Peppie23

[1.1] RemoteTech v1.6.10 [2016-04-12]

Recommended Posts

Yeah I feel the same about RemoteTech. I like the concept and have successfully set up some communication networks. It feels really great the first few times when you do that and watch the network nice and tidy. But the problem is the satellites drift away way too fast even if you have set up a very closely synced network. The game engine itselfs limits how fine tuned the orbit can be and it gets quite tedious and boring to have to manually correct it every few hundred days.

I hope there could be a feature like ground satellite control that the system automatically synchronize and adjust the orbits of satellites, provided you have launched them properly within a reasonable margin of error(eg. 5 seconds). That would relieve a lot of tedium of maintaining the networks.

You have my 2 cents.

I'd be happy enough if ground control would be able to tell me "check constellation on planet x because it's about to break", then I'd load sat 1, and one click flight computer so it'd a) oriented pro/retro B) fire until it'd drift back into the proper place. Go to sat 2 etc

Now I have an alarm every 50 days for kerbin, if only I could do maintenance every 500 days.... I'd be 10x less bored.

Share this post


Link to post
Share on other sites
Now I have an alarm every 50 days for kerbin, if only I could do maintenance every 500 days.... I'd be 10x less bored.

Jesus Christ that's horrifying. Here I am trying to suggest solutions for a problem I run into about once every two game years, and you live like that?

Damn, let's meet up or smth, you need a hug ASAP.

Share this post


Link to post
Share on other sites

Every 50 days? Wow. My sats never go out of sync that fast. I can go decades. One tip for fine tuning, use the thrust limiter to get the orbital period as close as possible. Once you get in the ballpark, limit the thrust to as low as you can get it and adjust some more. What ever you do, don't use overpowered engines, that will make it harder. The Ant engine is perfect for Oscar B based sattelites, the 48-7S is ok but still a little too strong.

Share this post


Link to post
Share on other sites

I personally have no need for any sort of SMA alignment feature, but I understand why some people would like that.

It's up to the devs to decide whether such a feature is a worthwhile expenditure of their time to implement. Anyone not a current dev for RT2 who wants this feature is welcome to try and code it themselves and submit a pull request

Share this post


Link to post
Share on other sites

While it's not strictly RemoteTech related, I am wondering if the folks having trouble building stable constellations are using PreciseNode.

Share this post


Link to post
Share on other sites
Sounds like I'm not busy in real life?

This kind of thing is encouraging you to find a better orbit and a better way of fine tune your orbit. My 4-satellite equatorial Kerbin network still preserves a pretty decent square shape even after 10 years of game time without adjusting at all (and no save file editing - only adjusted by their own engines upon initial setup). How much maintaining would that cost? I think you find the checking/adjusting costly because you haven't found a better way of doing it.

Optimizing orbit is part of fun that RT brings me.

Better way? Other than manually fine tune the orbit period? Yes it's totally possible to achieve rather stable network but it's just very grindy and unnecessarily time consuming. It's okay to do that with first 1 or 2 networks. But once you do the same thing for the 20th time, it becomes very tedious. Optimizing orbit is fun the first few times, after that, it becomes a chore and repetitive. To each their own though. Some people like grinding in MMOs and I absolutely hate it.

- - - Updated - - -

Every 50 days? Wow. My sats never go out of sync that fast. I can go decades. One tip for fine tuning, use the thrust limiter to get the orbital period as close as possible. Once you get in the ballpark, limit the thrust to as low as you can get it and adjust some more. What ever you do, don't use overpowered engines, that will make it harder. The Ant engine is perfect for Oscar B based sattelites, the 48-7S is ok but still a little too strong.

Yep, or better just use RCS, limit thrust, capslock fine tune.

But even with that, it's quite tedious to switch back and forth the satellites to have the closest orbit period possible. I did that with the first several network I set up and then felt I was just repeating the chore with no real progression or new discover. It stopped being fun.

Share this post


Link to post
Share on other sites
While it's not strictly RemoteTech related, I am wondering if the folks having trouble building stable constellations are using PreciseNode.

Yep.

The issue is, even if you make it really close, it's never perfect, and that's all it takes.

