AX.25 Layer 2 has been conceived as an ARQ link. (A)cknowledge (R)e(Q)uest does not exclude sharing the channel or not.
In spanish, amateur is written as "aficionado". Radioaficionado does not have, in spanish, the slightest derogatory hint that the early "ham operator" english denomination had about the early radio amateur operators, sometimes more enthusiastic than knowledgeable. And not all amateur radio operators are equally educated, we are certainly lucky to have quite a few PhD's in our ranks, quite knowledgeable in software, hardware and math. They are as much "aficionados" as anyone who takes a hobby seriously. To name just one example, the TAPR-93 DSP TNC, which did not become popular nor cheap. Of course, there is a scale effect, short production runs are hardly able to become cheap. The best term I have seen coined for evaluating a digital radio link is "spectral footprint", which takes into consideration occupied bandwidth, period of channel occupation and net traffic moved. Pactor II is an ARQ mode that also uses FEC, block coding and interleaving. Its bandwidth may be smaller than packet, especially those hard keyed FSK modems like the XR2206, on which frequency switching transients may have the lion's share on bandwidth occupation. A far better solution was the old KAM hardware modem, I built almost a copy myself as part of a homebrew RTTY/packet terminal unit, using a CMOS Johnson counter, which actually did work very well with BPQAX25. It did CPFSK and worked very well on the air. I had an old marine radio crystal which could get 1200, 2000 and 2200 Hz tones from, and I hard wired my counters to switch with the RS-232 input signals. Yes, I am a hardware guy. Q15X25 belongs to layer 1, where modulation and channel coding is done. It uses BCH code FEC, the same thing I am advocating, and is about as wide (2000 Hz) as Pactor III. Q15X25 uses AX.25 as its layer 2, as TNC's approved by the FCC do. I thought that Q15X25 would be a good substitute for the old Bell 103 modem. I sadly see it did not. Tomi, OH2BNS had little support with it, working alone mostly, as far as I know, and when he stopped working on it, there still were some unresolved issues. For me, it was brilliant to use 15 tones are separated by 125 Hz, each QPSK modulated at a symbol rate of 83.33 bauds to provide a raw data rate of 2500 bit/s in an occupied bandwidth of 2000 Hz. The 83 baud rate seems to be a good match to the average HF radio channel with multipath. But I still wonder, why did it not match Pactor III? In thruput, compression makes a difference, but what about the rest? Some people were lucky to convince Kantronics to include it on one of their latest TNC's, but I really don't know how succesful it was. GTOR may have been a past good alternative. Only some Kantronics TNC's carried it, so it did not become popular. AEA had already mimicked Peter's cry about the wolf coming (the old russian Peter and the wolf tale) and filed bankrupcy, acussing the Internet of the demise of HF packet. I am afraid it actually did have some "domino effect" on HF packet. Why do I believe the BBS model is OK? I used my PTC-II on pactor for a long time with FBB7 and it works as well as if it had been working with my old KPC-2. Maiko Langelaar, VE4KLM has written the code to use the PTC-II with JNOS, and quite likely, TCPIP. I have not tried it so far. You state that AX.25 would only be in use on VHF/UHF, and I see no need to corner it there, just a change of modem and channel coding is needed, that's why I have been willing to attempt to convince some software author to provide some packet standard packet interface to a more advanced sound card modem, be it kiss, or host mode, or 6PACK, so popular in Germany. That might allow to use BBS software with sound card modes. Or any other standard, if some software can be developed to convert from old standard interfaces to a new one. And even on VHF/UHF something could be gained with a better modem and a SSB radio using something like RFSM2400 or RFSM8000 as a layer 1 device instead of the old Bell 202 modem. It could be an SDR broader than 3 kHz, it is allowable on 6 meters, on 2 meters, and higher. Of course, up there, nowadays, it is hard to beat WLAN's. I did use Flexnet on DOS for some time, and soundmodem on Linux for a longer period, and it is astonishing how many modulation schemes you can run using standard AX.25 Level 2.0 On Windows, if we could use PAX, or any other more advanced layer 1 software as modem, if we could use 110 or 200 baud, or the FAE400 modem, to name a few, we might see a better solution than the old Bell 103 modem. It was meant for line, not HF radio, so it is quite surprising that nothing better evolved in 26 years to become a popular substitute. I have not tested robust packet, as I have no HF DSP modem or enough RAM in my PTC-II. If it uses the same bag of tricks as Pactor-II