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

[sebi.zzr]

42 minutes ago, Sarxis said:

1800? As in 18kPa? 

Yes.This is relevant only to celestial bodies with atmosphere (Q).
 

 

42 minutes ago, Sarxis said:

Getting into orbit from a moon isn't a problem.  Just wondering if there was a way to do a more efficient 'direct assent' out of SOI back to Kerbal. 

Like suggested before,get to orbital velocity as soon as possible,while not spending much energy gaining vertical speed.Then use "return from a moon" maneuver in "Maneuver Planner" and edit the node once plotted.I usually aim for 15 km orbit to clear the terrain.

If you want to use ascent autopilot,then you need to edit ascent path.

Edited February 20, 2017 by sebi.zzr

[Loren Pechtel]

32 minutes ago, Sarxis said:

1800? As in 18kPa? 

 

Getting into orbit from a moon isn't a problem.  Just wondering if there was a way to do a more efficient 'direct assent' out of SOI back to Kerbal.  Sometimes, depending where we land, 'up' will put us straight out of SOI into a return trajectory, but I didn't know if there was some way to use Mechjeb to orient the craft right from the surface into an efficient trajectory back home.  Unfortunately I can't seem to use 'Kerbal' as a target, otherwise I could use the Smart SAS to do some advanced orientation stuff. 

Unless you have an infinite TWR it's more efficient to go into orbit first.

The problem is gravity loss.  Escape velocity is a linear quantity, not a vector, you need the same final velocity no matter how you acquire it.  Burning horizontal is just as good as burning vertical.

However, your gravity loss is the gravity of the body * ( 1 - your fraction of orbital velocity)--integrate over your flight.

Lets say you are on the Mun and your engine can put out out 10 m/s.  By a simplistic answer you have to burn for 80.7 seconds to get away.  However, the Mun is pulling you back, you actually need to burn for 96.4 seconds to escape.  That is gravity loss.

Instead, lets try burning vertically for six seconds.  The rocket is heading up at 60 m/s - 10m/s due to gravity for a final velocity of 50 m/s, left to it's own devices it will stop rising after 30 more seconds and will go splat after 66 more seconds.

Instead of going splat lets burn horizontally for those 66 seconds and see what happens.

(Note:  I am doing a very crude brute force integration, the numbers are only approximate.  I am also assuming gravity doesn't drop off with distance.)

First 5 seconds, our average horizontal velocity is 25 m/s.  10km orbit is 547m/s relative to the surface.  Had I been standing still during this time I would have lost 8.67 m/s of vertical velocity but instead I only lost 8.0m/s.  I'm now heading up at 42m/s

Next 5 seconds, now the average horizontal velocity is 75 m/s.  That's 13% of orbit, I only lose 7.2m/s to gravity.  Heading up at 34.8 m/s

Five more seconds, 125 m/s, 23% of orbit, I lose 6.5m/s, going up at 28.3m/s

Five more seconds, 175 m/s, 32%, 5.7m/s, going up at 22.6m/s

Five more seconds, 225 m/s, 5.1m/s, up at 17.5m/s

275 m/s, 4.2 m/s, 13.3m/s

325 m/s, 3.4 m/s, 9.9m/s

375 m/s, 2.9 m/s, 7.0 m/s

425 m/s, 2.2 m/s, 4.8 m/s

475 m/s, 1.4 m/s, 3.4 m/s

525 m/s, .7 m/s, 2.3 m/s.

Oops, I didn't even need all of those 66 seconds.  I've spent 610 m/s of fuel to reach 550 m/s (and a hair more due to the 2.3m/s of vertical velocity but that's a smaller error than caused by my crude integration.)  Burning another 257 m/s will get me away for a total burn of 867 m/s compared to the 964 m/s of boosting straight up.

Of course rockets can't make a right angle change in thrust like that, it has to tip over gradually.  The basic logic applies, though.  You start tipping long before those 6 seconds are up.

This doesn't apply to bodies with an atmosphere, though.  Plowing through the thick air at high speed would squander fuel and very well might burn up your rocket.  Note, though, that MechJeb's default ascent path can get rather fiery--yet I've found it saves fuel.

