Now-defunct-thread-that-should-not-appear-in-google-search.

[Rokker]

Last time i checked Cliph is still working on kOS compatibility. So probably not.

[krenshala]

AdmTigerclaw, your setup looks a lot like mine, only I used five instead of six for that outer ring.

My basic setup is a ring of 5 (or 6) sats, equidistant from each other on orbits synched to the tenth of a second (thank you, Kerbal Engineer Redux!). Those sats have a semi-major axis of 1,584,953.3m (orbital period of 3 hours 0.0003 seconds). Each sat is at least one basic omni (2.5Mm range) and a 90Mm dish for Mun/Minmus comms. Early versions used smaller dishes to interconnect the sats to the ones before and after in line, but I've found the omnis work better for that.

I have found that if you take the time to get the orbital period synched to at least half a second of each other then even long periods of time warp lead to very little drift due to rounding errors. Of course, I haven't done much long duration spaceflight yet, so it may make more of a difference once I do.

[AdmiralTigerclaw]

AdmTigerclaw, your setup looks a lot like mine, only I used five instead of six for that outer ring.

My basic setup is a ring of 5 (or 6) sats, equidistant from each other on orbits synched to the tenth of a second (thank you, Kerbal Engineer Redux!). Those sats have a semi-major axis of 1,584,953.3m (orbital period of 3 hours 0.0003 seconds). Each sat is at least one basic omni (2.5Mm range) and a 90Mm dish for Mun/Minmus comms. Early versions used smaller dishes to interconnect the sats to the ones before and after in line, but I've found the omnis work better for that.

I have found that if you take the time to get the orbital period synched to at least half a second of each other then even long periods of time warp lead to very little drift due to rounding errors. Of course, I haven't done much long duration spaceflight yet, so it may make more of a difference once I do.

I keep the dishes off my lower orbit sats and reserve those for my KEOsats. Then I aim all my KEOsats at a 'hub' like my Mun <> Minmus intermediate orbital command station.

[Peppe]

My biggest concern is connection reliability and redundancy.

I don't much like working on a long range probe if at the critical time in flight, the connection drops out because there's a hole in the network.

And the biggest cause of holes is the rounding error from things going on the rails. Without time warp, a well placed constellation takes a long time to drift out of phase. Once Time Warp kicks in, the entire thing becomes a mess in the time it takes to send a ship to minmus.

So I try to set up a network that maximizes uptime while mitigating any down time ASAP.

Uptime tends to result from a combination of coverage and link redundancy. The more links I have between sats that have wide coverage area, the better the uptime.

In this case, my primary goal is to keep the primary link to KSC up to about 99.99% (Mathematically, there WILL be a gap at some point in the game, meaning it is impossible to achieve perfect 100%)

In order to ensure I have coverage and connections, I consider that my satellites have a very wide field of view of kerbin, while also being close enough that I can get them to cross-communicate with each other unless they're ocluded by kerbin itself.

This way, even if they bunch-up in their orbits due to time warp, they still have LoS and range to talk to other satellites in the constellation.

Like this group:

http://img.photobucket.com/albums/v430/admiraltigerclaw/SpikeSatArray_zps9a50f8c1.jpg

I forget what sat set that was, but even though the two up on the top side of the picture are bunched up, it's clear by the cat's cradle of connections that even if they get wonky, down time remains very low.

And in the image, the sats that are in lower standby orbits have higher frequencies, which allows them to swing around and clear gaps in coverage that could pop up due to drift. That's not intentional however, just a side-effect of popping a bunch into LKO first, then seeing how many I need to place.

Going for simplest is a good goal, but in my opinion, it's too easy to 'break'.

I think the ideal is probably an orbit just under the antenna limit. A larger orbit means a longer period, so the rounding/accuracy issues with orbits is reduced.

For accurate orbits though to me it is VOID > Flight Engineer > Mechjeb

I don't have mechjeb installed, but it only shows down to the second right?

Here is Flight engineer vs void:

With VOID i use the line below orbital period, semi-major axis. If you want to eliminate drift i don't think you can beat that last meter of fidelity VOID gives.

I think the 1/10 second ker shows is less accurate than the millimeters VOID shows. If take that out to a very high and long orbit you should be able to get orbits that drift won't be a problem for the life of your game. If you are patient you can use the milliliters as it saves the values displayed when you pause to leave for space center, but controlling the satellite you can only really narrow it down to a meter unless you want to rapidly puase/resume to get the final cm/mm equal.

