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

[Makeone]

Has anyone experienced a almost total lockdown of the game when MJ tries to calculate(?) REALLY long orbit periods and such information? I have noticed it in the past and also currently, although i am not using the most current version of the MJ, actually just the 'hotfix' that came after 23.5 got released as it seems to show all the deltaV information, even with new engines and NFPP's electric engines. However, sometimes, even as simple as going to Mun, breaks the thing, and badly. Only way out is sign out and in from windows so that forces ksp.exe to close.

For example, i just tried to catch an asteroid that was about 40Mm Pe from kerbin, so no biggy, right? Asteroids orbit was, in kerbins SoI, Pe was 40Mm and Ap was ~1Gm(negative), when i tried to burn and close my speed differential, suddenly MJ showed my orbit around kerbin to be Pe 40Mm and Ap..several Gm, orbital period was 10 000 YEARS, so no wonder MJ stalled? Usually when orbital speed goes over the escape velocity, even MJ starts to show negative Ap distances/height.

The lockdown also appears if i try to burn more directly to target (like using Interstellars engines) and technically escape of kerbol, but again, orbital period goes to...well, VERY long.

[daawgees]

Has anyone experienced a almost total lockdown of the game when MJ tries to calculate(?) REALLY long orbit periods and such information? I have noticed it in the past and also currently, although i am not using the most current version of the MJ, actually just the 'hotfix' that came after 23.5 got released as it seems to show all the deltaV information, even with new engines and NFPP's electric engines. However, sometimes, even as simple as going to Mun, breaks the thing, and badly. Only way out is sign out and in from windows so that forces ksp.exe to close.

For example, i just tried to catch an asteroid that was about 40Mm Pe from kerbin, so no biggy, right? Asteroids orbit was, in kerbins SoI, Pe was 40Mm and Ap was ~1Gm(negative), when i tried to burn and close my speed differential, suddenly MJ showed my orbit around kerbin to be Pe 40Mm and Ap..several Gm, orbital period was 10 000 YEARS, so no wonder MJ stalled? Usually when orbital speed goes over the escape velocity, even MJ starts to show negative Ap distances/height.

The lockdown also appears if i try to burn more directly to target (like using Interstellars engines) and technically escape of kerbol, but again, orbital period goes to...well, VERY long.

I personally have never experienced this, either in the past with MJ or currently. This issue you're describing would lead me to believe that it could POSSIBLY be hardware related, but without more detailed information it would be hard to say.

[OrbitalDebris]

I think I may have a new bug to report, I'd like someone else to confirm first though. The issue I'm seeing is that with build 213, RasterPropMon can no longer communicate with mechjeb - the usual screen that provides access to the MJ functions shows the "autopilot not installed" message that normally shows up when MJ is not installed. MJ is installed, otherwise I have a ghost living in my PC that's rather good at putting rockets in orbit

I haven't touched RPM, so I'm guessing that something has changed in MJ that makes it incompatible somehow. Looking at ksp.log, I see a few error messages; Log is at https://onedrive.live.com/redir?resid=962FD212EEE06DAC!11402&authkey=!AJGPDhleD_Va0fg&ithint=file%2c.log

edit: I just tried with the dev build of RPM and still see the issue. I don't want to cross-post, so if Sarbian or r4m0n can confirm that this is not an MJ issue I'll go over to the RPM thread and cause a ruckus there

Edited April 9, 2014 by OrbitalDebris

[shadowlessman]

It has been fixed

Thanks a lot!

[Starwaster]

I think I may have a new bug to report, I'd like someone else to confirm first though. The issue I'm seeing is that with build 213, RasterPropMon can no longer communicate with mechjeb - the usual screen that provides access to the MJ functions shows the "autopilot not installed" message that normally shows up when MJ is not installed. MJ is installed, otherwise I have a ghost living in my PC that's rather good at putting rockets in orbit

I haven't touched RPM, so I'm guessing that something has changed in MJ that makes it incompatible somehow. Looking at ksp.log, I see a few error messages; Log is at https://onedrive.live.com/redir?resid=962FD212EEE06DAC!11402&authkey=!AJGPDhleD_Va0fg&ithint=file%2c.log

edit: I just tried with the dev build of RPM and still see the issue. I don't want to cross-post, so if Sarbian or r4m0n can confirm that this is not an MJ issue I'll go over to the RPM thread and cause a ruckus there

Did you also install KSO or something else that installs RPM? Something that overwrote your RPM files? I have problems using the MJ functions in the RPM monitors using the version that shipped with KSO. In any event, checking the RPM thread is the way to go, pretty sure they have a fix there.

