Jump to content

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


Recommended Posts

I am playing a new career in 1.2.2, and I made things really difficult for myself. I am trying to land on the Mun in a very specific spot (within 100m of target), using as little fuel as possible, using low thrust engines (TWR of 1.5 at Moon). First, I set the inclination so that the orbit goes over the target location exactly (using Trajectories addon to account for Mun rotation), then lower the periapsis above target location so that my trajectory, clears all obstacles just by 100 meters or so. And final part is executing the burn at the right time. This is the part I can't figure out.

I set up a maneuver node, using Trajectories again to tell that my trajectory will hit the landing site after the burn is complete. The node it tells me that my horizontal velocity kill burn will take 98 seconds (570m/s). But even if I burn at half-burn time (49 seconds away from node), I overshoot the target site. Obviously, I need to burn earlier. But when?

Something tells me this should be pretty simple math. I have a marker at the landing site, so I can always tell the distance. How do I figure out the distance from landing site when I need to start that 98 second burn so that I end up right over target when the burn is done?

P.S. Oh, and I tried Mechjeb autopilot. Crashes the ship half the time, and it isn't nearly as precise as I need it to be.

P.P.S. Maybe Kerbal Engineer has some readout that would help with this? I found suicide burn timings, but not ones for maneuver nodes.

Edited by aluc24
Link to comment
Share on other sites

Hey,

I use mechjeb too, only just for the calculation, if you make a custom window with suicide burn and with an impact time, you'll see the time you've left before impacting the body, the suicide burn number gives you the time the suicide burn is finished on impact. So when it gives you 5 seconds, you're hovering above the body 5 seconds before impact.
When it gives you -5 seconds, it says you're impacting the body just 5 seconds before you finish the suicide burn.

 

I usually time it on 1 seconds before impact 
 

Link to comment
Share on other sites

3 minutes ago, DrLicor said:

Hey,

I use mechjeb too, only just for the calculation, if you make a custom window with suicide burn and with an impact time, you'll see the time you've left before impacting the body, the suicide burn number gives you the time the suicide burn is finished on impact. So when it gives you 5 seconds, you're hovering above the body 5 seconds before impact.
When it gives you -5 seconds, it says you're impacting the body just 5 seconds before you finish the suicide burn.

 

I usually time it on 1 seconds before impact 
 

But I'm not talking about killing the vertical velocity - I'm talking about the horizontal one.

Link to comment
Share on other sites

Just now, aluc24 said:

But I'm not talking about killing the vertical velocity - I'm talking about the horizontal one.

You kill both with them, suicide burn means you kill you're velocity, when impacting on a body, your suicide kills both a long as you aim retrograde :) 

Yep of course you need to burn yourself into a impact trajectory. But after that mechjeb calculates everything for you. The time on impact calculation is inter graded somehow into the suicide burn calculations, which gives you a very precise timing of burning. Also, you can always orientate your craft a a bit lower or higher, which results in extending or lower your time of burning and impact.
 

And when you want to land on the right place, that's still a difficult one, depends how large the body is and how long you have to burn. But I believe trajectories mod does take account for that (not sure)

Link to comment
Share on other sites

4 minutes ago, DrLicor said:

You kill both with them, suicide burn means you kill you're velocity, when impacting on a body, your suicide kills both a long as you aim retrograde :) 

Yep of course you need to burn yourself into a impact trajectory. But after that mechjeb calculates everything for you. The time on impact calculation is inter graded somehow into the suicide burn calculations, which gives you a very precise timing of burning. Also, you can always orientate your craft a a bit lower or higher, which results in extending or lower your time of burning and impact.
 

And when you want to land on the right place, that's still a difficult one, depends how large the body is and how long you have to burn. But I believe trajectories mod does take account for that (not sure)

Well, impact trajectories are not exactly fuel-efficient... And if I make it fuel-efficient by making the impact trajectory very shallow, it is difficult to tell if some mountain or hill won't be in the way. As I said, I usually make my deorbits just dozens of meters above ground, for efficiency.

