[aprssig] APRS low-power-local ALT input channel
wes at johnston.net
Mon Sep 27 10:38:41 CDT 2004
I did some stats with 4 home stations (10 minutes each) and 4 trackers (1 minute each) and came up with
4 trackers x 60 packets/hour each= 240 packets/hour, or 240 seconds
4 homes x 6 packets/hour each = 24 packets/hour or 24 seconds.
So local users end up TRANSMITTING for 264 seconds out of 3600 seconds per hour, 7.3% utilization on 144.99 if everyone in my local area ran 144.39+600. This number is a little higher than Bob's. But it's in the ballpark.
Pete, you are right about packet being bursty, but as I have watched my kenwood radio's track, I see that packets do not come out at exactly 60 second intervals. Other stations being heard cause my radio to hold off TX if, when it is time for it to TX, another station is being heard. Since my car hears different stations than another local's car does, sometimes my car waits 60 seconds to tx, sometimes it waits 60-70seconds. My neighbor's car does the same. And we gradually ratchet towards and away from one another's time slots. The fact that my car hears different stations (and assumes the channel is busy) serves to randomize our TX slots. Now if we ran our TNCs in full duplex (doing away with carrier detection as a collision avoidance measure), then my car would tx every 60 seconds no matter what and if two stations ever transmitted at the same time, they would _always_ transmit at the same time. So in the case of an alt input digi, we are using the packet signals on 144.39 and carrier detection to randomize the access on 144.99.
You are also right about specifying how many we can DECODE direct. Just as I specified above that my car can hear dx signals and those signals (decodeable or not) serve to randomize my TX time, simply being able to have the CD light on the TNC will serve as an effective means of CSMA. If a distant station's signal drops for 2mS, I've lost the packet... but there is no way my TNC would try to seize the channel in 2mS... slot time and persist wouldn't allow it. So yes, you are correct, even if I can't decode a packet I can still use it to make CSMA work.
>And all of those other stations (300 per hour according to your count at
>your QTH)? You are not competing with them for transmission on RF, only
>with their packets that are being digipeated by your near-by WIDEn-n
>which you do see.
This I have to take issue with. My mobile drives past a tree or signpost causing the digi to hear my signal drop 10dB for a millisecond or two. Now if there were no other stations (or dx digipeaters) that the local digi could hear, then the FM receiver would track the signal strength fluxuations and the TNC at the digi would continue decoding me. But there are other digi's in the distance... and when my signal picket fences, the local digi looses me in the noise floor. It sounds like you are basing all your ideas on a home station, I dunno.
>Finally, let's look at your digipeater design as implemented at a
>WIDEn-n site. Every time the WIDEn-n side transmits, your local
>receiver is desensed and cannot copy anything on the "local" side.
Right again.... We either need vertical separation of the 144.99 and .39 antennas, or to run cavities. Or we need to not implement a full sevice 144.39 digipeater at the alt input digi site. The only reason we need to TX on 144.39 (if we have no cavities) is to forward a packet we just heard to the main 144.39 digipeater. In this case, our 144.99 digipeater is unavailable for RX while it is TX'ing the packet it just heard. This means we have to double my 7.3% number above to 14%.
Stephen Smith said that this alt input digi idea was no different that installing MIC-e on a voice repeater. And we are setting up a voice repeater here in our town for mic-e, and we are planning on using 70cm to forward the packets to 144.39 digipeater. Now if we can forward packets to the digi from the alt input site in such a way that it does not desense the alt input's receiver, we have a winner. If we end up with multiple alt input digipeaters in a metro area, we need to take care that they can all hear each other. This is getting convoluted though.... I fear that by trying to cover all the bases and possibilites I am muddying the water.
>Also, as pointed out before, the PocketTracker
>(sorry Scott for misspelling, if it is) only supports 144.39 and 144.34.
>You have excluded them from your configuration with promoting 144.99 for
The maker of the pocket track has said that it will use 144.39 and (144.99 or 144.34). It appears he selected 144.34 at the request of a couple of his first customers and because it was close in prox to 144.39. He has offered to sell a pocket tracker on any freq you ask for... so this is really a non-issue.
Pete, I will also conceed that you are right about the correct fix is to get people to run shorter paths. That solves the problem... but I (and others) have talked until we are blue in the face and people still run crazy long paths out of ignorance or out of spite (as one particular station near Statesville NC does ;-) ). There is a point when you have to trump these unchangeable humans with the network itself.
What does it hurt for someone to setup a alt input digi? As long as it listens before it digipeats on 144.39, it does not interfere with anyone on 144.39. The same number of stations will be in the same areas, so there is not going to be any more traffic on APRS. We can sit here and spiral ourselves into a tizzy hashing out who's numbers are right and pick nits till the cows come home. In the end, if I QSY my moble to 144.99 and transit to a digi which then relays that packet back to 144.39 for me, what if the difference to the stations on 144.39? They still see one copy of my packet on 39. Maybe, though, they see packets from my mobile with more regularity b/c I won't be competing with stations I can't see on 39, and therefore more of my packets get thru. The only problem would be that if some other group or system was on 144.99 and I interfered with them. But Bob has said that 144.99 +/- needs to be coordinated locally... that way it can be declared as a part of the aprs system.... ( and no, I'm not trying to say it's _my_ frequency)
I suggested over a year ago an idea like this where I tied the PTT line of the TNC to the mic-up input on my 2m radio. setting up multiple alt input digi's on different frequencies around town and programming those freq's into my radio's memory in some random order would also serve to randomize. But yes, I'd still be transmitting in the blind on any of those frequencies.... but that would be OK if I had coordinated them as part of this SYSTEM.
And finally, if CSMA works so well, why did they switch to DAMA in europe?
More information about the aprssig