and III do, it could be another suitable substitute. It is made by SCS, yes, and requires a PTC to plug it in. Period. Nowadays I cannot accept anymore that a raw 300 baud modem is the solution, and many quitters, sadly, seem to corroborate this lately. Convolutional coding is used not only by space communications, but by Pactor II and III, and by WSPR, and it _*does work well on HF*_ Block codes and interleaving are used by many professional protocols, like DRM, DVB-T, DVB-S, DMBT, and actually, not in vain. Enhanced ATSC uses two layers of Reed-Solomon coding, so, why leave those signs pass as unseen ? The best seem to be LDPC and turbo codes, but there might be some patent issues with them, of which I am not sure right now. So, I was happy using Pactor-II to do BBS forwarding in HF, of course, on different frequencies to the mainstream forwarding ones, and it did work well for the links to the US, Europe and Africa, sometimes with as low power as 25 watts (PEP) on a dipole or a vertical antenna for twenty meters. It was the alternative for me, and I am glad I used it. It meant a 10:1 spectral footprint improvement. Something else is the mess the forwarding routes have become, but that is a different topic. 73, Jose, CO2JA --- Charles Brabham wrote: > PACTOR, being an ARQ mode is incapable of sharing a frequency with > more than one other station. That, along with the extreme bandwidth > and lack of effective signal detection makes PACTOR unsuitable for > digital HF networks on anything but a very limited scale. - A few > afficianados can play around with it, but in that case as the network > grows, more and more participants cop out and use the internet > band-aid to cover up for the mode's basic lack of suitability for HF > networking. > > Or they do like WinLink and run roughshod over their fellow hams, > operating what amounts to a QRM mill that takes up more and more > spectrum as the "network" grows. > > HF Packet, warts and all, is currently the only digital mode that a > serious HF network can be built upon. The secret to this performance > edge is AX25, which allows multiple stations to share a single > frequency. The more reasonable bandwith there is also a positive > factor that appeals to responsible amateurs who know how to play well > with others. > > They call this "spectral efficiency" and if your mode of choice does > not have it, best to keep it for keyboard use and leave the > networking to the networkers. > > It is fashionable to diss Packet radio and AX25 - but none of the > detractors have been able to demonstrate anything that does HF > Packet's job any better... In fact, nobody has come up with anything > yet that even works as well. Performance talks, and "fashionable PC > attitudes" walk when actual networkers look at the available digital > modes. > > That's the way it is... Maybe someday there will be an actual > improvement over AX25 and Packet for HF networking. When this > happens, I'll be one of the first to put the new system on the air > and into actual use. BUT I have witnessed and been part of several > efforts to improve upon AX25 and Packet over the last couple of > decades, and what has been found in every case so far is that it is > awfully easy to sit around and diss AX25 Packet for HF networking, > but not so easy to come up with something that actually works as > well, much less any better. > > If there was anything actually better out there, the HF digital > network would already be using it and AX25 Packet would only be found > on the VHF/UHF bands. > > But there isn't, so... > > 73 DE Charles Brabham, N5PVL > > [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]> > > > > ----- Original Message ----- > *From:* Jose A. Amador > <mailto:[EMAIL PROTECTED]> > *To:* digitalradio@yahoogroups.com <mailto:digitalradio@yahoogroups.com> > *Sent:* Sunday, August 03, 2008 9:16 PM *Subject:* Re: [digitalradio] > Re: Has anyone looked into FPGA-based digitalmodes? > > I believe that both the AX.25 and the BBS model are OK, but that the > packet channel coding is a disaster in the sense that a single > erroneous bit trashes a frame. That fires up the retries chain that > are so detrimental to the link capacity, and may sever it as well. > > Pactor does a _LOT_ better, as it is able to use frames with errors > that would be useless on packet using different FEC mechanisms. > Source compression may help as well, as FBB and WL2K do. If the > signalling speed can be made to match the channel and the protocol > yield capabilities under a certain level of errors, a huge relative > improvement can be achieved. > > That is the big adventage of WL2K, the use of Pactor II and its > better channel coding. The rest is much alike the old BBS system, > reworked. > > I believe that something that achieves similar results to those > stated above will certainly be a step ahead. > > 73, > > Jose, CO2JA
