[Tutorial] Geostationary Orbit Constellations for Dummies [MechJeb]

[phoenix_ca]

The title says it all. I would put this wee guide on the wiki, but the wiki is rather...restrictive. I also couldn't find any simple guides like this for KSP anywhere after a lot of searching, so here it goes.

First off: No complex math. You don't even need to know how to multiply or divide. The only skill you need is basic addition and subtraction, and an ability to read the data MechJeb gives you (yes, you can fly the thing yourself the entire way, though I can't imagine why anyone would want to do this eight-or-so times without using at least Smart A.S.S., which is what I use for all my orbital maneuvers for this after the ascent. It will get repetitive the more satellites you put up).

Second: I say geostationary because I'm comfortable with it, I'm not interested in arguing semantics, and geo is from the Greek for "earth" though typically in the sense of "land" or "ground", not specifically this planet we meatbags call home.

Third: I assume here that the reader has a firm grasp of orbital maneuvers, including phase changes (adjusting the altitude of an orbit), plane/inclination changes, basic ship piloting, and perhaps even a cursory knowledge of what all the lovely buttons do in MechJeb. There are tutorials already for all of this, so I'm not going to cover any of it.

Finally: The goal of this tutorial is not just to describe how to get a payload into geostationary orbit, but how to get it at more-or-less a precise longitude with a minimum of fuss. That is, one burn from a parking orbit to GEO, or close to GEO, with minimal adjustments if the user wishes.

Step 1: Understand What a Geostationary Orbit Is

...and the difference between "geostationary" and "geosynchronous". Really, I can't do a better job here than Wikipedia, and for the sake of brevity, I won't go into detail. Short version: a geostationary orbit will keep some orbital body above the same longitude at all times. Put more simply, to an observer on Kerbin, a satellite in geostationary orbit will remain stationary in the sky (well, hopefully, or at least with so little deviation it's unnoticeable). A practical example: Geostationary satellites are used to provide streaming data to large numbers of clients, such as in the case of satellite TV providers. The simplicity of using a simple receiver that doesn't need to track the satellite far out-weighs the cost of putting a satellite in GEO.

Now, Kerbin's GEO altitude is 2.870Mm. Get a piece of space junk up at that height, such that its apoapsis and periapsis are both at 2.870Mm, and you have yourself a GEO.

Step 2: Build a Rocket

Anything that can get your desired payload to an altitude of 2.870Mm above Kerbin will do (that's Mega-meters, or 2.870 million meters). Test it out with your payload, see if it can get you there in one shot, with fuel and an engine to spare, preferably, for fine-tuning; I have a soft spot for the NERVA for this task. An easy way to do this is to just set MechJeb's ascent autopilot to 2,870km and let it go. If you get your payload there with fuel to spare, you're in good shape and have enough delta-v to get your payload there. Note that any rocket that has enough delta-v to escape Kerbin (e.g. Mun or Minmus launch vehicles) will likely have enough delta-v to get to a stable GEO (the circulisation burn could mess with this; I haven't done any specific math on this point, nor do I plan to).

Step 3: Launching

I've seen more than a few mentions of GEO constellations requiring precise launch timing, as one would use for a rendezvous with a spacecraft in low-orbit. This simply isn't the case. If we consider the case of a direct ascent to GEO, and a similar ascent profile for each launch, each satellite will end-up in roughly the same position because, relative to the rotating surface of, the final point of the ascent hasn't moved. So, the bottom line? Launch whenever you darn well please, as we need to fix this later anyway.

This first launch is a dry run. We're out to collect data, and little more, you may even wish to do this in a test save if you're as OCD as I am about not littering space more than necessary. Launch your vehicle to a low-altitude of your choosing, just write it down. This is where all subsequent launches will be going. I usually go with 200km (eventhough this has the mild downside of being a little slower because of time compression limits...I don't really mind). Once you are at your parking orbit (you did remember to write that altitude down, right?) and inclination is at 0 (or within a few degrees), open MechJeb's surface information window. At the very bottom you will see Longitude. This the the key piece of data we'll be using.

Step 4: Transfer

Wait until your craft is approaching 0 degrees longitude. Then commence a burn to enter a Hohmann transfer orbit with apoapsis at 2.870Mm. Warp to the AP, then bring your periapsis up to 2.870Mm as well. If you are using a relatively low-thrust engine (or even a high thrust one; you'll need to expend quite a bit of delta-v to circularise this orbit), you may end-up burning for so long that your apoapsis begins to rise again. That's okay, get it close, we've got our data.

Note the current longitude. We'll call this x.

And for the sake of actually giving an example, let's say we arrived at a new longitude of 138 degrees. Ta-da! That 138 degrees is very, very important. I hope there are lightbulbs going off now...

Step 5: Interpreting What the Heck I Just Told You To Do

Here's where the math comes in. The beautifully easy, peasy math. Let's say that we now want to deploy a payload such that when it's in GEO, it's (approximately; done right this method should get one within a few degrees longitude) at 0 degrees longitude. We simply do: 0 - x = longitude of burn. In this case, we get -138 degrees. So, following the same pattern as our test flight, exiting our parking orbit of 200km at -138 degrees longitude (or at least, starting our burn there) will have the satellite arrive in GEO on or close to a longitude of 0 degrees.

