that problem now.

Edit: I tried your method and it works there too. I'll go and post your information.

[futrtrubl]

I understand then.

Out of curiosity and for future reference, what's the inclination range that rendezvous with target will definetly work? I realize that it would vary from planet to planet, or moon to moon.

15 degree inclination is really steep for a body the size of Ike anyways. Might work okay at Kerbin with a 15 degree inclination, haven't tried.

15 degrees is 15 degrees, doesn't matter what planet. Heavier planets will take you more dV to later match inclinations. Launch to rendezvous will put you in the same orbit just different inclination, so no matter the inclination you will need to match planes at the ascending/descending node, which is also the same thing as matching velocities at closest approach in this situation.

[AgentMOO]

467 fixed my node problem. Thank you!

[Galane]

Launching to rendezvous with an inclined target will almost never work, can almost never work. You have to launch at a certain time in order to match the plane, but you also have to launch at a certain time to get the right launch phase angle. These times will almost never coincide.

NASA and other Earthly rocket launchers do exactly that. They wait until the time when the target orbit and the launch site are at the proper coincidence in order to use the least amount of reaction mass. Launch windows.

To launch to ISS required a northerly "dogleg" to reach the orbit of the ISS which is inclined to the degree it is so that it will pass over Russia's launch site. Most other Shuttle launches angled to the south in a mostly straight path that was more efficient from Canaveral's 28.4556° N latitude.

That's why Columbia didn't go to ISS, it was too heavy to carry a useful payload to the station even with the super lightweight external fuel tanks. Which is what led to its demise due to flying with an old "lightweight" tank on its last mission instead of one of the newer tanks with revised and fresher insulation.