Is there another method that doesn't require impact trajectory? I remember using a formula back from my Orbiter days, s = vt / 1.5, where s equals braking distance (or distance from target site), v is initial speed, and t is burn time (I can get that from maneuver node). For some reason it doesn't work for me, I always land short of the site.

Link to comment
Share on other sites

You need to consider it as a high speed rendezvous,  make sure you will be near enough you target and at closest approach burn to kill velocity. I'm assuming MJ have readouts for time to closest approach and burn time to kill velocity. 

 

Link to comment
Share on other sites

9 minutes ago, Spricigo said:

You need to consider it as a high speed rendezvous,  make sure you will be near enough you target and at closest approach burn to kill velocity. I'm assuming MJ have readouts for time to closest approach and burn time to kill velocity. 

 

I can't find it... I probably haven't unlocked that part of Mechjeb functionality in my career yet. Does Engineer have this readout? I can't find it either.

Link to comment
Share on other sites

2 minutes ago, Spricigo said:

This one have it in a simple and elegant way :

Like I said,  assumed MJ and KER had it also,  but can't check it at the moment 

Alright, but I assume I need a flag over there to have it as target? What if I only have a waypoint (I'm using Waypoint Manager)?

Edited by aluc24
Link to comment
Share on other sites

Yes,  you need something to target.  I was assuming that was available.  

At this point I really see no reason to such precision, other than the challenge of it. So I have only 2 answers left:

a) accept you won't be able to archive total 

b) practice, practice, practice 

 

Link to comment
Share on other sites

19 minutes ago, Spricigo said:

Yes,  you need something to target.  I was assuming that was available.  

At this point I really see no reason to such precision, other than the challenge of it. So I have only 2 answers left:

a) accept you won't be able to archive total 

b) practice, practice, practice 

 

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

What about manual calculation? I mean, if the burn time is known, then it should be simple math to determine the point to start that burn to end up in specific location?

Link to comment
Share on other sites

1 hour ago, 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.

What about manual calculation? I mean, if the burn time is known, then it should be simple math to determine the point to start that burn to end up in specific location?

Even without something to target that mod will provide more accurate burn times than stock, but if you want to calculate manually then using the ISP and Current mass of your lander and the DeltaV required you can rearrange the Rocket Formula (deltaV = 9.81ISPln(initial mass/final mass) to spit out the final mass, then use the fuel flow (seen in right click on engine while burning or in VAB) to calculate how long it will take to burn the difference (1u = 0.005t)

Edited by Rhomphaia
Link to comment
Share on other sites

1 hour ago, omelaw said:

right click on waypoint

P.S. I just bring a rover to there, because that kind of contract actually need to visit more the one places

If I right click on waypoint, it gives me a dialog to "activate navigation" and "delete waypoint". Not to select it as target...

 

1 hour ago, Rhomphaia said:

Even without something to target that mod will provide more accurate burn times than stock, but if you want to calculate manually then using the ISP and Current mass of your lander and the DeltaV required you can rearrange the Rocket Formula (deltaV = 9.81ISPln(initial mass/final mass) to spit out the final mass, then use the fuel flow (seen in right click on engine while burning or in VAB) to calculate how long it will take to burn the difference (1u = 0.005t)

I know how to calculate burn time, but could you please tell me how to calculate the distance the craft will travel until it finishes that burn?

Link to comment
Share on other sites

3 hours ago, 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.

What about manual calculation? I mean, if the burn time is known, then it should be simple math to determine the point to start that burn to end up in specific location?

The kind of contracts I always decline. Too much trouble,  too time consuming. 

As I said just accept you won't have total efficiency. If you need to land further and then hop closer so be  it,   just bring more fuel. 

Link to comment
Share on other sites

2 minutes ago, Spricigo said:

The kind of contracts I always decline. Too much trouble,  too time consuming. 

As I said just accept you won't have total efficiency. If you need to land further and then hop closer so be  it,   just bring more fuel. 

