How to nail these high-precision, low-twr, super-efficient landings?

[Carl]

Quote

Guys, thanks, but as I said, I worked out the formula I quoted above, and it has gotten me very close to the precision I want. I now just burn 1km earlier than the formula provides, and then eye-ball the throttle when I'm down to the last hundred of meters of velocity above the target zone. My record today was 76 meters away from the target marker (103 second burn). I don't see why I should bring more fuel or accept worse accuracy if this formula +1km gives such great results  Of course, it would be better if I could get an auto-pilot to execute that burn when the distance from target hits the threshold, but even manually, it is close enough for practical uses.

 Like this

By manually eyeballing the last few hundred meters your doing exactly what we said, accepting a less efficient profile to get better results.

[Spricigo]

aluc24

 

Nice to see you found a working solution. 

Don't mind if we keep the discussion going.  It may be useful for other people in similar situations and part of the fun. 

[aluc24]

4 hours ago, Carl said:

By manually eyeballing the last few hundred meters your doing exactly what we said, accepting a less efficient profile to get better results.

Well, I mean, the only sacrifice in performance is that I set the throttle to 90% instead of 100% in the last few kilometers of distance... I do the whole burn and landing in a single, continuous burn. The loss of performance is marginal, I think. But technically you are correct  All I wanted was to work out a way to make this calculation so that I don't have to eyeball the whole burn.

[Kerbart]

On 1/15/2017 at 11:18 AM, aluc24 said:

Sometimes I get contracts to land in a very specific spot on Mun, and then do some science there... Thus my need for precision.

Aren't those shown as targets on the navball? Now you have something to aim for! 

[Carl]

Quote

All I wanted was to work out a way to make this calculation so that I don't have to eyeball the whole burn.

The problem is those errors i mentioned earlier are gong to make every decent subtly different you'll only really know for sure what data to put into any calculation after you start the burn, and of course as the burn continues and your angle of descent changes as you continue the descent the variables change. It's not really a sum you can do in advance and doing it mid flight would be very tough. There's a reason computers are preferred for this. Honestly though, building a new design and putting a rover down would be far and away the better option, you can easily get a landing within a few hundred meters as you've allready shown, a rover could then haul you the rest of the way. I'm not sure why your having issues with MJ btw. The only reasons i've seen the landing autopilot screw it up is:

 

1. Too low a starting orbit, you really want to be in a 50km+ orbit for preference.

 

2. Setting an acceleration limit, MJ can;t work that into the sums it uses so it underestimates the burn.

Edited January 16, 2017 by Carl

[aluc24]

11 minutes ago, Kerbart said:

Aren't those shown as targets on the navball? Now you have something to aim for! 

Well, they are showing up in navball, but they aren't considered legit targets - BurnTime addon doesn't see it as rendezvous, and it doesn't provide the calculation to time the burn start.

 

10 minutes ago, Carl said:

The problem is those errors i mentioned earlier are gong to make every decent subtly different you'll only really know for sure what data to put into any calculation after you start the burn, and of course as the burn continues and your angle of descent changes as you continue the descent the variables change. It's not really a sum you can do in advance and doing it mid flight would be very tough. There's a reason computers are preferred for this. Honestly though, building a new design and putting a rover down would be far and away the better option, you can easily get a landing within a few hundred meters as you've allready shown, a rover could then haul you the rest of the way. I'm not sure why your having issues with MJ btw. The only reasons i've seen the landing autopilot screw it up is:

 

1. Too low a starting orbit, you really want to be in a 50km+ orbit for preference.

 

2. Setting an acceleration limit, MJ can;t work that into the sums it uses so it underestimates the burn.

Well, yes, my orbit before the burn is usually 5x5km (sometimes lower if there are no obstacles). 50km+? That's wasting a ton of fuel...

As for the rover, yeah, it can be made, but what if I want to make a Mun base later on? Flying new modules in will require accuracy down to 10-20 meters (unless I stick rover wheels to everything, which is kind of silly).

Edited January 16, 2017 by aluc24

[Carl]

Going from 50x50 to 5x5 uses fuel too so efficiency wise the difference is minimal, (it may be nonexistent but i'd have to sit down and think through the physics involved, it's complicated).

[Kerbart]

Just now, aluc24 said:

Well, they are showing up in navball, but they aren't considered legit targets - BurnTime addon doesn't see it as rendezvous, and it doesn't provide the calculation to time the burn start.

Aaah, well, that's where practice makes perfect.

