Re: [MGSN-discussion] TNC?
Status: Beta
Brought to you by:
jwcutler
|
From: James C. <jw...@st...> - 2003-10-24 17:44:58
|
> 1) What hardware does your TNC at Stanford consist of? > 2) Which specific tasks does the TNC perform? In the HAM world, the TNC (terminal node controller) traditionally performs (de)modulation, bit synchronization, frame sync, and packetization. TNC's tend to be fairly intelligent with small microprocessors embedded in them. They often have a specialized interface to allow you to configure their AX.25 packetization capabilties. Their is a group at the U of Wuerzburg that is also putting FEC (forward error correction into a TNC). We tend to operate our TNC's in KISS mode (KISS-Keep It Simple Stupid). TNC's in KISS mode do not perform many of the standard packetization functions. They pass the AX.25 packets on to the CPU that is connected to them so that the CPU can handle them. We connect our KISS TNC's to Linux boxes that have kernel support for AX.25. TNCs are dinosaurs and I expect them to be extinct soon. General purpose CPUs are replacing all of their functionality. For instance, a CPU running Linux with a sound card can perform all the functions of a standard TNC. We're combining these software radio/modem systems with virtual machines. Check out this paper: http://swig.stanford.edu/pub/publications/space/RCSGSOJCutlerMS.pdf. With this technology, we should be able to provide incredible multi mission support in ground stations. > 3) Is it sufficient to use the MX909 (GMSK Modem Data Pump), and a computer for > the AX25 coding? Yes it should. We'll need to figure out how to export an interface to your system through Mercury. We haven't labeled all the pipeline outputs yet. In our AX.25 world, we've been using a TNC export and KISS export. So, the default is that the GS exports KISS packets (with AX.25 packets encoded into them). Can we make your computer interface to the MX909 do the same thing? Other interfaces we'll be exporting include IP, IP over AX.25, etc. And actually, as we implement the virtual machines described above, the GS and Mercury will no longer have to worry about explicitly knowing about the interfaces. They will be handled by the VM. I hope this helps. Sorry for the late reply. We just finished building a ground station here and we're shipping it to Fairbanks, Alaska today! That should give us and the MGSN some good northern coverage. Can't wait to get your station up and running too! Also, I just received some CPU hardware and we'll be setting up some of the initial MGSN central servers here shortly. Hope all is well! Cheers! --Jamie |