Promethean Guidance and telemetry systems

Team Discussions about Guidance and telemetry systems and construction.

Ok guy's
If we are going with the gumstix SBC then I'm getting ready to start configuring the computer to work with it. There are several RTOS options for linux so we need to figure out which one we are going to use so I can build a kernal and get started. I have an Xbuntu system here I can work with.

Monroe

what type of telemetry system are you using? my offer stands to help you with the error correction parts. unless you have access to and understanding of turbo codes i highly recommend using RS+viterbi FEC it'll be worth the hassle for your downlink power needed (only need ~4dB SNR for a BER<10^-6). admittedly the decoders for those are complicated and you might not be able to scrounge one, but as long as you have a few spare fpgas there should be open source solutions you can use.

Right now we are staring directly at Gumstix and the GPStix add on as flight computer and tracking.(near space platform) I want a live feed of HD video and remote control of different aspects of the mission as well as the normal cut down ect... At this moment I have no idea if the Gumstix can handle all that but the guy's like the gumstix so I'll go with it. I'm thinking a 900MHZ downlink at internet speeds to make it really interesting. The CPU will have to handle some stablization chores too. Think thats to much?

Monroe

I like this SBC http://emea.kontron.com/products/boards ... tx+25+sbc/
But if the gumstix's can handle all that im game!

I don't understand the systems you speak of that does not mean I'm not interested but your going to have to speak a little slower untill I can catch up

Sure we would love to have your help but we have to get on the same page first.

if you're using off-the-shelf telemetry equipment then it's not really relevant, if you're using an in house designed fpga or microcontroller system to make a custom data stream then what i'm talking about would be useful. forward error correction is how nasa makes it possible to communicate with its deep space spacecraft without spending massive amounts of power on their transmitters. basically you have this hardware that encodes a mathematical formula that makes the data have redundancy, and your ground-side hardware uses this redundancy to correct any bit errors introduced by the noisy transmission environment, greatly reducing the signal to noise ratio required to communicate with your spacecraft (theoretical net gain is on the order of ~3.3dB for the system i'm talking about). you have about 7dB of reduction in SNR and about 3.6 dB in increased bandwidth from the extra symbols. if you're using off-the-shelf this stuff is probably built in, especially for the video downlink, but if not you'll have to put it on or rely on a measly CRC for error detection, with no way of recovering the underlying data and hence needing more power.

TerraMrs,

We will implement some form of forward error correction. Reed-Solomon coding is nice because it is byte oriented. I have also implemented the BCH method that is defined in the H.261 video coding standard. BCH is inherently a bit-serial algorithm, but it can implemented with byte operations by using look-up tables. This is true of most bit-serial algorithms.

Dave
Team Prometheus

Great! now we are all on the same page! So what RTOS are we going to choose? So I can try and keep up I need to build a terminal and get started with that. I'm a little rusty on linux and I want to get a system I can work with going.

Monroe
Team Prometheus

Let's go!

if you don't already have an RS core i'll recommend the RSTK toolkit from sourceforge, or the open cores implementation (both conventional not berlekamp). both work fine, all you have to do is add the transform and inverse transform matricies and write an interleaver (which is easy and not required). i would also like to note that there are two ways of calculating the generator polynomial coefficients and the RSTK kit uses the method that isn't what's defined in the CCSDS standard (don't remember exact formula right now), so if you want to use that you'll have to manually change the polynomial coefficients (easy). on the other hand, if you're using the toolkit decoder maybe you won't even need the transforms... the open cores i believe is not a (255,223) E=16 by default but is easily modified. however, from my reading it seems that smaller code block sizes (E=4 or so) are better when you're not using concatenated codes.

TerraMrs
What is an RS core? I could possiably make sence of the rest If I knew what that was Thanks!

Monroe

Edit: I googled RSTK Toolkit Sourceforge and did not find anything how about a link? I'll see if I can seacrh Sourceforge

Edit2: Ok I found that: Reed-Soloman Core Compiler Great!

Off Topic: What RTOS are we going to use guy's?

cool. glad you found it. we're using it over here at loral for the project i'm working on (new ccsds compliant electronics) so you know it'll be good (though like i said requires some changes for CCSDS).

Have a look at this: www.cubesatkit.com wht do you guys think? Vax's Idea!

Monroe

I wonder what Salvo is? I'm having a look.