To further illustrate, here's a table showing the (exceedingly simple in comparison to most math stuff with orbital mechanics in KSP) math that gives us the longitudes to burn at for all eight orthagonal points, giving a nice, evenly spaced set of 8 satellites (negative values used so they correspond correctly to longitude as displayed by MechJeb):

 

 

0 - 138 = -138
45 - 138 = -93
90 - 138 = -48
135 - 138 = -3
180 - 138 = 42
-135 + 138 = 3
-90 + 138 = 48
-45 + 138 = 93

 

And unless I completely messed-up (I'm a little sleep-deprived of this writing) that should make sense. See? I keep my promises. Not so much as a multiplication sign. But wait! There's more!

Step 6: So You Don't Know How to Add

No problem! All you need to know is how many degrees you want to separate your satellites by. Then just initiate your burn from points that are equally spaced by longitude. So if you don't really care where exactly your satellites end-up above Kerbin, you could simply do transfer burns at 0, 45, 90, 135, 180, -135, -90, and -45 degrees longitude. Just make sure you do the same thing with the same rocket every time. Changing the throttle, thrust (i.e. engines), mass, anything, will throw everything off, requiring you to go back to Step 3 for that rocket.

When you're finished, you're free to correct the orbit as you normally would, by reducing periapsis, increasing apoapsis, and returning, burning normal or antinormal to adjust inclination, etc. I've managed to get within a few degrees longitude of my intended destination (about a degree or two off; close enough for me), with eccentricity of 0.0003 and inclination of <0.1 (which is the standard margin of error for inclination of real-world GEO satellites).

I hope that made some modicum of sense, and that it's somewhat helpful. I've found this to be an exceedingly easy way to have a nicely-spaced constellation of comsats around Kerbin. You know, for Kerbin TV...or whatever they do to rot their brains.

Edited June 1, 2016 by phoenix_ca
Fixing typos

[djnekkid]

Not too bad, but here is my 2cent :-)

If you want a geostatenary net with say 8 sattelites, with mechjeb assisted launch, (com releay mod anyone??) Then i find it easier to build a simple rocket that can get you into a LKO, any altimtide is ok, but like the OP, say 200km. Let mechjeb launch it for you. Name it i.e. geostatcom1. When in orbit go again to VAB, and rename it to geostatcom2. Then notie the LPA in mehjebs acent module, roundevouz with the prev sat, but add or substract 45 to that number (if you want a grid of 8)

Redo this until you have 8 sats in a nice LKO.

Now go to number1, start a timer (cellphone or otherwise) and set it to 30sec or 1min, depending on rocketpower and how stressfull you want your next 15 mins to be. Raise the AP of the first sat to 2864,4Mm. At a good timing raise the AP, hence the stopwatch or timer.

When all 8 hve their AP raised, go to nr1, and warp until AP-30sec, wait until AP is less then 1sec away, and circulize the orbit. Asumeing that it takes less then the 1minute, now go to the next one and the AP should be just. Few seconds away. Rince and repeat, and voila, theree are your geosat net in 8simple launches.

IIR did i do this with a KW 1x4m tank with an aerospike underneth, with a remotesat, antenna, mechjeb computer, the smll 165liter blue and white novapunch tank, and that 60n thrust novapunch engine.

My next set of missions are to do the same with a spaceplane. That will not be an easy task. And the ones that will stay at about sunrise and sundown will have the interplanetary antennas as well :-)

[phoenix_ca]

Hrm, but is there a way to do that easily so that your satellites end-up near a per-determined longitude? That's the main "benefit" of the method above. Granted, it has absolutely no real use other than saying "I can do that"...or if you have an active imagination and like to think you just hooked-up the TV and satphones for the Kerbals.

[WoodenBiplane]

I love this!

[Tinman]

Now to figure out the altitude to do a Geostationary orbit on all of the other planets

[djnekkid]

Now to figure out the altitude to do a Geostationary orbit on all of the other planets

I _think_ most of the gratival parameters are known, as well as rotational speed. Then use a (online) calculator to figure it out. I did it with Duna, and quite magically is it the same as of Kerbin

[KleptoKat]

Great work. Really simplify's it.

Could you include a chart somewhere for the rest of the planets geosynchronous orbits? I would love you, as if I already don't.

[phoenix_ca]

I'm using my phone right now, so can't really tell you what the altitude for each other celestial body is. I can tell you that the original procedure is still valid for any body with a stable geosynchronous orbit. ISA MapSat's in-game Kerbalpedia has this information readily available.

[Azivegu]

i am currently working on a geostationary explanation for the wiki. The last few days i have been adding/recalculating the geo orbits and adding special sections into the pages on orbits. Ill try to incorporate as much of your work as i can (and i will of course give you credit )

[maltesh]

I _think_ most of the gratival parameters are known, as well as rotational speed. Then use a (online) calculator to figure it out. I did it with Duna, and quite magically is it the same as of Kerbin

It's a bit higher above the surface of Duna than Kerbin's is (2,880 km vs Kerbin's 2,868.7) on the other hand, Duna's a smaller planet, so the actual radius of the stationary orbit over Duna is smaller.

