Post Reply 
 
Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
That Miraculous Device: Onboard Radio
Author Message
Newt Offline
Flying Buttress
***

Posts: 134
Joined: Dec 2014
Post: #1
That Miraculous Device: Onboard Radio
This thread's focus will be on the design of the satellite's radio, and though some discussion of the ground station will be inevitable, that should probably be reserved for a separate thread to be begun later.

The main purpose of the radio is to communicate between the media CPU, and through that relay data, in the form of photographs and likely other sensor inputs, to the ground station(s) on request. The antenna will most probably be required to be omnidirectional, in view of the current design, which calls for a spinning satellite, whose axis points to the sun.

Currently, I am trying to look generally into information about how to design, and importantly evaluate, the radio that we can know what downlink speed to expect, how powerful of a signal we will need et cetera. Some of us here are hams, or have other experience working with and building radios, and that will be valuable, but as far as I know, currently the radio is a somewhat vague. We will probably be depending on hams (yet to be clearly arranged), for most of the communication and ground stations, and so we will probably aspire to create a system comparable to other amateur stations, and we should use such designs probably as a template to ensure compatibility.

Resources:

Rhoark of the KSP forum posted this link, to a poorly formatted but, I think useful, evaluation of some cubesat radio work. Not all of it is relevant to the discussion here, but some of it certainly is enlightening.

Additionally, here is an amateur radio faq page. Again, not all is relevant, but some of it is helpful, and I think that we could get some good insights from the work of AMSAT.
(This post was last modified: 12-19-2014 10:08 AM by Newt.)
12-19-2014 10:04 AM
Find all posts by this user Quote this message in a reply
Bunsen Offline
Junior Member
**

Posts: 18
Joined: Dec 2014
Post: #2
RE: That Miraculous Device: Onboard Radio
For anyone who already knows a thing or two about ham radio, but hasn't designed a satellite link before, Jan King has assembled the most terrifyingly glorious mass of Excel sheets in history, for the purpose of calculating all the calculables relevant to space communication. (Note: I don't work in the sort of field that uses Excel for real work, so a 2.1 MB spreadsheet seems unusual. If that's normal for you, well, you have my condolences.) http://www.amsatuk.me.uk/iaru/AMSAT-IARU...v2.5.3.xls

It's a bit overcomplicated for a simple cubesat (want to work up a link budget for an X-band link to a deep space probe just the other side of Mars that may not be pointing in quite the right direction? No problem!) but once you understand what each fill-in-able blank there means, you pretty well have the technical bases covered.

Newt's quite right that the satellite should use an antenna that's reasonably close to omnidirectional. A spacecraft that can't communicate when things are going wrong (e.g. the attitude control logic is feeling wonky) is a short-lived thing. So long as one is moderately patient about obtaining those photographs, a reliable link can be set up rather cheaply.

There are some technologies just showing up in the "dirt cheap and Arduino-compatible" market that could really help with this -- I'm looking really hard at Semtech's LoRa line of direct sequence spread spectrum radio chips (and HopeRF's integrated modules based on them). It's looking like one could have a low-speed telemetry/command link that would be very, very gentle on the satellite's power budget and not even require directional antennas on the ground, at a cost of around $10 per radio (plus antennas and such). I haven't figured out if there are any unique licensing hurdles for using that on an amateur-licensed spacecraft.

Omnidirectional antennas on both ends won't get you a link fast enough for convenient downlinking of sizable images, though, so there's still a place for a system that demands a tracking ground station with high-gain antennas.
12-19-2014 05:27 PM
Find all posts by this user Quote this message in a reply
Newt Offline
Flying Buttress
***

Posts: 134
Joined: Dec 2014
Post: #3
RE: That Miraculous Device: Onboard Radio
Truly a most glorious spreadsheet, Bunsen, thank you for it. I will surely play a bunch with it. When it took about 30 seconds to download I was quite surprised, but It will I am sure be helpful, it already may be.

Recently I have been reading about antennae, trying to figure out which frequencies we will best be able to transmit on based on the size of the sat, and which antenna types will be best for our situation. There is a lot of material, most of which is geared to Earthbound stations (understandably), and it takes a long time to sift through so I do not have very much clearly to say about that at this time.

