APRS ALT-PRIORITY Input Channel for LOCAL stations 10 Jul 2004 ------------------------------------------------------------------------- Updated 19 Oct 2005 Alternate frequency input digipeaters is a great and easy improvement for local APRS reliability, AND it can be done independently of what any other digipeater owners do. It is simply the additon of an alternate locally coordinated APRS input frequency to give local stations priority over the local 144.39 channel. Packets heard on this very quiet low dutycycle channel are then digipeated in occasional short bursts over on 144.39. Advantages: - Drastically reduces collisions for locals (improves reliability) - Is a local issue that can be implemented locally and independently - Can use any existing TNC' with NO software/firmware upgrades to digis. - Or, with software can do WX conversion to standard WX format for Kenwoods Implementation is simple: 1) Existing digi hardware on 144.39 is unchanged 2) Add a new TNC and Receiver on the local alternate input (144.99 if avail)which then digipeats those packets over to 144.39 Additional options: 3) Connect the original TNC PTT, and DCD lines to new TNC external CD line. Connect its PTT to the original TNC's XCD. 4) Connect the new TNC TX audio and PTT to the 144.39 TX This way the alternate RX channel TNC waits for 144.39 to clear, before it transmits. The cross connected PTT's and DCD's to the external CD lines assure they will not collide when both trying to use the XMTR. Though for only a few low powere trackers, none of this is necessary. The TNC settings are: MYCALL: 99ANAP The first two bytes are the Freq (144.99 here) and then some local abbreviation (Annapolis here) because we want all these TNCs to have a unique call so that they are distinct on FINDU. BText: Standard DIGI position report. BEACON: Every 10 minutes direct. (it only serves direct users!) UNPROTO: APALT (direct to the APALT TO-CALL with no VIA path) XCD: External CD enabled (if used) FIXED AND WX STATION SETTINGS: The users of this back door alt freq simply include the callsign of this digi in their outgoing path. For example: VIA 99ANAP,WIDE2-1. Note, with home stations designating a specific digi, the path is traceable even if this is a dumb TNC with no callsign substitution possible. Since there are so few locals compared to the 95% of other distant stations that would hit this digi, the probability of a collision is drastically reduced. MOBILES and LOW POWER TRACKERS: This alternate local channel also drastically improves the probability of a mobile or low power tracker being heard. Being on this input-only channel, they only have to complete with the very rare local packets, instead of the other 95% of QRM packets from out of area. When their packet is digipeated over from the alt-channel to the 144.39, their local packet reliability has improved by an order of magnitude or more. SOFTWARE IMPROVEMENT #1. For an additional benefit to Mobile Kenwoods that cannot currently decode the RAW stand-alone Peet Brothers WX packets, a nice level of improvemnt can be gained with a processor that collects all the non-standard APRS WX reports and weatherless position and positionless weather, and converts them all to the standard COMPLETE format before dumping them on 144.39. Also, bundeling then in groups of 5 makes for a nice neat channel burst of weather and one full screen display on the Kenwoods LIST. Nothing is lost in thie bundling because WX is only flashed on the Kenwood display raw anyway when first received and the user has always had to "select" a WX station on his radio front panel to see the actual weather anyway. But going to all the trouble of a PC at the site instead of a dumb TNC might make this option simply not worth it. SOFTWARE IMPROVEMENT #2: The software can also implement the OBJECT bundling included in APRS1.1, further improving channel efficiency. See the links on the APRS 1.1 SPEC. This added capability makes a PC more valuable at the digi. STAND-ALONE ALT-DIGI: Alternatively, in an area that is saturated with too many digis already, simply convert one HIGH digi that already hears too much over completely to this new method. One TNC, One radio. It just receives on the "fixed" local priority channel and TX's on 144.39. It can have the 555 bundler or not and it may or may not have a separate CSMA RX on 144.39. Reason being, if it is high enough, it probably hears wall-to-wall packets on 144.39 anyway, and the CSMA gives no real advantage. Think of this concept as a back-door into your local digi that is almost two orders of magnitude less vulnerable to collisions. Also, when the digi does re-transmit these local packets, they can be inserted seamlessly using CSMA into the existing 144.39 while fully honoring the collision avoidance of the digi which is well situated to hear all other potential signals on the channel. Thus everyone wins... HERE IS A CONDENSATION OF FOLLOW-ON DISCUSSION FROM THE APRSSIG: * If the frequency of 144.99 is available in your area, see if you can get it coordinated for this application. It then lets all users simply use the +600 KHz offset on their radios * ANY TNC can be used, since no smart digipeating is needed to get the packet from the alt channel to the 144.39 channel. The digi's actual call is used. If there is more than one in an area, use SSID's 99ANAP-2 etc... Even if there are two such inputs that are in RF range of each other, they doubtfully will hear the same ground level fixed users AND there are no dupes because the given site is specifically called out in the path by callsign. * Notice that EVERY ONE still listens to everything on 144.39 so nothing is lost and the channel operated normally even for visitors. * Since only fixed stations or local mobiles use this new alt-channel input, the locations of these new alt-channel inputs can be much more sparse than normal digis because with the clear channel, the locals packets do not have to overpower distant QRM as they do now. * Fixed and WX stations are transmitting in the blind on the alternate input channel but since there are only so many of these stations in the range of this alternate channel input, it has been shown that the chance of a collision even with 14 home stations and 6 WX stations all in simplex range is down around 3% or less. Listening first only changes this by 0.3%, because any one station cannot hear most of the other stations sharing the input anyway. This guaranteeed low collision rate without CSMA compared to the very high rate you otherwise would get on 144.39 even with CSMA is what makes this so attractive. * There really is not much advantage to dual CSMA on both channels because because the loading on the alternate channel locally is so low. The probability that you are colliding with a local packet is very low. And the purpose of this system is to give LOCAL priority! * Home stations serving as local WIDE1-1's are not affected. They still will hear mobiles on 144.39 and although they send the packet via the back-door on the alternate channel, the packet still gets to 144.39 as before. * By activating this second channel and maintaining a few digis there this does provide a backup channel for passing traffic during emergencies or special needs. Or even maintaining home station BBS's for end delivery (on-air BBS mail reading is not desired). Though you cannot use the alt-digis function for reading mail because they collide with 144.39 traffic. But the FREQUENCY is available for direct connects if needed. * The advantages of this alternate channel are because of its low usage. If a few local stations become active and begin heavily using the channel say with APRS messages, then some of the advnatages to others are lost. Users should consider QSYing to the main 144.39 in this case. Or conversly, let the single heavy user such as the EOC stay on the alt channel so that he DOES get priority CSMA service at the digi without other 144.39 competition... * Igates continue to function normally. But the best 24/7 IGate should clearly be on this alt-channel so that it gets the advantage of good CSMA at the digi for its packets.. Though as usual, good placement and minimization of duplicate IGates is good practice. ANALYSIS OF BENEFITS: --------------------- In our area, with a 120 Height above average terrain, there are no more than about 17 fixed home stations (3 are WX) and 3 stand-alone WX stations. These add to about 52 packets per 30 minutes or about 1 every 35 seconds for a probability of collision on the non-CSMA input channel of about 3%. The benefit is determined by how much better this probability is than if they transmitted on the 144.39 channel and it's much worse probabilities. -------------- de WB4APR, Bob