Remote Tech - How to get 100% synchronous orbits?

[Cairol]

Today I started with Remote Tech.

I'm currently setting up my Kerbin orbital com sat network.

I'm using three sats for full coverage in a geostationary orbit (2,868.7x km)

Problem is: I can't get 100% perfect orbits.

Even when using RCS in minuscule bursts for fine tuning, I have a margin of about +/- 50 m per sat.

That means, my three sats will eventually desync and move towards or away from another noticably in a matter of a few Kerbin days / weeks.

I don't want to rearrange all my sats every couple of in-game hours (or even minutes if high acceleration is used during missions)

Especially when I set up further sats over other planets ect.

How do you get a "perfect" orbit so you would not have to readjust all sats constantly?

How do you maintain your sat networks?

(Hope this doesn't belong to "Support (modded installs)" as it's not a bug or tech support request...)

Edited July 3, 2015 by Cairol

[LABHOUSE]

Higher orbit, lower twr will get it closer

[SoWeMeetAgain]

Today I started with Remote Tech.

I'm currently setting up my Kerbin orbital com sat network.

I'm using three sats for full coverage in a geostationary orbit (2,868.7x km)

Problem is: I can't get 100% perfect orbits.

Even when using RCS in minuscule bursts for fine tuning, I have a margin of about +/- 50 m per sat.

That means, my three sats will eventually desync and move towards or away from another noticably in a matter of a few Kerbin days / weeks.

I don't want to rearrange all my sats every couple of in-game hours (or even minutes if high acceleration is used during missions)

Especially when I set up further sats over other planets ect.

How do you get a "perfect" orbit so you would not have to readjust all sats constantly?

How do you maintain your sat networks?

(Hope this doesn't belong to "Support (modded installs)" as it's not a bug or tech support request...)

This will not be a perfect solution either, and perhaps you're already doing that anyways.

If your burns directly into the direction you want to go are always "too much" you can instead split it up into two bursts at steep angles.

So instead of burning prograde you burn mostly "upwards" and just a little bit forwards, a steep angle upwards (say 80°), then you burn a second time with the same angle from before, downwards (so 160° downwards from your first burn). The vertical parts of the burns should cancel out and leave you with the two horizontal parts which you can make smaller than one direct horizontal burst depending on your angle. Or you make their sum slightly bigger and do a third burn in the other horizontal direction.

Can one limit thrust on RCS ports?

[TheCanadianVendingMachine]

Use thrust limiter on engines. If you have a limit of 1%, that means at max thrust it'll only give 1%

This also means that you can have 0.1% thrust. This will help a lot

[Streetwind]

I used Kerbal Engineer Redux to get the orbital period readout. I adjusted that one to match as close as possible. It doesn't matter if one satellite's apoapsis or periapsis is slightly different, the thing that really matters is that they all take the same amount of time to go around Kerbin. That way, they may oscillate a bit with respect to each other, but overall the network will remain stable.

[Eric S]

I can usually get them stable for years, but even then I want better.

You can use Hyperedit or save file edits to park the satellites perfectly, but if you do that, you need to be careful never to make them your active ship, as the pass from on rails to physics and back to rails will throw the orbit off due to FP rounding errors.

I find the best way to fine tune the satellite's orbit is to use a very weak engine (single LV-1 ANT or Ion engine), thrust limited as low as you can get it and then burned on minimum thrust. Oh, and you want the satellite to be very stable, no flexing, as the flexing will throw off the orbit.

[Starhawk]

The best way as I understand it is to use time rather than distance. You need KER or some other mod which will display the orbital period.

Orbital period matters more than distance, and is easier to get very accurate. Using engines thrust limited to the minimum or RCS, try to get the orbital period as perfect as possible.

As I recall from when I tried RT, it's possible to get it close enough that you don't have to tend them for several years.

The other option is to edit the config file once you set up the orbits to make them perfect. I'm sure there's info about how to do this around. Probably in the RT thread.

Happy landings!

[Alshain]

It is impossible to get 100% accuracy but you can get close enough that it will take decades to fall out of sync. At synchronous orbit altitudes you have a lot of wiggle room around the planet anyway. You want to use orbital period (the amount of time it takes to orbit the planet). Actual orbital altitude is irrelevant in this case because there are several ways to get the same orbital period.

[Cairol]

I'll try it with a weak engine with low thrust and the "don't burn directly prograde/retrograde" seems also worth trying.

Sadly RCS thrusters cannot be limited in thrust. I thought they should be weaker than any engine burst.

But using a limited engine may actually work more precise.

Thanks for your helpful answers!

[Requia]

Some additional options: Instead of matching orbits, match the total of AP and PE, at long as they sum together the orbital periods will be exactly the same, no matter how weird and elliptical they are.

For limited thrust, the flight computer can do .01s bursts, faster than you can hit x after hitting shift. Between the two of these I've gone years without adjustments, perfection is impossible though cause of the AP/PE wobble when not on rails.

By using highly elliptical orbits, you can greatly resist drift, even once badly out of sync the sats will still give full coverage whenever in the upper part of their orbit, which is most of the time.

[windows_x_seven]

100%? Nope. Not possible. Even real-life satellites have to adjust their orbits once in a while.

If you don't want to go through that, however, use HyperEdit. Cheat your way to space.

[Requia]

even hyperedit isn't 100% cause of the way the game calculates orbits, AP and PE change slightly every physics tick.

[windows_x_seven]

even hyperedit isn't 100% cause of the way the game calculates orbits, AP and PE change slightly every physics tick.

I've used HyperEdit to set up my satellite constellation in geostationary orbits, and I didn't need to correct them. Ever. And it's been a long time since I've set the satellites up.

[Kryxal]

As long as they're off stationary by the SAME amount, it's not going to be a big deal, is it?

[blnk2007]

The best thing I can say to do is use RCS to fine tune orbits. Watch YouTube and build a four satellite deploying ship. Then, use Mechjeb to maneuver plan resonant orbits each time you need to drop another satellite off. Again, use Mechjeb to match velocities with the last satellite you dropped off. After you have dropped off all four satelites use RCS to match their orbital periods down to the seconds (This is much easier if you have used RCSBuildAid to balance you satellites . It is possible without hyperedit, plus it's more fun. 100%? No. Very close? Yes.

[Reddeyfish]

As long as they're off stationary by the SAME amount, it's not going to be a big deal, is it?

pretty much, as long as the orbit is more-or-less circular (or equally elliptic with all the other sats). If one sat is extremely elliptic while all the others are circular, there will be problems.

[LordCurlyton]

Or just launch lots of satellites to ensure that there will always be somebody able to talk to somebody else around the whole coverage area? Its what I'd do to compensate for human error. Enough redundancy should ensure that basically it would be well beyond any reasonable time before coverage is <100%. I know, its not the "skill" method, but comm satellites are cheap relatively speaking. Admittedly, Remote Tech is on my "to do" list, not my "have done" one.

[Cairol]

I really miss a "Thank you" button in this forum. ^^

As I saw another problem, I'll go with 4 orbital satellites instead of 3.

I can't place the 3 sats exactly evenly in a 120 degree angle from another just by eyeballing it.

-> Again: How do you usually do this?

So my Communotron 32 omni antenna on sat B with its 5,000 km range is currently a bit outside of its range

which means my connection is not stable.

A 2,869 km orbit means a direct distance of 4,969 km if all sats are optimally placed.

But currently, B is about 5,900 km away from A.

With 4 sats it should be OK with that 950 km margin it gives with my antennas.

[RCgothic]

The star shaped button to the right of 'Blog This Post' under a user's avatar allows you to thank them by boosting their reputation.