[RazgrizRaptor]

So I've been having a problem with the auto docking program, for some reason it always forces the craft out to 321m then it does nothing cause I think it loses the target, it's very stressful, upside tho is it forced me to learn how to do docking by myself. does anyone know how to fix this? (everything else seems to work perfectly fine).

[OrbitalDebris]

Did you also install KSO or something else that installs RPM? Something that overwrote your RPM files? I have problems using the MJ functions in the RPM monitors using the version that shipped with KSO. In any event, checking the RPM thread is the way to go, pretty sure they have a fix there.

Pretty sure nothing's overwriting RPM files because I deleted the old folders in order to test the dev build to see if it fixed the problem. I was just over at the RPM thread and it sounds like both MJ and scansat are misbehaving, so this might actually be another 0.23.5 incompatibility problem. Ah well, I'm sure it will get sussed out and in the meantime RPM is just a nice to have.

[Subcidal]

I know I'm a bit outdated now but, In .23 with RSS and the realism overhaul compilation + a few mods here and there mechjeb's auto ascent pilot is screwy. It starts off with a rather sharp turn right after ignition then begins its gravity turn and i'm horizontal around 6-8 km. Giving it large winglets for extra control just causes it to wobble all over the place trying to go off and do whatever in the world it's trying to do. It dosen't really seem to make much of an attempt at following the ascent target, though it usually goes towards it's general direction, Like east or west. That's about as close as it gets to going where it's programmed to go. Sometimes it will go straight south instead of east, Seems random.

Long story short, For some reason my ascent auto pilot has a mind of it's own, and I can't figure out why. It kind of seems like it's doing a mini ascent profile up to around 7KM in whatever direction it happens to end up heading. The ascent turn start/end settings seem to make no difference at all. Only saw one other person mention a similar problem and they were never answered. Hoping somebody knows a solution, Figuring somebody has ran into this problem before and I just can't find it.

Edited April 9, 2014 by Subcidal

[Jethro Bodine]

I know I'm a bit outdated now but, In .23 with RSS and the realism overhaul compilation + a few mods here and there mechjeb's auto ascent pilot is screwy. It starts off with a rather sharp turn right after ignition then begins its gravity turn and i'm horizontal around 6-8 km. Giving it large winglets for extra control just causes it to wobble all over the place trying to go off and do whatever in the world it's trying to do. It dosen't really seem to make much of an attempt at following the ascent target, though it usually goes towards it's general direction, Like east or west. That's about as close as it gets to going where it's programmed to go. Sometimes it will go straight south instead of east, Seems random.

Long story short, For some reason my ascent auto pilot has a mind of it's own, and I can't figure out why. It kind of seems like it's doing a mini ascent profile up to around 7KM in whatever direction it happens to end up heading. The ascent turn start/end settings seem to make no difference at all. Only saw one other person mention a similar problem and they were never answered. Hoping somebody knows a solution, Figuring somebody has ran into this problem before and I just can't find it.

I don't know if you're experiencing the same issue, but I had some similar problems. It turned out my craft was simply top-heavy, and I didn't have enough reaction wheels/control surfaces to counter it. Try throwing on a couple reaction wheels and see if that helps?

[Sabor]

Any plans on making the Runway a targetable landing point, and getting together with the guys behind the Steam Gauges to make the Hud highlight it as NASA's Orbiter's Hud does?

Like so:

[Subcidal]

I don't know if you're experiencing the same issue, but I had some similar problems. It turned out my craft was simply top-heavy, and I didn't have enough reaction wheels/control surfaces to counter it. Try throwing on a couple reaction wheels and see if that helps?

Well I made a larger rocket to see how it would react and it followed it's ascent path quite well. It twitched every now and then on the way up for some reason but it still flew alright.

But I still feel as if something is off in my game ATM with mechjeb. The initial rocket I had problems with launched fine vertically with SAS on, and was able to perform a decent gravity turn if i did it manually, Could probably do it a lot better but I don't fly ascents. With MJ on it's all over the place though, It feels like mechjeb is making a half-hearted attempt to steer it. Though like I said that larger rocket I just launched seemed to ascend quite well along it's path.

I've used MJ and FAR and whatnot for quite a while, Quite a few hours (hundereds) logged with both of them, New to RSS though. Something just feels off, I can't really put my finger on it. It feels like the ascent autopilot is being lazy or something. My original thought was that mechjeb didn't scale properly with RSS or something as it kind of seemed like it thought vertical was a 80-70 degree pitch to the east. It also rolls a lot more than I ever remember it doing before.

Were you using RSS when you had this issue? It's the only mod I can think of that would really affect mechjeb. Though nobody else seems to be having problems with it. Maybe i've just become extremely rusty since .21 haha.

[sarbian]

I just pushed #214 that fix the Toolbar. You'll also see a new checkbox in settings to hide the big Mechjeb button on the side when the toolbar is active.