From that screenshot this was saved in my persistence file for the craft SMA:

1302717.39773173

I went to my other 2 satellites and did the same thing and got them all to 1302717.xxx. Simmed a year at 100k and they were fine, though i did not write down their distances before reverting. They are in 700km orbits, which works fine, but if the goal is max long term stability I think the larger the orbit the better.

3 is a little light for redundancy, so I would probably just double it and make them higher. What is the worst case, sat just coming over horizon and KSC rotating away to the other side? 5mm antenna on 6 satellites @3500 altitude should give max overlap, an accurate orbit, and max equatorial coverage.

Here is my scratchpad:

[dlrk]

If KOs isn't implemented yet, what is the method for unmanned missions to distant areas?

[AdmiralTigerclaw]

@Peppe

Don't forget to account for one other thing when you're talking about Max Antenna Range.

It's not talking to the ground I'm wheeling for, but talking with other satellites.

If, for example, I put a satellite with the 5 Mm antenna up at say, 4,000 km, then there's no way it can talk with any satellite outside a little more than a third of the circumference around. I like my low sats to be low enough so that they can interlink.

As I said, part of my strategy for my constellation is redundancy through connection numbers.

It's based on my gameplay philosophy. Set it and forget it.

My philosophy is based on one simple thing. In real life, space infrastructure is maintained by hundreds, if not thousands of people watching things around the clock. Any deviations in the infrastructure are corrected almost immediately as needed almost non-stop.

As a player, I take all those jobs and roll them into one person, me. And whatever assisting supporting programs I may have. I simply do not have the ability to split my attention to handle the kind of micromanaging that is required to keep everything well-oiled like that. Nor do I have the time to fix things.

I'll probably go grab VOID if it provides me with even more accurate orbits, but the fact remains that as a Player, I don't have the ability mass-manage anything that is not 'set it and forget it'. One of the reasons why I prefer redundancy over raw accuracy.

The accuracy WILL collapse, I can't escape that. I can slow it down, but that's just a stopgap measure. But I can orchestrate a mess that effectively has no holes in it save for freak events on par with The Stars Being Right.

That's my philosophy/strategy.

[Peppe]

@Peppe

Don't forget to account for one other thing when you're talking about Max Antenna Range.

It's not talking to the ground I'm wheeling for, but talking with other satellites.

If, for example, I put a satellite with the 5 Mm antenna up at say, 4,000 km, then there's no way it can talk with any satellite outside a little more than a third of the circumference around. I like my low sats to be low enough so that they can interlink.

As I said, part of my strategy for my constellation is redundancy through connection numbers.

It's based on my gameplay philosophy. Set it and forget it.

My philosophy is based on one simple thing. In real life, space infrastructure is maintained by hundreds, if not thousands of people watching things around the clock. Any deviations in the infrastructure are corrected almost immediately as needed almost non-stop.

As a player, I take all those jobs and roll them into one person, me. And whatever assisting supporting programs I may have. I simply do not have the ability to split my attention to handle the kind of micromanaging that is required to keep everything well-oiled like that. Nor do I have the time to fix things.

I'll probably go grab VOID if it provides me with even more accurate orbits, but the fact remains that as a Player, I don't have the ability mass-manage anything that is not 'set it and forget it'. One of the reasons why I prefer redundancy over raw accuracy.

The accuracy WILL collapse, I can't escape that. I can slow it down, but that's just a stopgap measure. But I can orchestrate a mess that effectively has no holes in it save for freak events on par with The Stars Being Right.

That's my philosophy/strategy.

Yeah i was starting on an edit to that. My thought was trying to figure the mean time to failure.

How would calculate mean time to failure? Or say worst case. Say we got things down to the meter, but were off on two neighbors in opposite ways. So net they are moving away from another at 2 meters an orbit. The orbit would be over a day, but we'll just call it a day. With 6 satellites failure would be when the 2 reach --- i guess antenna limit from eachother. At 3500km orbits That would require a drift of ~900km? They should start at 4100 km away from eachother and 5000k should be the antenna limit, drift one away = 900,000 / 2 = 450,000 = kerbin year is 426 days? = 1056 years to failure.

I must be missing something or not enough people use VOID for fine tuning ;P

[Beetlecat]

If KOs isn't implemented yet, what is the method for unmanned missions to distant areas?

There is a self-contained flight computer that allows inputs like orientation and node execution --with optional signal delay.