[ragusila]

having a weird bug. It appears when I have mechjeb2 installed, if I delete it it goes away. In space, when I decouple something that doesnt have command ability (old stage, fairings, etc) that part/subassembly starts accelerating down (towards earth). If I switch to it, i cant switch back, as it says "part under acceleration". 

Is this a feature of mechjeb to clear the space of useless junk? Or is this a bug? 

How do I post logs? I have FAR, DR, RO/RSS Kerbalism and a bunch of part mods. The bug only dissapears if and only if I delete mechjeb. 

 

cheers

[Starwaster]

3 hours ago, ragusila said:

having a weird bug. It appears when I have mechjeb2 installed, if I delete it it goes away. In space, when I decouple something that doesnt have command ability (old stage, fairings, etc) that part/subassembly starts accelerating down (towards earth). If I switch to it, i cant switch back, as it says "part under acceleration". 

Is this a feature of mechjeb to clear the space of useless junk? Or is this a bug? 

How do I post logs? I have FAR, DR, RO/RSS Kerbalism and a bunch of part mods. The bug only dissapears if and only if I delete mechjeb. 

 

cheers

No, there's no such 'feature' in MJ2. 

You mean if you're in orbit and decouple something it starts doing that?

Then it's time for you to post logs as per the first post.

[Mandella]

10 hours ago, ragusila said:

having a weird bug. It appears when I have mechjeb2 installed, if I delete it it goes away. In space, when I decouple something that doesnt have command ability (old stage, fairings, etc) that part/subassembly starts accelerating down (towards earth). If I switch to it, i cant switch back, as it says "part under acceleration". 

Is this a feature of mechjeb to clear the space of useless junk? Or is this a bug? 

How do I post logs? I have FAR, DR, RO/RSS Kerbalism and a bunch of part mods. The bug only dissapears if and only if I delete mechjeb. 

 

cheers

I had this bug a while back, and it seemed to have something to do with an improper Kopernicus install. Or at least when I fully uninstalled (as in deleted the folder) and then reinstalled Kopernicus it cleared up. It's been a while, but if I remember correctly I might have somehow had a duplicate Kopernicus folder inside of Mechjeb.

Before I cleared the bug, I did come up with a workaround. Hit escape the instant you undock and quicksave. Go all the way back out to the main menu, then reload the save. Everything should be all right then, although the errant part will still have picked up a few meters per second difference in velocity.

Oh, and in my bug, the falling part could well have a command pod too, so it's not just for space junk..

[Loren Pechtel]

What is the status of advanced transfer to another planet?

I sent a bird to Moho.  11,000m/s of nuclear thrust at a 200km orbit about Kerbin.  (LKO would be a bad idea.  It dipped almost 40km during the ejection burn.)

The ejection burn was supposed to be ~2000 m/s.  After it burned down to about 620 m/s left the remaining burn went up rather than down!  After burning ~3000 m/s the orbits appeared to intersect (but I didn't get an intercept circle) but MechJeb still wanted to burn 700 m/s on the same vector that had been dropping the periapsis.  I finally aborted that burn and tried to fine tune the intercept.  Over 2000 m/s?!  Most of that was a big plane change--this was supposed to be a lowest energy intercept, there shouldn't have been one!  After that burn I had a decent intercept--but barely enough fuel for capture, not enough to reach the target orbit.  3x the fuel that would have gotten me into low Moho orbit yet it wasn't enough.

Alt-F12 shows nothing out of line, the log file was from before the 1.2 update--obviously not the right flight.

[brygun]

Greetings.

First off. Thank you. I am very much into the engineering aspects vs the twitch/flying. Espically when hyper details of 0.1 m/s make a difference for orbits.

