Ideal Commnet Altitude - What do you use?

[MaximumThrust]

Hi!

I like to set commnet relays to archive total coverage of the bodies surfaces. More a hobby/completionist thing than a real need.

Searching about it, i found that the bare minimum is the radius of the planet/moon. I'm trying to use two sets of 3, the first set at 45º, and the other at -45º. Or both 45º with LAN 0º and 180º. Forming a "X", each satellite apart 120º, with one RA-2.

Using Kerbin as an example, it would be 600 km. I tested it and it didn't felt right. The lines between the satellites goes too close to the planet, and vessels on the surface need angles too step into the horizon to have communication. Any increase in occlusion and it stop working.

I tested 720 km ( x 1,20 the radius), it became better, but with 6 satellites there are blackspots. They can be in the poles, or in the tropics, depending how I release the 2 sets. Making some tests, I found that 5x the radius is what looks better and somewhat more realistic, the angles in the horizon are far better for the vessels in the surface. For Kerbin this is 3000 km. It felt much more right, and I'll probably keep it for Kerbin and all the bodies. With a simple Communotron 16 I can get 100% signal anyway.

To you guys think this altitude is simple a waste of resources, or it is really more realistic? What altitude to you use?

[The Aziz]

Try the geostationary altitude maybe.

[Spricigo]

my rule of thumbs: 1.5*radius of planet

 

[Streetwind]

Your answer depends not only on the planet's size, but also on its terrain. For example, on a perfectly even sphere, a trio of satellites passing directly overhead ensures that you can always see at least one of them - regardless of altitude. Meanwhile, if you have some steep mountains and valleys, even six satellites in the same orbit passing directly overhead may not be enough unless the altitude is really high (for example, geostationary). That's why designing for 100% coverage is a fairly tedious and resource-intensive task; try limiting yourself to 99.9%.

The really interesting points are those furthest away from the orbital plane of your satellites. For a low orbit constellation consisting of sets of polar satellites, they will be along the equator in between where the satellites pass, and these spots may shift over time if the satellites' orbits are not perfectly in resonance with the planet's rotation (see: sun-synchronous orbit). For equatorial satellites, they will be on the poles. Conveniently Kerbin's poles are flat as a billard ball, and so a trio of equatorial satellites will not have any dead zones on Kerbin's poles above a certain altitude. On other celestial bodies, like the Mun, this is an entirely different story; no amount of equatorial satellites will manage to cover the Mun's poles adequately, regardless of altitude.

 

Edited February 8, 2018 by Streetwind

[RizzoTheRat]

Bearing in mind I use RemoteTech so might be slightly different to standard...

For Kerbin I use a 4 satellite network at about 780km, as this rather conveniently gives me a 1.5 hour orbital period which is nice and easy to remember when setting them up.  This gives a decent clearance above the horizon to allow for a reasonable amount of drift, and 17 years in to the game I'm now thinking about needing to move them a bit, but Iv'e still got full coverage around thee equator. 

This has some useful info.

https://wiki.kerbalspaceprogram.com/wiki/Tutorial:Satellite_Coverage

[Spricigo]

1 hour ago, Streetwind said:

That's why designing for 100% coverage is a fairly tedious and resource-intensive task; try limiting yourself to when needed.

 

Is perfectly viable to start with a basic network 'full of gaps' if you can get signal often enough and long enough to do the necessary manoeuvrers/transmissions for your mission. As you needs for a better coverage increase, new satellites can be placed either to extend the network, replace the older satellites or just for redundancy.

[Aegolius13]

It seems like the most relevant factors are:

First you want to avoid gaps in coverage.  A higher orbit is better for this due to less occlusion from Kerbin.  This also depends on how many relays you use.  E.g., 4 evenly spaced relays will do better than 3.

Second, you want enough signal strength for consistent control.  A lower orbit is better for this.  Of course this also depends on which antennas you use on both your relays and your probes.  I try to set up my network so that even if I forget an antenna on my probe (which never, ever happens, believe me), I can still get control through the built-in 5k antennas on probe cores.  According to the wiki entry on Commnet, with a 2G relay antenna and a 5k built-in, you can get signal about 3 million meters, or 25% of the way to the Mun.  

Third, other factor is that the higher the orbit, the less degradation in your alignment you'll probably see.  (Longer orbits = fewer orbits over time, and thus less relative movement among the relays due to minor differences in orbital period). 

So putting this all together, I tend to go with one of two options.  One is to set up a network of four 2gm relays in low (but not too low) Kerbin orbit - say 1 million meters or so.  The other option is to use a set of 100gm relays way out past the Mun, which can do double duty as deep space relays.  

[MaximumThrust]

Thanks everyone for the comments. In this gameplay I'll use 4x the radius, a little less than stationary (for Kerbin). I know this isn't the best option in game mechanics, but is whats looks better for me. IRL constellations for air-ground communication usually orbits relatively high. From the Wikipedia:

"Because of their low altitude, these satellites are only visible from within a radius of roughly 1,000 kilometres (620 mi) from the sub-satellite point. In addition, satellites in low earth orbit change their position relative to the ground position quickly. So even for local applications, a large number of satellites are needed if the mission requires uninterrupted connectivity."

I also set a powerful stationary relay above the KSC, so I can have better signal in my vessels launches and landings without adding antennas, and minimize the signal losses with unmanned (unkerbed?) spaceplanes. I really enjoy this, latter I'll mess with the range settings in difficult to make this networks more useful. Right now I'm using 1.1 occlusion, the maximum the interface allows.

Some pics: