KD0TLS: When packets collide...

  APRS, and packet operation in general, suffers from a phenomenon called "packet collisions" on RF. Essentially, this is a situation in which two stations transmit simultaneously  on the same frequency. Due to the FM "capture effect", only one will dominate, but that "dominant" packet is very likely to be rendered unreadable. Nearly all TNCs, modems, and trackers have some means of avoiding this by checking the frequency for activity before transmitting. In some cases, this is known as DCD (Digital Carrier Detection). If you are operating on RF with the squelch turned off (as is usually the case), you should take the time to ensure that this function is engaged. 

 While this is generally sufficient, it only works for stations that your receiver can actually hear within simplex range. Given our robust metro infrastructure, it's entirely possible for two stations on opposite sides of the metro to fail to hear the other. And it's also very possible for large gateways, which hear stations over a wide area, to pick up two stations transmitting at once with each outside of simplex range of the other. And it's not uncommon for big digipeaters, which can transmit across half the State, to fail to hear traffic from smaller stations in that coverage area. This is, in fact, one of the main purposes of digipeaters -- to extend coverage between stations that otherwise would not be able to maintain simplex communication. 

 This phenomenon is why "mega-stations", such as W0PZT-1 and W0YC-5, can't bear the load of covering the metro on their own. Below is a map of the current coverage of the Medina gateway/digipeater W0PZT-1. You can click on it to enlarge the image, if you wish. 

Direct coverage of W0PZT-1 as of 11/13/21

 As you can see, this single gateway hears stations over virtually all of the metro. You may easily conclude that W0PZT-1 makes all of the other metro gateways, and even metro digipeaters, redundant. This, however, would be false. 

 Both Forest Lake and Bloomington are in the direct reception of W0PZT-1, and it's not uncommon for mobile stations to run low power (like 5W). These mobile stations can rely on the digipeater network to make their operation viable. But in my experience, gateways like W0PZT-1 are quite capable of hearing these low-power stations on their own in these areas. So, what happens when two stations, each out of simplex range of the other, transmit their packets? Neither can hear the other, so they both believe the frequency is clear. But W0PZT-1 can hear both. A packet collision is what happens. Now consider all of the digipeaters in that coverage area IDing and re-broadcasting distant traffic. Each of these stations is trying to be considerate and only transmitting when the (local) frequency is clear, but they can't hear what they can't hear. 

 To a certain extent, these gateways attempt to hold on to packets that they can't directly gate, and instead digipeat them. But that only goes for packets that aren't rendered unreadable from packet collisions. The end result is that these "mega-stations" drop packets. 

 In a practical sense, these "mega-stations" (e.g. W0PZT-1 and W0YC-5) only provide baseline coverage for the metro. It's up to smaller gateways and digipeaters to pick up the traffic that the "mega-stations" miss due to their wide coverage areas. That isn't a flaw; it's the nature of a network. No single element does it all. Instead, they work together to meet the objective of providing reliable, redundant coverage to amateur operators.

 An operator may set up a modest RX-only gateway in their home to provide a service to the local amateur community, but they become disappointed when they look at their station's statistics and see that they "only" gated 20 directly-received packets over the course of a month. That's not a failure on the part of the gateway operator; it's how the network functions. That's 20 packets that otherwise would have been dropped. And that modest gateway is very probably picking up digipeated packets from far outside the metro that our overwhelmed "mega-stations" missed. 

 As one moves outside of the RF-congested metro, packet collision obviously becomes less of a concern. But it can still happen in the case of wide-coverage digipeaters, such as KD0JOU-2 near Litchfield. That modest RX-only gateway suddenly becomes the sole source of direct coverage for the local area, providing message services and propagation data to the wider amateur community. Because it's a collective effort.