Jump to content

MechJeb 2 - Patch test bed release (October 10)


sarbian

Recommended Posts

Well, for one, you have to turn off MJ to do that. So what's the point of Ascent Guidance if you have to disable it to do what you want to add to it?

Actually, usually MechJeb (in this patch branch) does not interfere with you changing roll during ascent and makes no attempt to roll during ascent. Except when your rocket has aerodynamic surfaces, in which case it immediately attempts to roll to 0, usually before it cleared the tower.

It typically doesn't interfere with you changing roll afterwards, but by then the damage, if any, has been done. :)

Link to comment
Share on other sites

Exactly. In my experience, it will let you change the roll, but it's trying to correct the ascent in the meantime - and sometimes the result seems to be that it doesn't always remain exactly on-course. Maybe it's only off by a degree, but that adds up by the time you make orbit. Rather than an inclination of 0 to the target, it can require some (manual) correction once orbit is achieved.

I guess my point was much the same as yours - if I have to do it manually anyhow, why would I need MJ in the first place? So a means of controlling roll during automatic ascent would be a good thing.

Link to comment
Share on other sites

All fair enough. The ascent module is one of my most frequently used parts of MJ. In general I roll to 90 manually (with MJ still enabled) as soon as I get off the pad, then feel free to walk away until circularization.

EDIT: is there a way to remove default menu items from the MechJeb list? I never intend to use rover autopilot, spaceplane guidance, or autodock, and my UI is getting rather cramped as-is.

Link to comment
Share on other sites

@56egreSS With Visual Studio or Unity Monodevellop ? I use VS but it should be the same for both. Copy the source somewhere, add the project files and open the .sln

Here is my project files. You'll need to change the "C:\Games\ksp-win-0-21-1" path from them. hack away and share your improvement :)

Build: 1 succeeded, 0 failed

You are the man. Thanks.

Link to comment
Share on other sites

I've noticed a lot of instability with MJ in lower atmosphere since installing the 0.9.6.2 FAR that came out a few days ago. I know, not really MJs problem but I thought I'd mention it. Air friction under 15,000m or so makes the gravity turn into flipflop city. Anyone using FAR either edit the ascent to start way higher than MJ wants to or fly manually till 15 to 20km then flip MJ on.

Link to comment
Share on other sites

...correction.

MechJeb sometimes will and sometimes won't attempt to get a specific roll during ascent, and I can't tell what exactly determines that, I'd need a series of experiments for this -- I'm just sure it's not the aerodynamic surfaces. This actually caused a well-debugged rocket to crash three times for me today. :)

Link to comment
Share on other sites

...correction.

MechJeb sometimes will and sometimes won't attempt to get a specific roll during ascent, and I can't tell what exactly determines that, I'd need a series of experiments for this -- I'm just sure it's not the aerodynamic surfaces. This actually caused a well-debugged rocket to crash three times for me today. :)

I also got that, only way to turn this around (pun intended) is to restart KSP.exe...

Link to comment
Share on other sites

Anybody encountered occasions that when auto stage on, MJ will switch to next stage too early. I'm having this randomly that MJ dumped my 2nd stage when it still have over 90% fuel in it......

Link to comment
Share on other sites

Anybody encountered occasions that when auto stage on, MJ will switch to next stage too early. I'm having this randomly that MJ dumped my 2nd stage when it still have over 90% fuel in it......

Yes, there are situations in which autostaging will do that... but I forget what exactly if you provide a full list of your staging or a picture showing it we might be able to figure out where the problem is.

bottom line is that there's something your staging that's confusing it ... (that doesn't necessarily mean that you have bad staging though that could also be the case)

Link to comment
Share on other sites

Yes, there are situations in which autostaging will do that... but I forget what exactly if you provide a full list of your staging or a picture showing it we might be able to figure out where the problem is.

bottom line is that there's something your staging that's confusing it ... (that doesn't necessarily mean that you have bad staging though that could also be the case)

Here is a video of the problem, my rocket is not too complicated i think.... basically the 1st stage is the launch stabilizers, SRBs & big engines. 2nd stage is procedural fairing & engine + tanks. 3rd stage is stretchy tank + engine. may be those procedural fairing & tanks casing the problem???

edit: ignore the mess up prior srb sep & stage 1 sep, that could be the turn shape is to sharp, was trying 40% in that video.

Edited by mwlue
Link to comment
Share on other sites

Nevermind, rewatched video. Might help if you showed the delta-v stats during launch. It might show if mechjeb is reading how much d/v in that stage is wrong.

ok, i recorded another one.... but i have no idea what's those values means :P

and MJ seems lost control of the rocket randomly..... like this time i lauched the rocket ok, then revert flight to launch, then MJ will lost control & crash my rocket. Sometimes simply revert the flight again will get it working properly again but most time i need to restart ksp..... weird...

Link to comment
Share on other sites

Errr ok, i noticed the problem at stage 2, the procedural interstage base is casing the problem.

in my original version, i notice stage 2 has 0 atmo time, i think this is the cause. for some reason, simply re-attach the engine to the interstage don't fix it.