[Shania_L]

Only issue would be where you want those ground stations. If all they have is the 500km antenna can probably only go 40-45 degrees above the equator before they are out of range. The further up north/south you go the less time they will be in range as the field of view is a bunch of circles moving across the equator -- at the top of the circles covered will go in and out as the satellites move. The satellites at 300k cannot see much further than the 45-50 degree north/south.

What information mod are you using for orbits? With low orbits you run into an accuracy issue. I use VOID to see accuracy of the orbit semi-major access down to the meter, but the game bounces that last meter around whenever it is in physics range, so when synced up my sats I do it down to that last meter.

At 300km your orbits are under an hour, so if you warped say 100,000x to line up a shot to Jool your orbits could move even in an ideal case 5-6 meters a day and pretty quickly they won't be a reliable relay.

Given our accuracy is ~1 meter per orbit, the higher you can take the orbit the longer their relative distance to each-other will hold and the longer your relay will operate without adjustments. If you have a healthy overlap in coverage say you are at 600km with 4 satellites, start them in a box shop and they won't fail until the sides collapse down to a triangle shape.

If you don't use void and instead use kerbal engineer or mechjeb that don't show semi-major axis or go on just orbital period down to the second/tenth of second your orbits relative to each-other could be much more than a meter off.

Land based polar dishes for deep space is an interesting plan and think that part is solid. I don't think there would be a blind spot. The sun may still get in your way.

Thanks for the reply Peppe,

The ground stations have only the 500km antenna extended throuought re-entry and landing, but once landed I will extend dishes and the 2.5Mm antenna. This I think is going to become a problem landing probes on other worlds, I have to shut down everything but the 500km just to avoid breaking the antennas.

I am not using any mods for orbit information, just the games own orbit displays. As such I am fairly happy getting a Perapisis/Apoapsis within 200-400m of each other, which is why I am having to build such redundancy in I expect a lot of drift.

As far as the land based dishes go, how accurately does RT plot the surface form of a planet? does it take an average mean surface and assume a perfect circle, or would landing next to a mountain give me a massive blind spot?

[match]

It's based on my gameplay philosophy. Set it and forget it.

You, sir, are a Lazy Perfectionist. Not to be confused with the OCD Perfectionist. We of the Lazy Perfectionist clan have a healthy respect for our OCD brethren. However, they spend far too much energy fiddling with it, which normally results in a firm, yet encouraging "Right there is perfect!"

I'm close to abandoning the minimalist approach. While creating infinite 100% coverage is possible in theory, soon I'll take the cue from modern networks and build redundancy instead. The best I've been able to achieve is around 400 days before failure with a 3 satellite setup using orbital period (Kerbal Engineer Redux) as the constant. I have nothing conclusive, but I was getting better results when minimizing inclination with respect to orbital period (Fun Fact... there is a difference between 1:29:60.0 and 1:30:00.0... who knew?). I played around with eccentric orbits some, but really found no improvement (more time between failure, but more effort to correct). I resolved to set Kerbal Alarm Clock for yearly maintenance every six months.

The only way to truly negate the effects of drift would be to place the antennae in a static location... ground stations. I was considering the polar solution with the tidally locked Mun as relay as the foundation, but building/placing two stations 750 m tall directly at the poles seems a bit of work to provide only half a solution. Could use two stations at the Mun's poles to reduce the height a tad, or build sister stations offset from the poles...still a daunting task. I wonder what the elevation at the poles are to offset the height need?

Well, I got my task for the day... off to the poles!!

[BigD145]

You, sir, are a Lazy Perfectionist. Not to be confused with the OCD Perfectionist. We of the Lazy Perfectionist clan have a healthy respect for our OCD brethren. However, they spend far too much energy fiddling with it, which normally results in a firm, yet encouraging "Right there is perfect!"

There is no perfect. There is only drift.

[DerekL1963]

So... my RT2 installation seems to be half broken....

I encountered a glitch and had to wipe and re-install KSP and start a new game, only now RT2 (1.3.3, the version from Spaceport) is only half working. In sandbox and career modes, I get the correct RT2 display on the map and flight screens, but no 'tech bonus' (the 3.5km built in antenna) on any pods. In sandbox mode the antennas show up in the VAB, but there's no range or power consumption information. In career mode, the antenna's are not showing up in the tech tree.

Any ideas on where/how to fix this?

[sonicsst]

Do you have the ModuleManager_1_5.dll in your GameData folder?

[captainreynolds]

I've looked for this and have not found an answer anywhere.

