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

[JoshuaTheBestGuy]

7 hours ago, Brigadier said:

Do you have ModuleManager installed?

You know, that might be why.

[Brigadier]

5 hours ago, JoshuaTheBestGuy said:

You know, that might be why.

It is a dependency.

[sarbian]

New release MechJeb2-2.11.0.0.zip

• Landing AP fixes and improvements
• Plane landing AP fixes and improvements
• Rover AP fixes and improvements
• Updated Chinese loc

[Shoujo Q]

3 hours ago, sarbian said:

• Landing AP fixes and improvements

Does this fix the bug where Landing AP would not engage until after a save and reload, or until Mechjeb settings were reset to default?

[sarbian]

Yes.

[ThomasJoseph]

3 hours ago, sarbian said:

New release MechJeb2-2.11.0.0.zip

• Landing AP fixes and improvements
• Plane landing AP fixes and improvements
• Rover AP fixes and improvements
• Updated Chinese loc

Does this fix the frame rate drop on landing? Whenever I use the landing guidance, the frame rate drops to 2-3fps. I was just about to file a bug report on it. 

LMK,

Thanks!

[sarbian]

1 hour ago, ThomasJoseph said:

Does this fix the frame rate drop on landing? Whenever I use the landing guidance, the frame rate drops to 2-3fps. I was just about to file a bug report on it. 

LMK,

Thanks!

It does.

[ThomasJoseph]

1 hour ago, sarbian said:

It does.

Thanks for being on top of it! You rock!

[xTheCanadian]

Thanks buddy! You're the best! I'm not smart enough to play this game without my good buddy MechJeb!

[Jebidiah kerman]

i cant go into iva. is this intentional, or is it a bug?

[Brigadier]

16 hours ago, Jebidiah kerman said:

i cant go into iva. is this intentional, or is it a bug?

Why do you think this is a MechJeb problem?  You may want to read the topic linked in my sig block on how to report problems and provide some more info.  You haven't provided much to go on here.

And, if you read the first post, you'll see that Sarbian does not provide support if you don't post your logs.  Again, see that post linked below.

[antipro]

hi, I would like to ask for a feature:
when you use the "Landing Guidance" window and hit the "Pick Target on Map" button a yellow "triangular" cursor shows Coordinates and Biome. Then you choose the position and it become red.

Can you just add another information, the slope, with that yellow cursor?
it would be extremely useful.

Edited September 1, 2020 by antipro

[Gilph]

Hi, I've been noticing an issue with the Rendezvous AP.

I am trying to rendezvous with another object around 10-15 km away. Our planes and orbit ecc are matched, and a simple Hohmann transfer will intercept very nicely using only about 10 m/s. But, the AP will decide to say "close to target, plotting intercept" and use hundreds of deltav in a loop, trying to get closer. Usually when that happens, i can disable AP and do the hohmann transfer, warp until about 2 km away, and reenable ap to match velocities. That doesn't seem to work anymore and I have to manually match the velocities.

Is there a way to determine if a hohmann transfer will create a separation close to zero and prefer that method over the close to target intercept? Or, detect that I'm already on that trajectory and will just match velocities at the end?

Thanks

[DerekL1963]

1 hour ago, antipro said:

Can you just add another information, the slope, with that yellow cursor?

I do not believe slope is available natively (directly) in-game.  I think SCANSAT can do what you ask.

[Gilph]

12 hours ago, antipro said:

hi, I would like to ask for a feature:
when you use the "Landing Guidance" window and hit the "Pick Target on Map" button a yellow "triangular" cursor shows Coordinates and Biome. Then you choose the position and it become red.

Can you just add another information, the slope, with that yellow cursor?
it would be extremely useful.

 

11 hours ago, DerekL1963 said:

I do not believe slope is available natively (directly) in-game.  I think SCANSAT can do what you ask.

Hi, I've had come conversations with @DMagic on the subject of slope. Slope is not a native data element, but it can be calculated. In the context of Scansat, it's done in a certain way to support their scanning functionality. They take lat and lon and get ASL data, then get the ASL data for adjacent locations and calculate the differences for ASL to calculate slope.  It sounded like if you have access to that  ASL data, anyone can calculate slope, but it appears not to be a trivial task.

[HansAcker]

40 minutes ago, Gilph said:

Slope is not a native data element, but it can be calculated

KerbalEngineer continously scans the ground to display the slope directly under the craft. I don't know how expensive these tests are but that could be an option for the target picker (where to display the value, though?) I wonder if the surface slope returned by a scan from orbit accurately reflects the high-resolution surface on landing or if it's just a rough "scaled-space surface" value.

TCA's autolander does it a bit different and scans the ground for a flat surface after the landing site has been selected.

Then again, do you really want to target a flat landing site automatically? Let Jeb stake out good sites and note the coordinates plant flags

Edited September 1, 2020 by HansAcker

[SciMan]

16 hours ago, Gilph said:

Hi, I've been noticing an issue with the Rendezvous AP.

I am trying to rendezvous with another object around 10-15 km away. Our planes and orbit ecc are matched, and a simple Hohmann transfer will intercept very nicely using only about 10 m/s. But, the AP will decide to say "close to target, plotting intercept" and use hundreds of deltav in a loop, trying to get closer. Usually when that happens, i can disable AP and do the hohmann transfer, warp until about 2 km away, and reenable ap to match velocities. That doesn't seem to work anymore and I have to manually match the velocities.

Is there a way to determine if a hohmann transfer will create a separation close to zero and prefer that method over the close to target intercept? Or, detect that I'm already on that trajectory and will just match velocities at the end?