But even with that, it's quite tedious to switch back and forth the satellites to have the closest orbit period possible. I did that with the first several network I set up and then felt I was just repeating the chore with no real progression or new discover. It stopped being fun.

Hit the nail on the head.

Edited by Deimos_F

Share this post


Link to post
Share on other sites
Better way? Other than manually fine tune the orbit period? Yes it's totally possible to achieve rather stable network but it's just very grindy and unnecessarily time consuming. It's okay to do that with first 1 or 2 networks. But once you do the same thing for the 20th time, it becomes very tedious. Optimizing orbit is fun the first few times, after that, it becomes a chore and repetitive. To each their own though. Some people like grinding in MMOs and I absolutely hate it.

Nah I don't find it that time consuming. Here's my better way - I bring one rocket that carries all 3 or 4 satellites that I expect to put on the same orbit. Then put the main rocket into a carefully calculated elliptic orbit so that Ap is the target orbit satellite height, and orbit period is 2/3 or 3/4 of the target satellite orbit period. Then rocket releases one satellite each time it reaches Ap, and satellite circularize itself and fine tune after release (and this fine-tune is just to get your semi-major axis the correct value - no orbit matching). The satellite will naturally be separated 120 degrees or 90 degrees apart by this method. No need to worry about rendezvous or matching up anything at all. One rocket for all satellites on the same orbit. For satellites on other planet it's often easier to do the opposite, i.e. put Pe as desired, set period to be 4/3 or 5/4 as desired - still the same idea.

The only down side is that there's a little bit trouble upon launch as the rocket would be highly non-aerodynamic at the top to put all satellites there. But that is just sparing more dV and have a slower launch - not a great deal, comparing to the tedious orbit matching between different launches (I agree that this is tedious).

Edited by FancyMouse

Share this post


Link to post
Share on other sites
Nah I don't find it that time consuming. Here's my better way - I bring one rocket that carries all 3 or 4 satellites that I expect to put on the same orbit. Then put the main rocket into a carefully calculated elliptic orbit so that Ap is the target orbit satellite height, and orbit period is 2/3 or 3/4 of the target satellite orbit period. Then rocket releases one satellite each time it reaches Ap, and satellite circularize itself and fine tune after release (and this fine-tune is just to get your semi-major axis the correct value - no orbit matching). The satellite will naturally be separated 120 degrees or 90 degrees apart by this method. No need to worry about rendezvous or matching up anything at all. One rocket for all satellites on the same orbit. For satellites on other planet it's often easier to do the opposite, i.e. put Pe as desired, set period to be 4/3 or 5/4 as desired - still the same idea.

The only down side is that there's a little bit trouble upon launch as the rocket would be highly non-aerodynamic at the top to put all satellites there. But that is just sparing more dV and have a slower launch - not a great deal, comparing to the tedious orbit matching between different launches (I agree that this is tedious).

I launch my networks using a rocket with 4 satellites on top. I cover them with a fairing to keep it aerodynamic, and jettison when I get about 50-60 km high

I use an upside down quad connector, with decouplers for each satellite. Works quite well. I usually put the rocket in a somewhat lower orbit, and then use the 2 ant engines on the satellites to raise and then circularize the orbits. Lots of dv on the satellites, even though they only have an oscar tank

However, I agree about it being tedious after the first few. You could use MechJeb to remove some of the tedium of the initial orbiting, but I wish there was a way to circularize the orbit at the exact correct altitude, once you get it "close enough". I've sent a message to the author of Hyperedit, since this seems to be a variation on what it can do.

LGG

Share this post


Link to post
Share on other sites
Nah I don't find it that time consuming. Here's my better way - I bring one rocket that carries all 3 or 4 satellites that I expect to put on the same orbit. Then put the main rocket into a carefully calculated elliptic orbit so that Ap is the target orbit satellite height, and orbit period is 2/3 or 3/4 of the target satellite orbit period. Then rocket releases one satellite each time it reaches Ap, and satellite circularize itself and fine tune after release (and this fine-tune is just to get your semi-major axis the correct value - no orbit matching). The satellite will naturally be separated 120 degrees or 90 degrees apart by this method. No need to worry about rendezvous or matching up anything at all. One rocket for all satellites on the same orbit. For satellites on other planet it's often easier to do the opposite, i.e. put Pe as desired, set period to be 4/3 or 5/4 as desired - still the same idea.

