[1.0] XT Landertron - Smart Retrorockets for Landers and Spaceplanes v0.08 Oct 10

[komodo]

9 hours ago, Kerbas_ad_astra said:

I'd be happy to accept a PR, but I'm curious about why you changed some functions to be "virtual".  (I have some experience programming, but I'm not very familiar with C# specifically.)

Also, I'll need to get some modeling experience to add a model for the landertron-box, although I guess I could just add it to all probe cores with a Module Manager patch.

(I'd be interested in seeing this implementation too, to be honest!)

I wonder if you couldn't hijack a current model as a stand-in. Maybe something like an octo2 core set to surface mount only? The larger battery?; this mod did that originally for the retros themselves, with welded parts to stand in.

Incidentally, does anyone have a copy of landertron old enough to have said cfgs in it? I can't locate it on my ,  WAIT, I found a old copy on my laptop, back when it could run KSP (0.90 era). Apparently there was one welded part, a decoupler with ant engines, a separatron copy and an RT-10 copy. Hm. Still, if anyone else is looking for them, I can load them somewhere, as I was unable to find them online last time I went looking. They were rather handy, being somewhat smaller than the current batch of admittedly really cool parts.

[Booots]

On 2016-03-28 at 4:49 PM, Kerbas_ad_astra said:

I'd be happy to accept a PR, but I'm curious about why you changed some functions to be "virtual".  (I have some experience programming, but I'm not very familiar with C# specifically.)

Also, I'll need to get some modeling experience to add a model for the landertron-box, although I guess I could just add it to all probe cores with a Module Manager patch.

The 'virtual' keyword just allows for derived classes to override that method with an identically-named one in their instance. Basically, it lets me create something that, in the eyes of the landertron controller module, behaves exactly like any other landertron but has different underlying logic (ie. for calculating available thrust since I can't just look at the engine on the same part). I'm submitting a PR now, so you can see what I mean in the code.

The SmartParts pack has a handy altimeter part that I've appropriated through a MM file to be my stand-in. Maybe just get permission and retexture?

[Kerbas_ad_astra]

Good news!  Landertron compiles against the 1.1 prerelease binaries and it works.  I'll wait until the full release comes out before releasing 0.13, but I'll have it pushed to GitHub before too long.  Unfortunately, it probably won't have a part to implement the LandertronBox module, but until I figure out how to make my own part, I might weld together the spectro-variometer with the small reaction wheel.  (Smart Parts is -NC licensed, which is incompatible with the other model licenses.)

[Kerbas_ad_astra]

And here's the XT-LB Landertron Box, coming in Landertron 0.13!

[Booots]

34 minutes ago, Kerbas_ad_astra said:

And here's the XT-LB Landertron Box, coming in Landertron 0.13!

That looks great! I kinda like the hacked-together look of it. Could I recommend rescaling it down so it can fit on small 0.625m landers without looking like they're about to fall over?

[Kerbas_ad_astra]

It's actually already half-sized compared to the original reaction wheel.  I didn't mean for it to be so, but that's how it came out of UbioZur.  It's about the size of a Mystery Goo cannister:

[Kerbas_ad_astra]

Version 0.13 is out!  Featuring 1.1 support and the new Landertron Box.

[austinryan95]

Quote

 

Thank's for the 1.1 update. One of My favorite mod's to use with remote tech. Keep up the good work!

 

[ManuxKerb]

Hi

Thanks for a version who does work in 1.1 BUT in my install (stock but Landertrons) they fire much too late and it always crashes into the ground. I would guess it has something to do with the sea level and the actual ground level.

It says for eg.65 meters above sea level (graslands grounds) and just before that it starts to fire. I belive if it had the 65 m it would work but as the ground is that high it just crashes into it. 

If needed i can provide craft file....

 

Edited April 25, 2016 by ManuxKerb

[Kerbas_ad_astra]

Please do provide a craft.  I've tested the landertrons at KSC (ground level is about 70 meters ASL), and they work fine (the program works by raycasting to terrain, so it should work no matter the ground's altitude).

[ManuxKerb]

Hey again,

I made a new texture for the Landertron box and also added search tags into the parts.