I use KER and some spreadsheet calcs to set up my satellite constellations. EG, if I'm aiming for an orbit with a period of 1h30m and 3 sats, I'll inject my launcher into a 1h20m elliptical orbit with the correct apoapsis. It's then a case of dropping off a sat at apoapsis every third orbit and burning with the sat's on-board RCS until the orbital period is correct. My most recent constellations are accurate to milliseconds.

Edited July 4, 2015 by RCgothic

[blnk2007]

I really miss a "Thank you" button in this forum. ^^

As I saw another problem, I'll go with 4 orbital satellites instead of 3.

I can't place the 3 sats exactly evenly in a 120 degree angle from another just by eyeballing it.

-> Again: How do you usually do this?

So my Communotron 32 omni antenna on sat B with its 5,000 km range is currently a bit outside of its range

which means my connection is not stable.

A 2,869 km orbit means a direct distance of 4,969 km if all sats are optimally placed.

But currently, B is about 5,900 km away from A.

With 4 sats it should be OK with that 950 km margin it gives with my antennas.

http://ryohpops.github.io/kspRemoteTechPlanner/

That website will help you a ton. It will do the math for you as far as seeing what will actually be covered by your antennas. I suggest not pushing your satellites apart to the antenna's max range. You'll need a little wiggle room.

Also, check out resonant orbit calculations and Scott Manley's orbital math.

[ElWanderer]

I'll try it with a weak engine with low thrust and the "don't burn directly prograde/retrograde" seems also worth trying.

Sadly RCS thrusters cannot be limited in thrust. I thought they should be weaker than any engine burst.

But using a limited engine may actually work more precise.

RCS thrusters output a lot less when in fine control mode (I.e. hit caps lock). If you have four 4-way blocks you can right-click disable two of them, assuming you only need translate forward/backwards. Mounting an ant or two spiders and limiting the thrust to 5% (I think that's the minimum) may allow for even finer control, though I've not tried it.

[match]

As I saw another problem, I'll go with 4 orbital satellites instead of 3.

I can't place the 3 sats exactly evenly in a 120 degree angle from another just by eyeballing it.

-> Again: How do you usually do this?

So my Communotron 32 omni antenna on sat B with its 5,000 km range is currently a bit outside of its range

which means my connection is not stable.

A 2,869 km orbit means a direct distance of 4,969 km if all sats are optimally placed.

But currently, B is about 5,900 km away from A.

With 4 sats it should be OK with that 950 km margin it gives with my antennas.

Using a bit of math, you can calculate angle by targeting the other sats and finding the distances between them. You can solve the remainder of the triangle with altitude measurements of the sats. (Kerbal has a 600 km radius.)

Quick and dirty method is to equalize the distance between all three sats, given they are at the same altitude (and inclination!).

With RT, setting (and respecting) the tolerances for distances in your constellation is a priority. Like you have discovered with your Communitron 32, going over just a bit will have bad consequences. If you are approaching a distance limitation with your deployment, best to rethink it and consider what you are trying to accomplish.

As far as getting your birds to sit pretty exactly where they should, well that is a matter of theory vs application... 2868.75 km orbit may seem like perfection, but in reality 100% coverage is the goal, and there are far more ways of accomplishing that than a synchronous orbit. (If using a synchronous orbit to relay to Mission Control, why need a second one also in synchronous orbit? If their primary purpose is redundancy, then remove the need for "synchronous" on your other satellites.)

[Cairol]

Yup. I'm just watching this:

Very helpful videos.

My MK2 satellites now have a small engine that works perfectly for fine adjustments.

Didn't know that RCS thrusters in "precision mode" will output less. That may make me omit the engine again in MK3,

because I actually don't like to have a fully fledged rocket engine on my sats.

With your help I was able to set up 4 of my "old" MK1 COM sats and sync their orbital period with a tolerance of +/- 200 ms.

By using the RCS thrusters in "precision mode" I'll try to tweak this further.