Can I get a bit of feedback on that ? If it sounds ok I'll ask r4m0n to push a new official with 213 & 214 fix so all user can enjoy a better experience. ( I ll have a quick look at the docking AP and may add a fix to that too ).

[MK3424]

No issues found here!

[BudgetHedgehog]

FWIW, I tried the docking AP yesterday on 213 and found no problems, despite the craft being horribly imbalanced. Hooray for 214!

[AndreyATGB]

#215 works fine as far as I can tell. Seems to be the same as old MJ with the toolbar addon worked.

[Dimetime35c]

The toolbar update fixed my issues =) I'd like to suggest maybe adding a feature that some mods have that like will alert you when a new update has been published.

[sarbian]

With the help of Padishar I tracked why the dV stats and mass where different while in the editor and in flight. Now they should be consistent in #216

[Ashiataka]

I have tried to install MechJeb by copying the contents of the .zip folder into the GameData folder, but when I start a new game the mechjeb button doesn't appear on the right hand side. Does anyone know what might cause this?

Thank you.

[sojourner]

^Did you attach the mechjeb part to your rocket? (ar202 under control parts)

[Ashiataka]

Okay, yes found it and it works. I think this should probably be mentioned in the first post or on the installation instructions here http://kerbalspaceprogram.com/mechjeb/.

Thank you.

[Mystique]

"force roll" issue seems to be fixed, cool

[Brainstorm]

I have tried to install MechJeb by copying the contents of the .zip folder into the GameData folder, but when I start a new game the mechjeb button doesn't appear on the right hand side. Does anyone know what might cause this?

Thank you.

I've got the same problem. the "Ar202" is in my parts and also attached to my rockets. Before i updated to from 0.23 to 0.23.5 it worked and the menu showed up on the right side at the launch pad. Put after the update to 0.23.5 the menu is gone. I first thought it's because of the toolbar but after deinstalling the toolbar and also de- and reinstalling the whole KSP the menu still didn't show up.

[NathanKell]

On MechJeb and FAR/RSS/RO/whatever

First, topheavy is GOOD. Topheavy means you're stable (CoM in front of CoL). It's like a dart or paper airplane--the heavier the nose, the more stably it flies.

Now, here's how to fly an ascent with FAR/RSS (Other than turn end, this applies to Kerbin as well). With FAR, you must learn what a gravity turn actually is. Protip: it's not "fly up to 10km, pitch over 45 degrees." What a gravity turn actually is, is letting gravity pitch you over (hence its name). The way this works is that, once you reach a decent airspeed (100m/s or so, say--less if your TWR is KSP-normal rather than real-life-normal) you pitch over about five degrees and then don't touch pitch again. As gravity pulls your velocity vector downwards towards the horizon, aerodynamic forces will keep your rocket aligned with the velocity vector, and your rocket will pitch to follow. Your aim is to be horizontal at apogee.

Note that as it stands this can't actually be done with MechJeb, since MechJeb manually controls pitch all the way and keys it to altitude. However, you can set MJ to fly a decent approximation. Here are the steps.

*First, set your "turn start" to "the altitude at which you reach ~100m/s." This will be something like 1km if you have a 1.5TWR at liftoff, 1.5km if 1.2TWR, or something like 0.5km if 1.7. If you have a 2+ TWR at liftoff (from solids, say) you probably want to start your turn at 0.01km.

*Second, you want to set your "turn end" altitude and your "turn shape" such that deviation from your velocity vectory (the surface prograde indicator) is minimized. For your average two-stage rocket with a decent second-stage TWR (i.e. starts around 1, ends at maybe 3 or 4--that is, like Titan II rather than modern LVs) you might want to set turn at at 150km and turn shape at 50.

*Third, you fly a test ascent with "show ascent path" turned on. Watch how close the purple target (and your rocket's nose) is to the velocity vector. If it gets out of line by more than 5 degrees when below 80km, adjust turn end and turn shape.

*Fourth, if you find yourself circularizing after apogee (very common with low-thrust second stages), once you reach apogee you need to play with "end angle" to keep your vertical speed zeroed. Don't touch it until after apogee, mind. Then raise it until your stage is pointing vertically enough to offset gravity; as you speed up you gradually lower that angle until it reaches 0 (no up-thrust needed) at circularization.

[Jethro Bodine]

With the help of Padishar I tracked why the dV stats and mass where different while in the editor and in flight. Now they should be consistent in #216

Just installed #216 and have only done a quick cursory test so far, but everything seems fantastic. Great job on the Toolbar integration! Thank you, thank you, thank you!

BTW, did you ever get any word on Autom8? Just curious.

[EdisonMaxwell]

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.