I already have a very large satellite network in place, with each satellite having 4 communotron 16's, and one communotron 88-88. I did it mainly for fun, but it uses all stock parts. Here is a picture of it. http://i.imgur.com/FeFI2B5.png

I did it mainly for fun, but would like to use remote tech on it. However, having just installed the remotetech mod, it evidently retracts all my antennae that were open previously when I focus on a satellite, and when I try to do anything, it says "No Connection to send command on." What does this mean?

[DerekL1963]

Do you have the ModuleManager_1_5.dll in your GameData folder?

No, and I figured out that was the problem about three minutes after I posted here when I saw Module Manager mentioned in another thread.

It's downloaded and installed and KSP is loading even as I type.

[Peppe]

You, sir, are a Lazy Perfectionist. Not to be confused with the OCD Perfectionist. We of the Lazy Perfectionist clan have a healthy respect for our OCD brethren. However, they spend far too much energy fiddling with it, which normally results in a firm, yet encouraging "Right there is perfect!"

I'm close to abandoning the minimalist approach. While creating infinite 100% coverage is possible in theory, soon I'll take the cue from modern networks and build redundancy instead. The best I've been able to achieve is around 400 days before failure with a 3 satellite setup using orbital period (Kerbal Engineer Redux) as the constant. I have nothing conclusive, but I was getting better results when minimizing inclination with respect to orbital period (Fun Fact... there is a difference between 1:29:60.0 and 1:30:00.0... who knew?). I played around with eccentric orbits some, but really found no improvement (more time between failure, but more effort to correct). I resolved to set Kerbal Alarm Clock for yearly maintenance every six months.

The only way to truly negate the effects of drift would be to place the antennae in a static location... ground stations. I was considering the polar solution with the tidally locked Mun as relay as the foundation, but building/placing two stations 750 m tall directly at the poles seems a bit of work to provide only half a solution. Could use two stations at the Mun's poles to reduce the height a tad, or build sister stations offset from the poles...still a daunting task. I wonder what the elevation at the poles are to offset the height need?

Well, I got my task for the day... off to the poles!!

How about 20km of drift over a 7 year 100k warp?

Javascript is disabled. View full album

Later hitting escape rapidly to dial in the millimeter mark I got 2 sats semi major access 20 millimeters apart. Using 48-7s with thrust tweaked to min and burning near 90 degrees off path.

I guess I am a perfectionist, but I will not resort to save game Editting for no drift during the life of my game.

[DerekL1963]

Aaannd.. installing Module Manager did indeed fix the problem.

Moral of the story, do not re-install KSP and a metric buttload of mods without having at least *finished* that first cup of coffee.

[Peppe]

Thanks for the reply Peppe,

The ground stations have only the 500km antenna extended throuought re-entry and landing, but once landed I will extend dishes and the 2.5Mm antenna. This I think is going to become a problem landing probes on other worlds, I have to shut down everything but the 500km just to avoid breaking the antennas.

I am not using any mods for orbit information, just the games own orbit displays. As such I am fairly happy getting a Perapisis/Apoapsis within 200-400m of each other, which is why I am having to build such redundancy in I expect a lot of drift.

As far as the land based dishes go, how accurately does RT plot the surface form of a planet? does it take an average mean surface and assume a perfect circle, or would landing next to a mountain give me a massive blind spot?

Radio line of site is a perfect sphere @ height of the ocean.

The breakable antenna/dishes only occur in atmosphere, so that would only affect a few locations in the kerbol system. It will be harder to maintain contact with un-manned probes around kerbin's surface than on the mun's.

I have not tested delayed commands using action groups. Put if you can set the flight computer delay to sometime after your landing time you could setup a re-entry, queue up a command to extend the antenna, remove the delay and retract the antenna's/solar panels etc, land -- wait for queued command to extend antenna/panels. Probe back online on surface.

I'd still be concerned where you land your probe on kerbin with just the 300k equatorial sats. A probe on the ground in kerbin won't be able to see them above those same latitudes i mentioned before.

[AdmiralTigerclaw]

The only way to truly negate the effects of drift would be to place the antennae in a static location... ground stations.

Mobile Surface Command Truck.

[cfds]

If now only the solar panels and antenna would not break of on loading the vehicle...

[BigD145]

I've looked for this and have not found an answer anywhere.