Come on, man, if you don't know how to calculate it, just say so. I'm trying to figure out a precise, efficient solution here, not "just bring more fuel".

Link to comment
Share on other sites

Are you familiar with calculus? Distance is just the integral of velocity. You are ending with a velocity of zero, so that makes it a little easier -- but this is in two dimensions, so that makes it a little harder (especially since your burn direction changes over time). Your initial estimate for the burn is 98 seconds, so in that amount of time you will gain about 9.8 * 50 / 6 m/s of downward velocity (80 m/s guesstimate calculated in my head) -- so you will need an additional 15 seconds or so of burn to counter that. So my initial estimate would be 570 m/s * 105 seconds / 2 = 30 km almost on the button. But you could help yourself out by doing the experiment once and getting a more accurate burn time. The problems are: you are interested in only the horizontal distance and velocity (and those numbers are annoying to separate out), the thrust/acceleration vector is not constant, and the munar gravity acceleration increases non-linearly from 0 to 1/6G (as your speed drops from orbital to 0).

Edited by bewing
Link to comment
Share on other sites

I finally worked it out. The formula is  BrakingDistance =  V /2 * BurnTime /1000

V is in orbital velocity in m/s;

BurnTime is the time it will take to burn down that velocity, in seconds;

/1000 is just to convert the braking distance into kilometers.

 

I tried it out just a few minutes ago, and it landed me within 700m of the target zone in a 102 second burn (I need to work on my timing to get closer). Hopefully this will be useful for someone too. :)

Edited by aluc24
Link to comment
Share on other sites

20 minutes ago, bewing said:

Are you familiar with calculus? Distance is just the integral of velocity. You are ending with a velocity of zero, so that makes it a little easier -- but this is in two dimensions, so that makes it a little harder (especially since your burn direction changes over time). Your initial estimate for the burn is 98 seconds, so in that amount of time you will gain about 9.8 * 50 / 6 m/s of downward velocity (80 m/s guesstimate calculated in my head) -- so you will need an additional 15 seconds or so of burn to counter that. So my initial estimate would be 570 m/s * 105 seconds / 2 = 30 km almost on the button. But you could help yourself out by doing the experiment once and getting a more accurate burn time. The problems are: you are interested in only the horizontal distance and velocity (and those numbers are annoying to separate out), the thrust/acceleration vector is not constant, and the munar gravity acceleration increases non-linearly from 0 to 1/6G (as your speed drops from orbital to 0).

True. I wonder if it's possible to work out a formula that would account for all these variables. I worked out a simplified formula (see my reply just above), which worked pretty fine, but I would be interested in more accurate solution as well. I am not familiar with calculus, though.

Link to comment
Share on other sites

23 minutes ago, aluc24 said:

Come on, man, if you don't know how to calculate it, just say so. I'm trying to figure out a precise, efficient solution here, not "just bring more fuel".

How to calculate is the easy part.  You need deltaV equal to orbital velocity to stop.  You can use a maneuver mode to find out how much time it takes. 

The difficult part it's to execute the burn,  it helps nothing the fact you are approaching with enough velocity to pass through you target in a blink, or the fact that maybe your target area its not exactly in your current trajectory because a hair of deviation in the other side of the orbit.

You may not like  it,  but you need to do a suboptimal approach to have time for those adjustments. And that is how people deal with those superefficient precision approaches, they trade efficiency for precision until they get a good balance. 

Link to comment
Share on other sites

Quote

You may not like  it,  but you need to do a suboptimal approach to have time for those adjustments. And that is how people deal with those superefficient precision approaches, they trade efficiency for precision until they get a good balance. 

This. As someone with an engineering background let me tell you, you want precision like that at closing velocities measured in an initial value of 100's of meters per second you need a computer to handle it from start to finish you cannot tiem it by hand period.