Shot5177.jpg

then i moved that interstage base up to stage1 seems fix it, i have atmo time in stage 2 now

Shot5178.jpg

and yes! autostage seems ok now. but MJ continue to f-up my rocket...... it really hate my rocket...:(

Link to comment
Share on other sites

Mwlue, aside from anything else, your TWR is way too low. You're not going to make orbit with 3000dV and a TWR of 1.4 or less. You need 1.8 minimum but try not to go over 2.0 until out of the lower atmosphere. Also turning below 10 to 12km (the thickest atmosphere) with large or wobbly rockets which don't have much reaction authority also goes badly. From what I can see you have one little SAS way up top. For an 800t rocket, that's diddly squat - MJ will not hold that against drag.

If you want MJ to do it, edit the ascent to start at 12km and see what happens

Link to comment
Share on other sites

Mwlue, aside from anything else, your TWR is way too low. You're not going to make orbit with 3000dV and a TWR of 1.4 or less. You need 1.8 minimum but try not to go over 2.0 until out of the lower atmosphere. Also turning below 10 to 12km (the thickest atmosphere) with large or wobbly rockets which don't have much reaction authority also goes badly. From what I can see you have one little SAS way up top. For an 800t rocket, that's diddly squat - MJ will not hold that against drag.

If you want MJ to do it, edit the ascent to start at 12km and see what happens

Look at all the pictures, not just the ones where he's been editing stuff. In that final on he's got it up to 5k delta-V. TWR should be ok too.

Unless he's using thrust corrector, in which case it would be lower in atmo than Mechjeb says. mwlue, out of curiousity, ARE you using Thrust Corrector?

Edit: Oops sorry, Watched video, you are turning a bit early. But the main problem is that your TWR is a lot higher at that altitude because you've burned a lot of fuel. Where it says 'limit acceleration to' change that to 20.

Edited by Starwaster
Link to comment
Share on other sites

Mwlue, aside from anything else, your TWR is way too low. You're not going to make orbit with 3000dV and a TWR of 1.4 or less. You need 1.8 minimum but try not to go over 2.0 until out of the lower atmosphere. Also turning below 10 to 12km (the thickest atmosphere) with large or wobbly rockets which don't have much reaction authority also goes badly. From what I can see you have one little SAS way up top. For an 800t rocket, that's diddly squat - MJ will not hold that against drag.

If you want MJ to do it, edit the ascent to start at 12km and see what happens

Look at all the pictures, not just the ones where he's been editing stuff. In that final on he's got it up to 5k delta-V. TWR should be ok too.

Unless he's using thrust corrector, in which case it would be lower in atmo than Mechjeb says. mwlue, out of curiousity, ARE you using Thrust Corrector?

Edit: Oops sorry, Watched video, you are turning a bit early. But the main problem is that your TWR is a lot higher at that altitude because you've burned a lot of fuel. Where it says 'limit acceleration to' change that to 20.

Thanks for the advice, i'm not using thrust corrector. Will try on those values and see how it goes.

Thank you.

Link to comment
Share on other sites

I agree that stock MJ's "land at a target" thing is a bit wonky when the planet has an atmosphere. Often, it does a de-orbit burn that looks spot on and the info box says it's within about 200m. But then for no apparent reason the ship switches ends and starts burning prograde until the orbit no longer touches the ground and keeps getting higher and higher, pointlessly wasting gobs of fuel. This doesn't happen on airless planets, just those with atmospheres.

However, this seems to happen only when you tell MJ to land at a spot that's too close ahead of where your ship currently is. MJ appears to need somewhere between 1/3 to 1/2 an orbit. So, the best way to avoid or at least minimize this behavior is to start the process with as much of an orbit to go as possible, like right after you've passed over the target so you get it the next time around. USUALLY this will prevent the spurious "wrong way" burning. However, what often happens instead is that the rocket will now burn at idle speed pretty much all the way down, saying it's doing a coarse correction of like 0.1m/s, but this number never changes and the landing error never decreases, so it's just a pointless waste of fuel again.

So all in all, I have the best luck with not only engaging the landing autopilot a full orbit in advance, but also then disengaging the autopilot immediately after the de-orbit/plane-change burn when the error is about 200m. Then I fly it myself the rest of the way down, including opening any parachutes (which MJ doesn't do very well, either). Once the chutes are open, I tell it to "land somewhere" and it goes down just fine from there.

Thanks for the reply, I'll give the full orbit burn a try. Really hoping to put a few modules together from separate craft on the surface and my manual attempts usually end up with them hundreds of km apart.

Link to comment
Share on other sites

Thanks for the advice, i'm not using thrust corrector. Will try on those values and see how it goes.

Thank you.

errr, what i did made myself look so dumb. The actual payload (CSM+LM) i'm going lift with this rocket is around 40 tons, but somehow i have put 70ton dummy weight into my test probe to simulate that.....

Shot5206s.jpg

so this rocket should have no problem putting my real payload to orbit \o/

Link to comment
Share on other sites

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