The only down side is that there's a little bit trouble upon launch as the rocket would be highly non-aerodynamic at the top to put all satellites there. But that is just sparing more dV and have a slower launch - not a great deal, comparing to the tedious orbit matching between different launches (I agree that this is tedious).

Well that's exactly my method as well. I launch 3~4 sats on one rocket and release them at apoapsis then change my orbit period to 4/3 to release the next one at apoapsis again. The thing is it's still tedious and time consuming to fine tune the orbit period of each satellite due to the game engine constraint, especially if you are doing this the 10th time. Seriously, in real life nobody jumps back and forth on different satellite fidgeting the orbit period. Here's the drift error data:

[TABLE=class: data, width: 100%]

[TR]

[TH=align: center]Period Error[/TH]

[TH=align: center]Drift Rate[/TH]

[TH=align: center]Time to drift out of contact (38°)[/TH]

[/TR]

[/TABLE]

0.01 seconds 0.00067° per orbit10 Earth years (33 Kerbin years)

0.1 seconds 0.0067° per orbit360 Earth days (3 Kerbin years)

1 seconds 0.067° per orbit36 Earth days (140 Kerbin days)

The problem is with the limitation of the game engine, you won't get the period error around 0.01 seconds stably. And even with 0.1 error, 3 kerbin years is simply too short because of the scale of the planet and universe in game. In real world, 1 second of period error isn't a big deal considering how big earth is but in the game, that simply won't work.

- - - Updated - - -

I launch my networks using a rocket with 4 satellites on top. I cover them with a fairing to keep it aerodynamic, and jettison when I get about 50-60 km high

I use an upside down quad connector, with decouplers for each satellite. Works quite well. I usually put the rocket in a somewhat lower orbit, and then use the 2 ant engines on the satellites to raise and then circularize the orbits. Lots of dv on the satellites, even though they only have an oscar tank

However, I agree about it being tedious after the first few. You could use MechJeb to remove some of the tedium of the initial orbiting, but I wish there was a way to circularize the orbit at the exact correct altitude, once you get it "close enough". I've sent a message to the author of Hyperedit, since this seems to be a variation on what it can do.

LGG

Let me tell you even with Hyperedit placing satellites in "perfect" orbit, it is still not accurate and the constellation deteriorate over time. You will see your orbit wobble a lot and orbit period flickering back and forth and all different on different satellite, even if you set exactly the same parameters for them. It's just the game engine that can't be that accurate.

Share this post


Link to post
Share on other sites
How do you calculate the orbital period?

You can see it with Kerbal Engineer or Mechjeb.

Share this post


Link to post
Share on other sites
Well that's exactly my method as well. I launch 3~4 sats on one rocket and release them at apoapsis then change my orbit period to 4/3 to release the next one at apoapsis again. The thing is it's still tedious and time consuming to fine tune the orbit period of each satellite due to the game engine constraint, especially if you are doing this the 10th time. Seriously, in real life nobody jumps back and forth on different satellite fidgeting the orbit period. Here's the drift error data:

[TABLE=class: data, width: 100%]

[TR]

[TH=align: center]Period Error[/TH]

[TH=align: center]Drift Rate[/TH]

[TH=align: center]Time to drift out of contact (38°)[/TH]

[/TR]

[/TABLE]

0.01 seconds 0.00067° per orbit10 Earth years (33 Kerbin years)

0.1 seconds 0.0067° per orbit360 Earth days (3 Kerbin years)

1 seconds 0.067° per orbit36 Earth days (140 Kerbin days)

The problem is with the limitation of the game engine, you won't get the period error around 0.01 seconds stably. And even with 0.1 error, 3 kerbin years is simply too short because of the scale of the planet and universe in game. In real world, 1 second of period error isn't a big deal considering how big earth is but in the game, that simply won't work.