• I tend to keep the target marker halfway between the center of the navball and the retrograde velocity marker as you're bringing back your horizontal velocity to zero (so the marker and the velocity vector are in the center of the navball)
• For horizonal distance you can use the equation d = ½×a×t² which of course solves for t = √2×d/a, multiply t with your horizontal velocity and you'll know the horizontal distance (throw in Pythagoras and your altitude and you know at what total distance you have to burn). Of course your deceleration will slightly increase as you burn off fuel, but it's still good enough for practical purposes (and you'll have a buffer for overshooting)
• If you're using another craft as a target and you practice this a lot, it will pay off to pay attention to where you land. I've borked a landing or two by landing right on top of the target (not thinking that I would I wasn't paying attention). So that answers the "how accurate" question I guess.

[Spricigo]

1 minute ago, aluc24 said:

As for the rover, yeah, it can be made, but what if I want to make a Mun base later on? Flying new modules in will require accuracy down to 10-20 meters (unless I stick rover wheels to everything, which is kind of silly).

If you want a Mun base later on prepare yourself for deal with a lot of other issues. e. g: coupling two modules together can be way more challenging than landing, because off several issues of docking port alignment/placement (I really don’t see one struggling to archieve such precision landing to be satisfacted with just klawing the modules together) . Also, at this point consider some modules to give the base some objective since the stock game lacks one (except the 'I want/should be cool' one)

Quote

That's wasting a ton of fuel...

 

Convenience has a cost.

But even at such inefficient situation how much your launch cost would increase? How much time/effort can be saved by skipping fine mid-course adjustments?  The interesting thing about efficiency its that different people will asses it in different ways. You may consider launch cost, other will count on recovering the vessel and count just the fuel cost and others yet  will just consider just how much game time.

As everything in the life, its a question of cost/reward. just that everyone have a different way to calculate it.

[aluc24]

On 1/16/2017 at 7:44 PM, Kerbart said:

Aaah, well, that's where practice makes perfect.

• I tend to keep the target marker halfway between the center of the navball and the retrograde velocity marker as you're bringing back your horizontal velocity to zero (so the marker and the velocity vector are in the center of the navball)
• For horizonal distance you can use the equation d = ½×a×t² which of course solves for t = √2×d/a, multiply t with your horizontal velocity and you'll know the horizontal distance (throw in Pythagoras and your altitude and you know at what total distance you have to burn). Of course your deceleration will slightly increase as you burn off fuel, but it's still good enough for practical purposes (and you'll have a buffer for overshooting)
• If you're using another craft as a target and you practice this a lot, it will pay off to pay attention to where you land. I've borked a landing or two by landing right on top of the target (not thinking that I would I wasn't paying attention). So that answers the "how accurate" question I guess.

Good advice. Thanks. I have yet to land on top of my target, but I'm getting there.

 

On 1/16/2017 at 8:26 PM, Spricigo said:

But even at such inefficient situation how much your launch cost would increase? How much time/effort can be saved by skipping fine mid-course adjustments?  The interesting thing about efficiency its that different people will asses it in different ways. You may consider launch cost, other will count on recovering the vessel and count just the fuel cost and others yet  will just consider just how much game time.

As everything in the life, its a question of cost/reward. just that everyone have a different way to calculate it.

Well, I agree, I guess I'm just a kind of precision freak, I like to plan missions in advance and do them lean. More challenging than the usual "moar boosters" approach. I still pack a 100m/s extra dV for every mission, but rarely more than that. I just like knowing that I performed something as efficiently as it is reasonably possible.

[Spricigo]

13 minutes ago, aluc24 said:

Well, I agree, I guess I'm just a kind of precision freak, I like to plan missions in advance and do them lean. More challenging than the usual "moar boosters" approach. I still pack a 100m/s extra dV for every mission, but rarely more than that. I just like knowing that I performed something as efficiently as it is reasonably possible.

I know what you mean. Sometimes I also want something well planned and well executed. But often I don’t want to cut the last bit of unneeded propellant in my design or to figure out the most efficient maneuver so I can do one more reposition mission with some satellite. Good think its that, for me at least, KSP is fun in both occasions.

[aluc24]

Just now, Spricigo said:

I know what you mean. Sometimes I also want something well planned and well executed. But often I don’t want to cut the last bit of unneeded propellant in my design or to figure out the most efficient maneuver so I can do one more reposition mission with some satellite. Good think its that, for me at least, KSP is fun in both occasions.

Yeah, I do that quite a lot too. Huh, maybe I should try Realism Overhaul when it's updated to 1.2.2  I heard every meter of dV is precious as silver there. Sounds like my kind of overhaul.