 

TLDR of why. If your burning to decelerate at a specific TWR with a mostly unchanging target elevation and with the same rate of change, (be that a fixed or a variable value), of angle of burn on decent for each attempt your landing point will be X meters in front of the point of ground under you when you start a burn. (X varying on all of the prior listed factors). What that means is at the moment you start the burn your projected landing point is moving across the terrain below at a rate equal to your orbital velocity. Even a slow orbit around the mun is a couple of hundred m/s. You cannot hit a 100m area when you need to time it to a few tenths of a second, human reaction times just aren't reliable enough for that. And thats assuming you can manually hold the desired profile. You can't, again no human is that acurratte. An error of 700m is simply phenomenal for what your attempting.

 

My advice, either accept a less efficient decent, or accept less accuracy and bring a rover, it's contracts like this that are why the rover exists.

Link to comment
Share on other sites

On the Design:

The lower the acceleration of a lander (Lower TWR), the harder it is to be precise on the landing, as all control inputs must be put in farther in advance, thus accidental delays have a much greater impact on the end result (possibly ending in impact)

Also, the lower the TWR, the more Delta V you will spend combatting gravity during your controlled decent (as suicide burns are much harder) while you are scrubbing off horizontal velocity, thus the less efficient the landing is.

As a rule of thumb for my landers, I design them such that it takes no longer than 100 seconds of burn to arrest orbital velocity.  on the Mun, that equates to a .5 Kerbin TWR (around 3.0 Mun TWR).  I've not done a lot of spreadsheet work to confirm that this is the optimal TWR, but it is the one that feels the right balance between efficiency and controllability during decent.  

Link to comment
Share on other sites

12 hours ago, Spricigo said:

How to calculate is the easy part.  You need deltaV equal to orbital velocity to stop.  You can use a maneuver mode to find out how much time it takes. 

The difficult part it's to execute the burn,  it helps nothing the fact you are approaching with enough velocity to pass through you target in a blink, or the fact that maybe your target area its not exactly in your current trajectory because a hair of deviation in the other side of the orbit.

You may not like  it,  but you need to do a suboptimal approach to have time for those adjustments. And that is how people deal with those superefficient precision approaches, they trade efficiency for precision until they get a good balance. 

 

7 hours ago, Carl said:

This. As someone with an engineering background let me tell you, you want precision like that at closing velocities measured in an initial value of 100's of meters per second you need a computer to handle it from start to finish you cannot tiem it by hand period.

 

TLDR of why. If your burning to decelerate at a specific TWR with a mostly unchanging target elevation and with the same rate of change, (be that a fixed or a variable value), of angle of burn on decent for each attempt your landing point will be X meters in front of the point of ground under you when you start a burn. (X varying on all of the prior listed factors). What that means is at the moment you start the burn your projected landing point is moving across the terrain below at a rate equal to your orbital velocity. Even a slow orbit around the mun is a couple of hundred m/s. You cannot hit a 100m area when you need to time it to a few tenths of a second, human reaction times just aren't reliable enough for that. And thats assuming you can manually hold the desired profile. You can't, again no human is that acurratte. An error of 700m is simply phenomenal for what your attempting.

 

My advice, either accept a less efficient decent, or accept less accuracy and bring a rover, it's contracts like this that are why the rover exists.

 

1 hour ago, Birdco_Space said:

On the Design:

The lower the acceleration of a lander (Lower TWR), the harder it is to be precise on the landing, as all control inputs must be put in farther in advance, thus accidental delays have a much greater impact on the end result (possibly ending in impact)

Also, the lower the TWR, the more Delta V you will spend combatting gravity during your controlled decent (as suicide burns are much harder) while you are scrubbing off horizontal velocity, thus the less efficient the landing is.

As a rule of thumb for my landers, I design them such that it takes no longer than 100 seconds of burn to arrest orbital velocity.  on the Mun, that equates to a .5 Kerbin TWR (around 3.0 Mun TWR).  I've not done a lot of spreadsheet work to confirm that this is the optimal TWR, but it is the one that feels the right balance between efficiency and controllability during decent.  

 

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.

Link to comment
Share on other sites

This thread is quite old. Please consider starting a new thread rather than reviving this one.

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...