- - - Updated - - -

Let me tell you even with Hyperedit placing satellites in "perfect" orbit, it is still not accurate and the constellation deteriorate over time. You will see your orbit wobble a lot and orbit period flickering back and forth and all different on different satellite, even if you set exactly the same parameters for them. It's just the game engine that can't be that accurate.

So it seems that a mod which will keep a satellite in a specific orbit is needed.

Any volunteers?

Share this post


Link to post
Share on other sites
But even with that, it's quite tedious to switch back and forth the satellites to have the closest orbit period possible. I did that with the first several network I set up and then felt I was just repeating the chore with no real progression or new discover. It stopped being fun.

Not really. I don't switch back and forth much at all. I just pick one satellite to be the reference, write down the Orbital Period, and switch to the other two and make them the same. If they need some adjustment I will expand or contract their orbit until they are in harmony again and then fix the orbital period.

Share this post


Link to post
Share on other sites

The problem is with the limitation of the game engine, you won't get the period error around 0.01 seconds stably. And even with 0.1 error, 3 kerbin years is simply too short because of the scale of the planet and universe in game.

You're quoting the table for a specific period 1.5h (or semi-major axis 1376km). If you find the error too significant, just make your satellite higher so that orbit period is bigger. There's a similar table in another page of RT tutorial that shows the error for synchronous orbit - that's hundreds of years for 0.01 sec, simply because the orbit period is much bigger. That should give you a hint that higher orbit tolerates more about the errors. Of course it can't be too high because of antenna range - that's the place you need to carefully design.

That's what I said choosing a better orbit is also part of the design (and part of the fun).

And no, higher precision is achievable. My Kerbin omni satellite network has semi-major axis 2Mm +- 1m, it's about 2.7e-4 degree per orbit error - even better than the 0.01 sec drift rate for the low 1.5h orbit.

Edited by FancyMouse

Share this post


Link to post
Share on other sites
You're quoting the table for a specific period 1.5h (or semi-major axis 1376km). If you find the error too significant, just make your satellite higher so that orbit period is bigger. There's a similar table in another page of RT tutorial that shows the error for synchronous orbit - that's hundreds of years for 0.01 sec, simply because the orbit period is much bigger. That should give you a hint that higher orbit tolerates more about the errors. Of course it can't be too high because of antenna range - that's the place you need to carefully design.

That's what I said choosing a better orbit is also part of the design (and part of the fun).

And no, higher precision is achievable. My Kerbin omni satellite network has semi-major axis 2Mm +- 1m, it's about 2.7e-4 degree per orbit error - even better than the 0.01 sec drift rate for the low 1.5h orbit.

Like I said, it is achievable. Limit thrust, RCS and capslock I can do that too. But it's tedious and time consuming. I can tweak that to perfection but with the same amount of time, I could do more other things, not to mention how much you have to do it when you set up multiple network.

The table I provided is exactly the data from the synchronous orbit and yet you see 0.1 second error will collapse the network within just short 3 years. That's pretty low tolerance compared to real world.

Moreover, I play career mode so high orbit isn't really possible in the beginning. With the starting Communotron 16, the highest orbit you can go without breaking the line of sight is around 840km. Even that is pushing the margin of error.

I have done some extensive testing. Even with Hyperedit, with exact same orbit parameter, the orbit shifts by itself and the orbit periods desynchronize at an accelerated rate as time goes by. There's simply no way to lock the orbit or at least have it oscillate within certain margin. Even on the tutorial page, it clearly says this:

It is nearly impossible to give two satellites exactly the same orbital period, because the period will change whenever the satellite rotates (for example, to keep facing the sun over the course of the year). For the last word in satellite synchronization, you may wish to edit your save file. Some RemoteTech players see this as essential to get around game engine limitations, others see it as cheating. You will have to decide for yourself.

Edited by hbkmog

Share this post


Link to post
Share on other sites
You're quoting the table for a specific period 1.5h (or semi-major axis 1376km). If you find the error too significant, just make your satellite higher so that orbit period is bigger. There's a similar table in another page of RT tutorial that shows the error for synchronous orbit - that's hundreds of years for 0.01 sec, simply because the orbit period is much bigger. That should give you a hint that higher orbit tolerates more about the errors. Of course it can't be too high because of antenna range - that's the place you need to carefully design.

