What is ManT1S and T1S?¶
A basic topology for a ManT1S-based distributed system is shown in this diagram:
A few things can be gleaned from this. First, the boards are connected together with a single twisted pair of wires. That's where the term "Single Pair Ethernet" (SPE) comes from. In this example, four boards are in the system, so there are three twisted pair segments connecting the boards (black and white in the diagram). A collection of T1S interfaces connected together like this is called a "mixing segment", in case you've seen that term and wondered what it was. Up to 8 boards can be networked together like this on a mixing segment according to the IEEE802.3cg-2019 standard — no router needed! While in my diagram I show short connections between nodes, the standard allows 15 m between nodes and 25 m for the complete mixing segment!
10BASE-T1S allows for 10 cm stubs teeing off of a network pair to connect to a node, and while you can do that to connect to a ManT1S, that's such a short stub distance that I decided it was much more practical to just provide the ability to easily daisy chain boards. That's why the screw terminal has four positions: it allows you to connect two wire pairs for daisy chaining. Boards at the end of a chain will only connect using two of the four screw terminal positions, while those in the middle of the chain use all four. You can think of it as an incoming and an outgoing pair, but that could give the wrong impression as if there was a "direction" to these. There isn't: all boards can equally talk to each other up and down the chain and power can be distributed from anywhere to everywhere. There are really only two connections on the board, they are just duplicated to make your wiring job easier. The picture below shows the marking near the screw terminal: there are two "+" connections and two "-" connections. Each pair needs to connect to a "+" and a "-".
These markings indicate the "networking" positive and negative members of the pair. You need to make sure your wiring consistently connects the "+" terminals of all boards together and the "-" terminals of all boards together respectively, so the boards can properly talk to each other.
When it comes to power distribution, I have decided it's best to follow the same polarity convention as for networking, to avoid confusion. While the ManT1S has a diode bridge on board and in principle doesn't care which polarity the distributed power uses, the polarity does matter for the ManT1S-Bridge. So keeping it consistent is a good idea in case you want to use the ManT1S-Bridge. Much more on that in a future update.
The diagram shows power for all boards injected at the bottom left board using the red and black wire there. These connect to the "PDL+" and "PDL-" pins you can see in the picture. This is the updated naming used on revision 5 boards. You may see these labeled "PDL1" and "PDL2" on older revisions in some of the campaign pictures, but I changed that to "PDL+" and "PDL-" in pursuit of the consistency mentioned above. All hardware shipped to backers will use this updated naming convention.
There's nothing special about the node where power is injected in the diagram, it's just an example. Power could be injected at any of the nodes, but to avoid problems, you should only inject power for distribution at one node of the mixing segment. If you just want to provide power to a single node without distributing it to other nodes, you can do that using the "V+" and "GND" pins instead. In a future update I will do a deep dive on the various ways you can power the system.
When a system is wired together as shown in the diagram, they are all networked and can comminicate with each other using standard TCP/IP network protocols. And they can all get power over the same two wires, injected at the most convenient node of the mixing segment. Distributed systems made simple!