Feel free to copy my improvements into the main branch! :-)

Picture:

http://postimg.org/image/knip8gsmv/

Download:

http://s000.tinyupload.com/index.php?file_id=60382921107693955584

So i did check the Download -> Link and File work 100 % on my side. Did successful download and extract it. Please try it from a desktop machine. On my phone it did not work.

Have fun!

Craft file for not working fireing : http://s000.tinyupload.com/index.php?file_id=38836306184604945378

Edited April 25, 2016 by ManuxKerb

[Cheesecake]

Quote

Download:

http://s000.tinyupload.com/index.php?file_id=60382921107693955584

Download does not work. The Folder has only 9KB and WinZIP says the Zip is broken.

[Booots]

35 minutes ago, ManuxKerb said:

I made a new texture for the Landertron box and also added search tags into the parts.

Feel free to copy my improvements into the main branch! :-)

That looks good! I like it!

1 hour ago, ManuxKerb said:

Thanks for a version who does work in 1.1 BUT in my install (stock but Landertrons) they fire much too late and it always crashes into the ground. I would guess it has something to do with the sea level and the actual ground level.

It says for eg.65 meters above sea level (graslands grounds) and just before that it starts to fire. I belive if it had the 65 m it would work but as the ground is that high it just crashes into it.

Are you coming in really fast on an atmospheric planet, like terminal velocity or higher?

The current model for when to fire does have a problem with high approach speeds on atmospheric planets because it assumes that the outside acceleration on the craft is constant from start of fire to finish which causes it to overestimate the effect from drag as that will decrease as the craft slows down. Once I'm through the end of the semester I'm going to try to write a new dynamic model that will estimate the craft's ballistic coefficient to better account for that.

[ManuxKerb]

43 minutes ago, Cheesecake said:

Download does not work. The Folder has only 9KB and WinZIP says the Zip is broken.

Shoot  i will reupload when i finish work in ca 6 hours ...

16 minutes ago, Booots said:

That looks good! I like it!

Are you coming in really fast on an atmospheric planet, like terminal velocity or higher?

The current model for when to fire does have a problem with high approach speeds on atmospheric planets because it assumes that the outside acceleration on the craft is constant from start of fire to finish which causes it to overestimate the effect from drag as that will decrease as the craft slows down. Once I'm through the end of the semester I'm going to try to write a new dynamic model that will estimate the craft's ballistic coefficient to better account for that.

Nope not that fast like 600 m/s or so. I just launch from pad like 5-10 km high and then let it the second stage fall back to ground... 

Does the craft file DL work?

Edited April 25, 2016 by ManuxKerb

[drtedastro]

Just tried the update and all worked great.

Tested the HGR "Soy-Juice" pod, dropping it from a tall gantry and the little rockets ignited and set it down nice a soft.

Thanks for the update and the Mod.

Cheers.

[Kerbas_ad_astra]

I've got the craft file.  When I test it, the landing burn seems hit-or-miss -- it sometimes fires in time (losing some of the legs and engines in the process), sometimes not.  I would not recommend a landertron for landing at speeds on the order of several meters per physics tick (600 / 60 = 10 m/frame), where things can go very wrong if the calculation (or the game) skips a beat.

Also, how are you getting to over 600 m/s?  I was doing about 150 when the engines fired.  I'm not sure I'd ever call 600 m/s "not that fast" -- that's flirting with Mach 2.

[Booots]

12 minutes ago, Kerbas_ad_astra said:

 I'm not sure I'd ever call 600 m/s "not that fast" -- that's flirting with Mach 2.

So after thinking about this more, the only way to get perfect landings every time for every craft would necessitate implementing an ODE solver. Fortunately, I think the guys over at the Trajectories mod do something like that for calculating atmospheric trajectories. I wonder if we could make their system work for us without adding too much lag to the game. Although I could handle the maths for that sort of ODE, actually implementing it in-game is beyond my modding abilities.