That's what I said choosing a better orbit is also part of the design (and part of the fun).

And no, higher precision is achievable. My Kerbin omni satellite network has semi-major axis 2Mm +- 1m, it's about 2.7e-4 degree per orbit error - even better than the 0.01 sec drift rate for the low 1.5h orbit.

Now that I have better tech I'm considering thrashing my initial 1.5 orbit ring and only having 3 sats for all kerbin soi relaying: a keo and 2 9-days Molnars . Bye bye omnis and constellations.... it was fun while it lasted. It's simply not worth upgrading my initial ring with better batteries, panels and orbits. I'd rather face occasional loss of signal with kos than launching and/or maintaining.

Anyway, that mod will still be on my wishlist. Because duna ;)

Edit: anyway the molnars have big dishes for outer planets.

Share this post


Link to post
Share on other sites
It is realistic - just like in real life your orbit will get perturbed and you need to fine tune regularly. I feel strongly changing the value directly a cheat, regardless of whether you edit the file yourself or via some tiny window.

But what's not realistic is that in real life you can have them run software that maneuvers them for you, and they can run it without you having to take time off other projects to do it. Even with a kOS stationkeeping script you still can't escape from the fact that KSP doesn't let you move a vessel that's outside your focus range. The smartest stationkeeping script in the world won't help when the vessel is on rails until you focus on it.

Share this post


Link to post
Share on other sites

Been out of the loop for a minute.. Is it possible yet to have an antenna recieve a signal without the capability of re-transmitting it? Seems a bit silly to me that every single object with antennas are capable of performing the job of a dedicated communications relay. Always something thats irked and honestly the only thing I can currently think of that I have always wanted changed as it seems mods have fixed everything else at this point.

Hopefully somebody knows a way to do this and can explain.

Share this post


Link to post
Share on other sites
Been out of the loop for a minute.. Is it possible yet to have an antenna recieve a signal without the capability of re-transmitting it? Seems a bit silly to me that every single object with antennas are capable of performing the job of a dedicated communications relay. Always something thats irked and honestly the only thing I can currently think of that I have always wanted changed as it seems mods have fixed everything else at this point.

Hopefully somebody knows a way to do this and can explain.

If you use a directional antenna, it can only communicate to one target.

Share this post


Link to post
Share on other sites
If you use a directional antenna, it can only communicate to one target.

If you're lucky enough to get another vessel within your cone (check right bottom buttons on map view) you can tx/rx to it as well.

@subcidal If you really want realism you should stick to ground stations mostly how NASA does. The real world TDRS constellation is to be put to shame by any means in ksp: 2 very big low range dishes (one client at a time, even if you could put them in the same physical place) and one omni. And basic coverage on Mars but that's it. I have more fun with remote tech.

Share this post


Link to post
Share on other sites
Now that I have better tech I'm considering thrashing my initial 1.5 orbit ring and only having 3 sats for all kerbin soi relaying: a keo and 2 9-days Molnars . Bye bye omnis and constellations.... it was fun while it lasted. It's simply not worth upgrading my initial ring with better batteries, panels and orbits. I'd rather face occasional loss of signal with kos than launching and/or maintaining.

Anyway, that mod will still be on my wishlist. Because duna ;)

Edit: anyway the molnars have big dishes for outer planets.

How does that work? I do not even know what a molnar orbit is?

Share this post


Link to post
Share on other sites
How does that work? I do not even know what a molnar orbit is?

My orbits are more like inspired in those orbits. I send a big satellite with a lot of dishes into a 90 degree inclination, very eliptic orbit.

Then this guy has 4 small dishes for Kerbin SOI (mun, minmus and relay through something to get to KSC)

4 medium for inner planets and duna

3 big, one for each outer planet

1 big, for any active vessel within Kerbol SOI :D

oh, and an omni for when I'm getting near periapsis. Anyway I plan on having 2 of these, so when one is offline the other one covers me.

pE3dVgA.png

OMG! I've been spelling this thing wrong the whole time! :blush:

In case you're interested, there's also tundra orbits (I hope I spell right this time) which are very similar.

http://en.wikipedia.org/wiki/Tundra_orbit

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.