My concern is that I don't see a way to edit "turn end altitude" under "Ascent path editor". I am using CKan and it is showing I am on MechJeb 2.6,0. This thread's post one at this time says that is the current version. I am using the mod RSS (real solar system) and the default of 119 km is still within the Earth's atmosphere. This gives a wee bit of an issue on efficiency. I am able to edit flight path angle and turn shape as text fields. The "turn end altitude" appears as just a number and non responsive to clicking for editing.  Various web searches just say to edit it but not describe an example. I've not yet collected logs as this seems to be a User Interface issue. My desired orbit is set to 200 km.

I have closely observed with RCS on to see that MechJeb isn't trying to steer up higher. I believe it is aiming for that "turn end altitude". If I take over* I make out all right. This shows the control devices are capable of steering.

* = Using SMART ASS with the ascent navball as a guide. Manually adjusting pitch and watching ascent profile to get a near match. The tail end a little high to get to the desired "turn end altitude" of outside the enlarged atmosphere.

1) Is there a place such as a config file I can manually change the "turn end altitude"?

2) Was/is the "turn end altitude" be something that can be edited? With the increase in mods with alternate solar systems this would be useful (to restore?)

3) Am I missing something else on how to adjust "turn end altitude"?

Edited February 25, 2017 by brygun

[Loren Pechtel]

29 minutes ago, brygun said:

Greetings.

First off. Thank you. I am very much into the engineering aspects vs the twitch/flying. Espically when hyper details of 0.1 m/s make a difference for orbits.

My concern is that I don't see a way to edit "turn end altitude" under "Ascent path editor". I am using CKan and it is showing I am on MechJeb 2.6,0. This thread's post one at this time says that is the current version. I am using the mod RSS (real solar system) and the default of 119 km is still within the Earth's atmosphere. This gives a wee bit of an issue on efficiency. I am able to edit flight path angle and turn shape as text fields. The "turn end altitude" appears as just a number and non responsive to clicking for editing.  Various web searches just say to edit it but not describe an example. I've not yet collected logs as this seems to be a User Interface issue. My desired orbit is set to 200 km.

I have closely observed with RCS on to see that MechJeb isn't trying to steer up higher. I believe it is aiming for that "turn end altitude". If I take over* I make out all right. This shows the control devices are capable of steering.

* = Using SMART ASS with the ascent navball as a guide. Manually adjusting pitch and watching ascent profile to get a near match. The tail end a little high to get to the desired "turn end altitude" of outside the enlarged atmosphere.

1) Is there a place such as a config file I can manually change the "turn end altitude"?

2) Was/is the "turn end altitude" be something that can be edited? With the increase in mods with alternate solar systems this would be useful (to restore?)

3) Am I missing something else on how to adjust "turn end altitude"?

The turn end altitude is normally within the atmosphere.  In stock it ends at 60km, so the fact that you are saying 119km means it took RSS into account.

Crazy as this seems it's the right course of action for most rockets.  A while back I played with various numbers, launching the same rocket over and over.  You are trading the gains from tipping over earlier with the losses of extra drag--and MechJeb's default actually seems to lie on the conservative side.  Going even hotter than the default turned out better with the 2.5m rocket I was testing, although the gains were minimal.

I recently sent a bunch of birds into a 200km circular orbit preparatory to interplanetary launch.  (MechJeb does not know how to do such burns cleanly and will drive the periapsis down during the burn.  The longer the burn and the lower the altitude the worse this is.)  MechJeb's ascent worked fine--they were going horizontal by 60km, continued to build speed until their apoapsis reached the target and then shut down and coasted until it was time for the circularization burn.  When your rocket is sheathed in fire on the way up it looks wrong.  That doesn't mean it is wrong.

At this point the only time I mess with the defaults is if I'm hauling something very draggy and even those don't have a problem with going horizontal at 60km.

Once I realized it's not garbage I've never had a desire to change the turn altitude other than when launching from airless worlds.  MechJeb can make some ugly orbits when asked for a final altitude well under the turn complete altitude.

[brygun]

*Eyes skew*

So its already set for what it thinks is a sufficiently thin atmosphere?

I know something is a bit off as the launcher doesn't get the payload to orbit altitude on MechJeb but does when I switch to Smart ASS as described above.

I do think I need to nudge the turn end altitude. Useful info but still one question is how does one edit the turn end altitude?

[sarbian]

1 hour ago, brygun said:

how does one edit the turn end altitude?

You press "Edit Ascent path" and then un-check "Automatic Altitude Turn".

You can find some advice about MJ and RO on RO own wiki

[brygun]

Thanks Sarbarian. That was a key part of what I missing. "Oh that's a button" eureka moment. Sometimes the simple answer answers.

Loren also thanks for that input. Its something for me to think about.

Launcher does get to orbit well enough now.

There is a phase where the flight path does fall below the 'target' marker of the ascent guidance. However, the test launch was successful.

I am pondering if a marginal TWR when first staged is a factor. The particular stage at that altitude/speed starts at barely more than 1 but is 2+ by  the time its done. At the higher TWR I didn't see the autopilot attempt to steer back onto the 'target' marker. Still the launch worked with final turn at 150 km for a 200 km orbit.

If it works is it broken? Well perhaps not in the way I thought. Am able to proceed.

Thanks again for MechJeb.

[Loren Pechtel]

5 minutes ago, brygun said:

Thanks Sarbarian. That was a key part of what I missing. "Oh that's a button" eureka moment. Sometimes the simple answer answers.

Loren also thanks for that input. Its something for me to think about.

Launcher does get to orbit well enough now.

There is a phase where the flight path does fall below the 'target' marker of the ascent guidance. However, the test launch was successful.

I am pondering if a marginal TWR when first staged is a factor. The particular stage at that altitude/speed starts at barely more than 1 but is 2+ by  the time its done. At the higher TWR I didn't see the autopilot attempt to steer back onto the 'target' marker. Still the launch worked with final turn at 150 km for a 200 km orbit.

If it works is it broken? Well perhaps not in the way I thought. Am able to proceed.

Thanks again for MechJeb.

A TWR of 1 on an upper stage is ok.  I've successfully flown stuff with a .5 on the upper stage, although MechJeb does poorly with it.

[JonathanPerregaux]

I noticed a problem using MechJeb2 with the Outer Planets mod (OPM). I can plot a porkchop course well enough and it works great. However, beyond a certain orbit (past Jool it seems), things go haywire with the Fine Tune Closest Approach module.

The other day I was heading out to Plock and requested a fine-tune while still near Kerbin, which plopped a course-correction past the orbit of Jool. The course correction looked perfect. However, when this maneuver node executed, the module shut down the engines as expected, then suddenly lost it and fired them full-throttle, displaying thousands of delta-V being required for something mysterious. Once this occurred, it was not possible to use fine-tune anymore. It refused to place a maneuver node anywhere.

Over time, I've noticed this appears to be a problem with the OPM planets (which I often try to visit). Heading to Duna, Dres, or Jool works fine (OPM transplants Eeloo).

Edited February 27, 2017 by JonathanPerregaux

[shibdib]

Figure it out EDIT

Edited February 28, 2017 by shibdib

[Avarax]

Hi i want use mechjeb in career mode without upgrade. I try edit code but i don't know what edit

 

I edit 

 

 MODULE
    {
        name = MechJebCore
        MechJebLocalSettings {
            MechJebModuleCustomWindowEditor { unlockTechs = unmannedTech }
            MechJebModuleSmartASS { unlockTechs = unmannedTech }
            MechJebModuleManeuverPlanner { unlockTechs = unmannedTech }
            MechJebModuleNodeEditor { unlockTechs = unmannedTech }
            MechJebModuleTranslatron { unlockTechs = unmannedTech }
            MechJebModuleWarpHelper { unlockTechs = unmannedTech }
            MechJebModuleAttitudeAdjustment { unlockTechs = unmannedTech }
            MechJebModuleThrustWindow { unlockTechs = unmannedTech }
            MechJebModuleRCSBalancerWindow { unlockTechs = unmannedTech }
            MechJebModuleRoverWindow { unlockTechs = unmannedTech }
            MechJebModuleAscentGuidance { unlockTechs = unmannedTech }
            MechJebModuleLandingGuidance { unlockTechs = unmannedTech }
            MechJebModuleSpaceplaneGuidance { unlockTechs = unmannedTech }
            MechJebModuleDockingGuidance { unlockTechs = unmannedTech }
            MechJebModuleRendezvousAutopilotWindow { unlockTechs = unmannedTech }
            MechJebModuleRendezvousGuidance { unlockTechs = unmannedTech }
        }
    }

 

but dont work. 