I already have a very large satellite network in place, with each satellite having 4 communotron 16's, and one communotron 88-88. I did it mainly for fun, but it uses all stock parts. Here is a picture of it. http://i.imgur.com/FeFI2B5.png

I did it mainly for fun, but would like to use remote tech on it. However, having just installed the remotetech mod, it evidently retracts all my antennae that were open previously when I focus on a satellite, and when I try to do anything, it says "No Connection to send command on." What does this mean?

Read the OP.

[AdmiralTigerclaw]

If now only the solar panels and antenna would not break of on loading the vehicle...

That's not loading, that's deployment. It just came OFF the plane.

[JewelShisen]

Are there any plans to bring back the AeroProbe or MicroSat from RT1?

[Aeon-Phoenix]

I've read a couple pages past to see if my problem got covered already. If it is somewhere beyond the page 150 below. Sorry in advance.

My problem: I started to set up a basic geosyncronous sat network around Kerbin in career mode. My first sats had two DTS-M1 dishes. The first dish you unlock in carreer mode. The first one went right on (thanks to my "skylab" - which relayed perfectly). the Second sat worked OK too. but i never got the third sattelite going as it refused to connect to the second sat to complete the network. I triplechecked if they are all targeted right. a couple times. to no avail.

Then i started a second network using three dishes each sat. Two DTS and the first non-stock dish you unlock (sorry the name slipped me). the DTS should have enough range to connect to the each of their neighbours and one would point to active vessel or mission control. Again first two sattelites went up alright. but again the third simply refuses to connect to the second sattelite launched. I checked electricity, line of sight, range... to no avail.

I mean they even get shown in map mode that there is a connection between them. but the signal does not get relayed to the first sat (even tho the 2nd in line perfectly works and IS connected to the last sat in the chain. I'm baffled and confused.

Or do you need a dish for EACH "channel"? AKA. one channel to control the sat and one channel to relay?

[AdmiralTigerclaw]

I've read a couple pages past to see if my problem got covered already. If it is somewhere beyond the page 150 below. Sorry in advance.

My problem: I started to set up a basic geosyncronous sat network around Kerbin in career mode. My first sats had two DTS-M1 dishes. The first dish you unlock in carreer mode. The first one went right on (thanks to my "skylab" - which relayed perfectly). the Second sat worked OK too. but i never got the third sattelite going as it refused to connect to the second sat to complete the network. I triplechecked if they are all targeted right. a couple times. to no avail.

Then i started a second network using three dishes each sat. Two DTS and the first non-stock dish you unlock (sorry the name slipped me). the DTS should have enough range to connect to the each of their neighbours and one would point to active vessel or mission control. Again first two sattelites went up alright. but again the third simply refuses to connect to the second sattelite launched. I checked electricity, line of sight, range... to no avail.

I mean they even get shown in map mode that there is a connection between them. but the signal does not get relayed to the first sat (even tho the 2nd in line perfectly works and IS connected to the last sat in the chain. I'm baffled and confused.

Or do you need a dish for EACH "channel"? AKA. one channel to control the sat and one channel to relay?

It's not clear what you're saying exactly, at least to me.

But think of the dishes themselves as talking face to face. When doing satellite aiming, two satellites must have a dish on BOTH facing the opposing sat as if having a conversation with each other. If you have a third satellite come in and aim a dish at them, it's going to be ignored. In order to speak with the third sat, sat 2 or 1 must have a free dish to aim at the third one.

The biggest mistake I see going on is that people aim one dish, but either don't realize, or forget that the receiver is a dish as well, and can't just receive from any direction and must be aimed, just like turning to face someone speaking to you.

It gets horribly complicated really fast if you're trying to chain satellite constellations using dishes. That's why I prefer using lower orbit omnis, and linking them to satellites in synchronous orbit, then having all the sats in sync orbit pointing at a hub that's half way to minmus.

Also, remember that your interlink has to chain back to a mission control. That is defined as KSC itself, or a vessel containing SIX (6) kerbals AND a large probe core. If your chain is broken somewhere else, then you don't have the uplink, even if you have connectivity.

I suspect some Kerban Error to be involved with the 'aim at active' command. In this case, you work on sat 1, tell it to aim at the active, then launch sat 2, and sat 1 aims at sat 2. Then you tell sat 2 to aim at the active and launch sat 3. Now sat 1 AND 2 aim at sat 3, which means sat 1 is not longer talking to sat 2. Further, because 'talk to the face' with sat 3 aimed at sat 2, sat 1's attempt to link is being ignored by sat 3. Oops, link is broken... TWICE.

Edited January 5, 2014 by AdmiralTigerclaw