The Arduino radio sounds like it might be worthwhile to look into. I will have to read some about that too, but it looks as though we will be using a Raspberry Pi computer for our image handling, so whatever radio must be compatible with that, unless maybe it is a backup/beacon that the flight computer runs (in which case it will need to talk to the flight computer)

We should reserve the ground station debate for another thread, but I will note that previous discussions have lead to the apparent consensus of trying to contact hams for most of the comm. That enables better coverage for a lower price, and looks practical to try. They will have directional antennae, tracking et cetera.
12-21-2014 03:44 AM
Find all posts by this user Quote this message in a reply
MBobrik Offline
Member
***

Posts: 203
Joined: Dec 2014
Post: #4
RE: That Miraculous Device: Onboard Radio
I took the time to fill the sheet with my uneducated guesses realistic values, and came to very implausibly looking numbers.
@433 mHz, 10 W dipole on the sat, 50 W biggest choice yagi on the ground, best LNA circuit specs I could google, I came to 500 kbps downlink speed. Which is an order of magnitude above initial K^2's 50 kbps estimate. Too good to be true, I guess... Could someone else, who got actually a clue in RF, try it too ?
12-21-2014 08:40 PM
Find all posts by this user Quote this message in a reply
Bunsen Offline
Junior Member
**

Posts: 18
Joined: Dec 2014
Post: #5
RE: That Miraculous Device: Onboard Radio
A 10W downlink is pretty unrealistic on the satellite (1-3W is more likely, depending on how deep you want to discharge the batteries during a nighttime downlink session and how much heat you can afford to dump into the satellite), and you want to build in some more margin than the sheet's recommended minimum-- 10dB at 5° or 10° elevation is probably healthy. Also, the LNA's ability to help will be limited by local ground-level background noise, unless the ground station is in the absolute middle of nowhere. I'd estimate that 10-50 kbps is plausible, depending on modulation (the traditional ham radio systems are pretty inefficient, but fancier systems are less broadly compatible) and how big a ground station you want to require.

The uplink is much easier, because electricity and heatsinks are practically free on the ground.
12-22-2014 01:18 AM
Find all posts by this user Quote this message in a reply
MBobrik Offline
Member
***

Posts: 203
Joined: Dec 2014
Post: #6
RE: That Miraculous Device: Onboard Radio
(12-22-2014 01:18 AM)Bunsen Wrote:  A 10W downlink is pretty unrealistic on the satellite (1-3W is more likely, depending on how deep you want to discharge the batteries during a nighttime downlink session and how much heat you can afford to dump into the satellite),
Well, given that we will continuously spend over 3.2 watt just for heating, I think we can afford running 10 Watt during the minute-two while the sat is in range.

(12-22-2014 01:18 AM)Bunsen Wrote:  Also, the LNA's ability to help will be limited by local ground-level background noise, unless the ground station is in the absolute middle of nowhere.

Which, I believe, should be our plan to begin with. I was thinking Canada one station, and rural Poland the other.

(12-22-2014 01:18 AM)Bunsen Wrote:  I'd estimate that 10-50 kbps is plausible, depending on modulation (the traditional ham radio systems are pretty inefficient, but fancier systems are less broadly compatible) and how big a ground station you want to require.

The best would be if you would put your numbers in the sheet, because, well, I don't have any significant experience with radio.
12-22-2014 05:03 AM
Find all posts by this user Quote this message in a reply
Newt Offline
Flying Buttress
***

Posts: 134
Joined: Dec 2014
Post: #7
RE: That Miraculous Device: Onboard Radio
Would 433Mhz really be practical? With a half wave dipole antenna I am getting near 70cm length.
12-22-2014 06:43 AM
Find all posts by this user Quote this message in a reply
Bunsen Offline
Junior Member
**

Posts: 18
Joined: Dec 2014
Post: #8
RE: That Miraculous Device: Onboard Radio
(12-22-2014 05:03 AM)MBobrik Wrote:  Well, given that we will continuously spend over 3.2 watt just for heating, I think we can afford running 10 Watt during the minute-two while the sat is in range.

10W of electricity doesn't produce 10W of RF out the antenna, though. With most transmitters, that will get you about 4-6W of RF, with the balance showing up as heat in your transmitter circuitry (I've got questions about that 3.2W heater plan, but that's for another thread). If you go heavy on the battery pack (which I doubt there will be space for in a 1U), you might be able to tolerate the power draw, but you still have to do something with that extra heat. And don't count on it happening in two-minute bursts.

Part of the reason for being a little conservative with the power/bandwidth balance is to make sure there will be long(ish) communication windows. If your link only works when the satellite is 30° above the horizon, you'll get windows on the order of 3 minutes long, and probably two of those per day for a given ground station. If it works down to 10° elevation, you get about 6 windows of 5-7 minutes per day. A 5° horizon, that goes up to maybe 7 per day and 6-10 minutes each.