[shibdib]

Ascent guidance with anything besides a single engine tube results in a wobble that eventually flips the craft. Even with an AoA of 0, it just starts wobbling at a few thousand feet

Edited February 28, 2017 by shibdib

[Foxster]

28 minutes ago, shibdib said:

Ascent guidance with anything besides a single engine tube results in a wobble that eventually flips the craft. Even with an AoA of 0, it just starts wobbling at a few thousand feet

Picture?

[Loren Pechtel]

54 minutes ago, shibdib said:

Ascent guidance with anything besides a single engine tube results in a wobble that eventually flips the craft. Even with an AoA of 0, it just starts wobbling at a few thousand feet

This makes me suspect that your rocket is unstable.  If the center of mass is unstable it's very prone to wobbling (this has nothing to do with MechJeb, you would find it worse flying it manually.)  A single engine makes it not quite so vulnerable to starting to wobble.

Simple test:   Put your rocket on the pad and simply launch it--don't control it with anything, MechJeb or manual.  See what it does.  If it's stable it should head to space on it's own.  Of course it won't reach orbit without some very careful planning--the question is whether it flies straight without being commanded to do so.

[shibdib]

seems to be the case  few SAS fixed the issue. Been awhile since I've played seems it got a bit more unforgiving

[Avarax]

can you somebody help me please ?

Edited February 28, 2017 by Avarax

[eberkain]

5 hours ago, Avarax said:

Hi i want use mechjeb in career mode without upgrade. I try edit code but i don't know what edit

 

 

[Avarax]

Thx

[WuphonsReach]

5 hours ago, shibdib said:

seems to be the case  few SAS fixed the issue. Been awhile since I've played seems it got a bit more unforgiving

Some fins at the back (or some draggy parts) behind the CoM (center of mass) are also important.  Vernors or linear RCS at the base of the rocket also helps.

I generally put a 1.25 or 2.50 meter reaction wheel at the base of the rocket, with verns/linear-RCS, fins, and the main engine strapped to it.  That lets me swap out the first stage fuel tanks without fiddling with fins or the RCS.  I'll usually auto-strut that to the "root" or "grand-parent" part for stability.

-------

I'm currently trying to figure out why my rockets fly well from 0-50km altitude, then start tail wagging as the trajectory flattens out in the upper atmosphere.  The main stage is almost empty of fuel, so it's lost a lot of its mass.  The thrust value is still in the 1.0-2.5 TWR range.  I'm thinking that maybe it's MJ trying to get back to inclination zero, but I'm not sure.

F12 to show aerodynamics doesn't show much at 50-60km altitude.

[Starwaster]

29 minutes ago, WuphonsReach said:

I'm currently trying to figure out why my rockets fly well from 0-50km altitude, then start tail wagging as the trajectory flattens out in the upper atmosphere.  The main stage is almost empty of fuel, so it's lost a lot of its mass.  The thrust value is still in the 1.0-2.5 TWR range.  I'm thinking that maybe it's MJ trying to get back to inclination zero, but I'm not sure.

F12 to show aerodynamics doesn't show much at 50-60km altitude.

What is your target altitude? I've seen that behavior but suspect that it's more to do with hitting the final stage of its ascent. I submitted some tweaking of the ascent mode quite awhile back and IIRC, what it's supposed to stay basically neutral with regards to attitude, either staying oriented to its last desired target heading or zeroing its AoA. (I forget which)

But I think a few months back some pull requests might have changed behavior in the final part of ascent.

[sebi.zzr]

@sarbian,i just got the chance to test dev.build #678 and diferential throttle works fantastic.
Thank you very much.