Thanks

I find I get the results I want much more consistently if I use the Rendezvous Planner and just hit the right buttons in the right sequence, instead of letting the Rendezvous AP handle things (which does the same thing, but with you theoretically only having to push one button).

What you would do when using the rendezvous planner is hit the Match Planes button, execute the node, then Hohmann transfer, execute the node, use RCS to trim the closest approach to ~ 150m (trim this value to suite the size of the thing you're intercepting, further away for bigger things), then hit Match Velocity at Closest Approach, and execute that node, and that should take care of things.

I'm sure you already know this, but for everyone else: From there you should be able to dock by setting a docking port part on the target craft as your target, and then select "control from here" on the docking port you wish to use on the craft you're controlling, then go to the Smart A.S.S. Target mode, and select "Par-". Then you should be able to use just the translation RCS controls to dock manually assuming that the other craft is not rotating. That's how I do it, I find that the docking AP is entirely not needed. Of course I started playing KSP already knowing the general ideas of how to dock spacecraft, since I played Orbiter previously to KSP. You might want to watch the docking AP do it a few times, but I can almost guarantee that you will be capable of doing a better job with some practice.

EDIT: Of course, I'm not saying there isn't a problem with the Rendezvous AP code, I guess what I'm doing is proposing a workaround until that issue gets addressed.

Edited September 1, 2020 by SciMan

[chd]

FYI, "land at target" seems to be still badly broken for me. Set a target, "land at target," and it's half a planet off.

[Tonka Crash]

22 minutes ago, chd said:

FYI, "land at target" seems to be still badly broken for me. Set a target, "land at target," and it's half a planet off.

Through atmosphere or vacuum? I've done 4 landings on Mun since the last release, 2 different landers within 5m of the target and 2 with another lander that was about 50m off, but consistent with how it always behaves for that ship.

[darthgently]

On 9/1/2020 at 3:46 PM, SciMan said:

...

What you would do when using the rendezvous planner is hit the Match Planes button, execute the node, then Hohmann transfer, execute the node, use RCS to trim the closest approach to ~ 150m (trim this value to suite the size of the thing you're intercepting, further away for bigger things), then hit Match Velocity at Closest Approach, and execute that node, and that should take care of things..

That is exactly the rendezvous method I've converged on by trial and error and it works great. It is amazing the kind of fine tuning that can get done to a future trajectory with just a few dv of rcs if you catch it early. One of my faves is to get an encounter with, say Minmus, then before you've even gotten a few kliks from Kerbin, tease the patch at Minmus into a polar orbit injection just using rcs translation moves. The trick is to sas hold prograde, roll so Kerbin is to your left and the top of the ship is up, or normal. Then set focus to Minmus and zoom in, hit caps lock to get into rcs fine mode, then use the starboard/port translations like radial in/out, and dorsal/ventral translation for norm/antinorm. Fore and aft for pro/retro obviously

 

Oh yeah, same technique for injecting into a retrograde orbit insread of pro for those rescue contracts where the target is orbiting the other way

Edited September 3, 2020 by darthgently

[Gilph]

On 9/1/2020 at 3:46 PM, SciMan said:

What you would do when using the rendezvous planner is hit the Match Planes button, execute the node, then Hohmann transfer, execute the node, use RCS to trim the closest approach to ~ 150m (trim this value to suite the size of the thing you're intercepting, further away for bigger things), then hit Match Velocity at Closest Approach, and execute that node, and that should take care of things

It's easier for me just to use Maneuver Planner to work around it. The only really tricky part is if the approach is so exact and I don't slow down in time, So when I do the match velocities at closest approach, i open the maneuver node editor and subtract 10 secs.

It seems like, recently, it will pick the "wrong" choice more often. I could even disable AP, do the workarounds to get like 1km away, enable AP, and it would make the node to match velocities and maintain the distance i specified, but that doesn't seem to work anymore. I look at some of the recent revisions in dev and noticed there were a few updates.to Rendezvous AP that streamlined things a bit, which may be causing it.

It's not terribly impactful, just need to validate the AP choices before going to the bathroom and finding out my vessel is out of fuel when I get back.

[chd]

On 9/1/2020 at 9:35 PM, Tonka Crash said:

Through atmosphere or vacuum? I've done 4 landings on Mun since the last release, 2 different landers within 5m of the target and 2 with another lander that was about 50m off, but consistent with how it always behaves for that ship.

Vacuum; this was on Minmus. However, there was an update to MJ since then, and I haven't verified whether I can recreate the same behavior with the current update.

[JJJacksonTyler]

"Intercept Target At Chosen Time" feature still broken. Spits out insane results. 

[DMagic]

On 9/1/2020 at 8:54 AM, Gilph said:

Slope is not a native data element, but it can be calculated.

Slope is one of those tricky things that can get pretty complicated. It's highly localized and there are many different ways of calculating the value and reporting the value.

SCANsat just asks KSP for the terrain height at a few points surrounding the center point, I think something like 5m distance. This is a little bit expensive to calculate, but not really significant in the scheme of things. (This is for the little instrument data window, the slope map is another matter, and is generally far less accurate).

I think Engineer looks at the angle of the actual surface geometry relative to a flat surface. This requires that the terrain geometry (this could have changed at some point) is drawn, which means it will only work close to the surface. But it is fairly cheap to do and should give a precise value for a given point. 

[DerekL1963]

When meausuring distance to a target between a target and a craft in atmospheric flight, it is the true distance (measured along the surface) or the chord?