However, with Duna, there is the minor matter of a 260-km-wide rock (with its 2000 km wide sphere of influence) occupying a synchronous orbit, sneaking up on any synchronous satellite that dares to venture out of the exact 1:1 resonance.

[Mykill Metal]

Step 4: Transfer

Wait until your craft is approaching 0 degrees longitude. Then commence a burn to enter a Hohmann transfer orbit with apoapsis at 2.780Mm. Warp to the AP, then bring your periapsis up to 2.870Mm as well. If you are using a relatively low-thrust engine (or even a high thrust one; you'll need to expend quite a bit of delta-v to circularise this orbit), you may end-up burning for so long that your apoapsis begins to rise again. That's okay, get it close, we've got our data.

you should edit the underlined part...a bit confusing as i'm pretty sure they should be the same number...then again, could be wrong

[Mighty1]

Here is what I do.

Launch a Sat deployer (crewd if using RT, probe if not using RT) with a coresponding number of sattelites. Get into a stable orbit. Bring the Ap to Geostat level (2868,75 km). Now, we know that 1 rotation of Kerbin takes 6 hours but you could use this information for any planet depending on its rotation speed. For Kerbin I usualy use 4 sattelites as a comm relay. So at Ap i bring my orbital period to 4 hours (can t make it 2 because I would go down into the atmosphere) - MeckJeb or KE help a lot - and then decouple 1 sattelite at the Ap and circualise it to a 6 hour orbit (RCS helps for fine tuning). After 4 orbits of your Sat deployer you have 4 evenly spaced sattelites all in 6 hour orbits.

Your relay network is complete.

Simple, and low math usage

[SmarterThanMe]

I do pretty much exactly what Mighty1 does, except that I get the Ap up to 2868.75, then I decrease the orbital period to 7/8 x 6 hours (5 hours 15 minutes), then let it go through an orbit, use the Sat Deployer to circularise up to geostationary, and deploy one of the sattelites. I then use RCS on the sattelite to tweak the orbit so that it's exactly 6 hours and that gives Sat 1 at geostationary. I then switch back to the sat deployer, and repeat the process (change orbital period to 5h15m, orbit, circularise, deploy, tweak, wash, rinse, repeat).

It gives an almost perfect network of geostationary orbits. In fact, it can work to give yourself a network of sattelites in any orbit approximately equal spaced. If you want more or less sattelites in your network, then change the denominator to whatever that number is, and the enumerator is that number + or - 1.

[linkxsc]

The 2 guys above me have it spot on. Its not only more efficient to do 1 launch for 4-6 satellites. But the satellites can be smaller and less costly for when the game starts tracking costs. Many of the rt comsats/kethane scanners i build, end up getting fielded like this. Also it saves on total launches if your playing with limiting yourself (say 1 launchpad and 3 runway launches per week)

Also it can be much easier tuning the orbits because you can kill any polar component of all 6 sats orbits while they are still attached. When the time comes you just need to launch them off and burn prograde

Edited May 5, 2014 by linkxsc
more to say

[draeath]

Really key point to this that simplifies things a lot...

Your inclination and eccentricity are nowhere near as important as your orbital period. If you only care about one number while trying to put stuff into synchronous orbits, that is the number to care about.

[MerlinsMaster]

Okay, I've used this tutorial before, and it worked well. But today I'm trying to put three satellites into equally spaced in a 600km orbit. I started in a 200km orbit, and raised the orbit of the first satellite to 600km at 0 degrees longitude. I raised the second satellite at 120E longitude, and the third at 120W longitude. But they are not equally spaced at all. In fact, two of them are very close to each other. Is there some important step that I'm missing? I've tried this a couple of times with the same result.

[maltesh]

Okay, I've used this tutorial before, and it worked well. But today I'm trying to put three satellites into equally spaced in a 600km orbit. I started in a 200km orbit, and raised the orbit of the first satellite to 600km at 0 degrees longitude. I raised the second satellite at 120E longitude, and the third at 120W longitude. But they are not equally spaced at all. In fact, two of them are very close to each other. Is there some important step that I'm missing? I've tried this a couple of times with the same result.

The instructions in the original post /only/ work if your destination is a stationary Kerbin orbit. Given that the destination orbit is intended to have an orbital period equivalent to that of Kerbin's sidereal rotation period, the rotation of Kerbin effectively cancels out when going to circular orbit of altitude 2868 km. They won't work for a 600 km altitude orbit.

[MerlinsMaster]

The instructions in the original post /only/ work if your destination is a stationary Kerbin orbit. Given that the destination orbit is intended to have an orbital period equivalent to that of Kerbin's sidereal rotation period, the rotation of Kerbin effectively cancels out when going to circular orbit of altitude 2868 km. They won't work for a 600 km altitude orbit.

Yeah. That does make sense. Thanks for pointing that out.

So are there any ways that you know of for positioning satellites at specific orbital positions on orbits that are different altitudes? I see lots of people with perfect-looking satellite constellations, and I'd love to know how to do that.