As a halfway solution, we could partially fix the drag estimation by using an average drag value (which turns out to be just a_drag_I/3) instead of the initial value. A time-averaged value would be better, but that too would need an ODE solved, I think. We would have to split the outside acceleration term in the code into gravity and drag components, but that shouldn't be too hard because gravity is nicely predictable.

[StevieC]

so I just drag the "Landertron" folder into GameData?

[ManuxKerb]

30 minutes ago, StevieC said:

so I just drag the "Landertron" folder into GameData?

yep :-)

[ManuxKerb]

3 hours ago, Booots said:

So after thinking about this more, the only way to get perfect landings every time for every craft would necessitate implementing an ODE solver. Fortunately, I think the guys over at the Trajectories mod do something like that for calculating atmospheric trajectories. I wonder if we could make their system work for us without adding too much lag to the game. Although I could handle the maths for that sort of ODE, actually implementing it in-game is beyond my modding abilities.

As a halfway solution, we could partially fix the drag estimation by using an average drag value (which turns out to be just a_drag_I/3) instead of the initial value. A time-averaged value would be better, but that too would need an ODE solved, I think. We would have to split the outside acceleration term in the code into gravity and drag components, but that shouldn't be too hard because gravity is nicely predictable.

Hi,

Ok, so after the first shock that landertron does not work really good (as far as i can gather from your talks) please clarify for me, how i'm supposed to use it correctly.

My idea was to ditch the parachutes on various planets with athmo and just use the landertrons to land in style (freefall till landertrons take over...). Will it work on worlds without atmosphere better ?

So here is my test setup:

3 man capsule, 4 landertrons, heatshield, battery and reaction wheel. Put in semi orbit up to 71 km. Let fall back to kerbin. Result: Speed about 500 m/s 5 km from water surface, landertrons fire like 300 m before water surface.  Speed around 200 m/s, All kerbals dead.....

So how im supposed to use it to get kerbals safely down ?

Thanks

Edited April 25, 2016 by ManuxKerb

[Kerbas_ad_astra]

In my Kerbin testing, I've found that they're quite reliable at a vertical speed around 50 m/s.  The discrete models (i.e. not the box) can do a little better, since they can calculate their delta-V capacity and firing times a little better, because there's no need to worry about fuel transfer.  It's also possible to manually fire the engines with an action group, if it's the timing you're concerned about.  (The short-landing and stay-put modes are much less sensitive to this sort of thing.)

For your plunge-from-orbit test -- first of all, your kerbals probably would have died with chutes as well, as your capsule probably would not have slowed down enough to open parachutes safely before hitting the ground (I speak from personal experience -- I've lost a few tourists that way).  The Mercury capsules went 190 km up, and 485 km downrange.  Drogue chutes plus some landertrons might have done the trick, since drogues can be opened at much higher speeds, ~500 m/s.

Landertrons (in soft-landing mode, anyway) are probably best applied like the retro-rockets on the Soyuz capsule or Mars landers -- a final automatic burn to take off whatever speed isn't handled by parachutes (even a few meters per second is a pretty nasty bump for many parts).  On Duna, this can be quite substantial, and I've seen lots of people swear by landertrons when they work with RemoteTech and cannot keep their hand on the throttle for the landing due to command lag.

[Booots]

I recently did a landing on Ike using RemoteTech where I scheduled a manoeuver to cancel my horizontal and all but 50m/s or so of my vertical and let the Landertron Box take care of the final touchdown. Landing on Duna was actually harder because I couldn't plan a manoeuver like that in the atmosphere but my drogue wasn't big enough to cancel my horizontal velocity and the Landertrons kept firing either too soon or way too late. They're not magical, but they can be very useful, especially with RemoteTech.

[gendalf]

Kerbas_ad_astra, why not make a separate forum topic if you've taken over the mod/can't update OP?
Is it still dependent on B.Dynamics mod ?

Edited April 27, 2016 by gendalf

[Kerbas_ad_astra]

I haven't made a new topic yet because I didn't really properly "take over" until now -- this is the first time that I've made two releases in a row.

All of the models are self-contained now (except for the landertron box, which references a couple of Squad's models), so there's no dependency on BD's mods.

[StevieC]

So, does the last build of LanderTron still work in KSP v1.1.2?