Communications framework | Communications Workgroup | Forum

Post edited 3:42 am – January 13, 2010 by Luke Maurits

Wikipedia article on HDV gives a "Compressed video bitstream rate"  of ~25 Mbit/s for HDV 1080i and an "audio data rate" of "Stereo (2-channel) at 384 kbit/s (192 kbit/s per channel)".  Is that the kind of info you need?

Video and voice will be by far the biggest requirements, telemetry bursts will be less than one kilobyte and will happen at most every 5 seconds or so.  Control channel commands will only be a few bytes as well.

EDIT: Just realised that HDV isn't hi def video per se, it's a standard for storing hi def video on magnetic tape.  Not sure if that makes the data rate higher or lower but it's at least a rough ballpark.

squid said:

1080i is probably much more than we need. trying to get 25mbit over such a long distance is going to be hard, especially since we need to be power conscious. we'd need to either sacrifice frame rate or picture quality to bring the data rate down

Fair enough, I have no problem with lowering our video standards, I didn't know what a realistic data rate to aim for would be.  Would 10 mbit be reasonable?  5 mbit?  1 mbit?

We would probably want to decrease frame rate rather than picture quality, at least for the long and boring part of the trip.

Ah, good job noticing that!

I see they have considered spacecraft TX powers up to 25 watts max.  I wonder if they chose that limit because they have a small power system to save weight (since they are just sending a small unmanned rover they have to worry about weight much more than us – we have about 10,000 kg to work with).  The fuel cells we are considering can put out 50 W continuously, although of course that will not be purely for comms it will be shared with computers and other things.  We could always run a bank of two cells in parallel, though, to get 100 W if we needed to.  The fuel cells are only 1.6 kg, the difference between two or three redundant sets of one cell and two or three redundant sets of two cells would be under 5kg, we can almost certainly spare it.  It looks like the bit rate scales linearly with TX power so a 50 W transmitter could get 20 mbit, I guess.  Mind you, those values are with large antennas Earth side (30 m), I don't know if we'll be able to manage that.

It looks like we may have to resort to transmitting lower than HD quality video and sound during the trip to lower our data rate and just storing all the HD footage on hard drives during the trip.  A pile of hard drives will use a lot less power than even a 25 W TX.  This isn't really a big deal.

This is excellent stuff to know.

Team Frednet's wiki has a page called Frequency Allocations with a table of "Frequency allocations to Space Operations (SO), Space Research (SR), Deep Space (DS), and Earth Exploration (EES)".  As a reference it cites "ECSS-E-ST-50-05C".  It turns out ECSS is the European Centre for Space Standardisation, so I assume that the frequencies in that table are okay to use for space comms in most of Europe.  That actual standard can be downloaded here.

The SO, SR, DS and EES definitions are given in that document and I think we most likely fall under "Space Operations", which is limited to the "S band", using 2025 – 2110 MHz for Earth-to-space and 2200 – 2290 MHz for space-to-Earth.

The only frequencies within the Earth-to-space range mentioned in the Australian license are "(e) 2000 to 2009 MHz; or (f) 2009 to 2010 MHz".  A very narrow range.  I presume this refers only to the carrier signal?  The actual modualated signal would be unlikely to stay within a 1 MHz range, right?  Also the only frequencies within the space-to-Earth range mentioned in the Australian license are "(l) 2193 to 2200 MHz" (of which only 2200 MHz is mentioned in the European standard).  There's very little overlap here.  Of course, the ECSS thing is a standard for space operations, not the actual law.  It may be legal to use other frequencies, just not standard.

At any rate, these figures should give you a sense of the rough range we are talking about – somewhere in the low 2000 MHz range.  Hopefully this is enough for you to make some rough calculations on.

FYI, that ECSS standard also talks about modulation a lot, might be an interesting read.

Quick question:  The near-far problem, which is apparently quite a problem for CDMA, won't really be an issue for us if we're the only people using our particular frequency range, right?

Argh, I completely forgot about something important: using the comms system for navigation!

There are two parts to this:

• Estimating distance: this is simply a matter of timing how long it takes a signal to get from the ship to Earth or vice versa and using the constant speed of light in a vacuum to estimate distance.  Normally this is done by having one end send a "ping" signal to the other and that end being programmed to immediately respond with a "pong".  If the time it takes the computer to generate the "pong" is known (tested in advance in a lab) then it can be subtracted from the time between ping and pong.  It seems a bit wasteful to dedicate an entire virtual channel to this.  If there were synchronised clocks on the ship and on Earth couldn't the ship just timestamp its telemetry packets and from this and the time of receipt on Earth we could get the one way travel distance?
• Estimating velocity: this is done based on Doppler shift.  I have no idea what's actually involved in this, how do we handle this issue?  I guess the receivers at both ends need to be able to dynamically shift frequecy as the spacecraft's speed changes?  Any system to do this should be able to give us the shifted frequency, which is all we would need, right?

Also, error correction.  We should definitely, due to latency, use Forward Error Correction codes, which avoid resends by embedding redundancy in the signal.  This will lower our data rate somewhat but it really can't be avoided.  I actually know a little about these (very nice maths behind them!) but not enough to be able to contribute much to the conversation.  I may be able to read up on it, though, I always did want to know more about them.

squid said:

yes i am at adelaide university. unfortunately i was caught up with some other things yesterday so didn't get much time to look at this.

might be easier for me just to come and in talk with you directly about it; i can be in wednesday afternoon and/or thursday if you are free then.

Thursday afternoon would definitely work better for me.  I'll be in from 12 onward but I won't have a lot of flexibility with regards to meeting place.  I need to stay in a lab the whole time because I have people coming in to participate in an experiment at intermittent times – but the lab is quite large, with a few couches, and I'm the only one in it, so it wouldn't be a problem for us to hang out in there and discuss CSTART stuff.  There's a whiteboard we could use and everything.  Running people through the experiment only takes about 5 minutes before and after so it shouldn't be too much of a hassle.  If this is cool with you I can PM you the building/room number.