That's important for more than just downlink capacity -- long periods of availability allow time to troubleshoot and acquire a link under less-than-perfect conditions. Short windows are very unforgiving; troubleshooting becomes kinda like trying to drive by strobe light, and getting a small problem ironed out can eat up much or all of a short mission.

With a 2W transmitter, more-or-less plausible guesses for other parameters (including a ground station built by a ham with money, i.e. the big 18.5dBic yagi and a good LNA), I'm finding an off-the-shelf 9600 baud amateur packet radio link (i.e. G3RUH FSK mode) having a +9.6dB margin at 5° elevation. That's pretty decent.

A good BPSK receiver should manage 38.4 kbaud with the same margin, but I don't know of an off-the-shelf source for such receivers or transmitters. The performance numbers I'm seeing for things like HopeRF modules suggest that they perform on par with the "Non-Coherent FSK" modulation described in the sheet, and therefore will get you about twice the speed per watt as the traditional ham system.

(12-22-2014 06:43 AM)Newt Wrote:  Would 433Mhz really be practical? With a half wave dipole antenna I am getting near 70cm length.

That's a full wavelength. The usual approach here is two quarter-wave strips of steel tape measure wrapped around the satellite and released by a resistor melting a nylon (or polyester, or Spectra, or what-have-you) line. It works pretty well as long as you don't overcomplicate the release mechanism.
(This post was last modified: 12-22-2014 07:36 AM by Bunsen.)
12-22-2014 07:30 AM
Find all posts by this user Quote this message in a reply
MBobrik Offline
Member
***

Posts: 203
Joined: Dec 2014
Post: #9
RE: That Miraculous Device: Onboard Radio
Well, as I said, I don't actually have any experience with RF. If you say that the transmitter amplifier got at most 60 % efficiency, then we have to go with that number.

However 9600 bps is grossly insufficient for our purposes, and even longer transition windows than I assumed don't make up for that.
Each sample produces cca 120 kilobit of stream each timelapse period, and we will have multiples of 3 of them on board. And we may have two cameras.

So we simply have to go for better solutions, like the one you mentioned :

(12-22-2014 07:30 AM)Bunsen Wrote:  A good BPSK receiver should manage 38.4 kbaud with the same margin, but I don't know of an off-the-shelf source for such receivers or transmitters. The performance numbers I'm seeing for things like HopeRF modules suggest that they perform on par with the "Non-Coherent FSK" modulation described in the sheet, and therefore will get you about twice the speed per watt as the traditional ham system.
(This post was last modified: 12-22-2014 08:53 AM by MBobrik.)
12-22-2014 08:51 AM
Find all posts by this user Quote this message in a reply
Newt Offline
Flying Buttress
***

Posts: 134
Joined: Dec 2014
Post: #10
RE: That Miraculous Device: Onboard Radio
As for the wavelength, even a quarter wave at 433Mhz is 17cm. Our longest side is 14cm. Is there a good reason to go with 433Mhz as opposed to say, 750Mhz, where a quarter wave equals 10cm?

Where are we getting these figures for communications windows? The ISS's orbit seems like roughly what we are guessing the sat's orbit to be like, but six windows, seven to ten minutes long, per day seems an unfounded assumption to me. Even with an ideally placed ground station, comm windows will be of varying lengths, ten minutes being the maximum of a spectrum that will range from zero to that figure. Based on this webpage, we maybe can expect three passes per day (at ten degrees over the horizon), depending on locations. Of course, if we have more stations, we have more windows, maybe that was what you were meaning.

Could we compress images? Maybe send down half sized images, or compressed JPEG images first, and then request the ones that look most worthwhile, and more if we have more time, or if a particular sample starts doing something interesting later and we want to trace its history. Sending all the images would be best, but might be a bit redundant, and slow.

I brought up electrical inefficiency before and it seemed a bit dismissed, but I think that we will need to carefully look into how much heat will be generated by all the systems and how fast it might be radiated into space. It might be the case that all our radios and other electronics heat the thing up plenty, or more than that. But that probably belongs in a different thread too. For now I think 60% sounds consistent with what I have come across,and am happy to go with it before we get better information about exactly how the radio will work.
12-22-2014 09:38 AM
Find all posts by this user Quote this message in a reply
Post Reply 


Forum Jump:


User(s) browsing this thread: 1 Guest(s)