In order to use the EDT, you need something to send it data such as a PC, the memory card would be for the Conductor but I believe that device is now a thing of the past.
Ah, OK. Well, no biggie. I was intending on using a PC (Raspberry Pi 2) in any case.
As for the hubs, you also need to take into account how many channels you are using. Each hub is locked into one PixelNet universe so if your channel count exceeds 4096 channels (assuming of course that you are starting at the first channel of the universe) you will need another hub for the carryover into the next universe.
Ah, right. The Zeus, too?
For me, the point is moot at the moment. I'm probably looking at 1275 channels starting at address 97, but it is certainly something to keep in mind.
I don't know what you mean by in-situ but with standard firmware, the SSC config is set by the utility on the bench with the jumper on the SSC set to program mode.
By "in-situ", I mean wherever the SSC is sitting connected to a hub. So one needs to use a PIC programmer directly connected to the SSC (or in my case, the Zeus), or does the utility run on a host connected to whatever is transmitting Pixelnet (in this case, the Etherdongle)?
Sending this much data over wireless in usually not a very good idea.
Tell me about it. I'm a network engineer, and I avoid wireless as much as humanly possible. That said, I'm using wireless transcievers on all four Lynx Expresses, because I have to unless I want to run multiple conduits under my walkways and direvway. It works very well. I was more or less hoping the Lynx Transciever could handle PixelNet.
Since you use FPP and the PI now, you would be better off just getting another PI and running FPP in multisync mode, making the PI at the megatree the slave. Then you could use WIFI to link to the remote PI and since it is only sync packets being transferred, traffic is minimal.
Well, that's an idea - possibly a really good one. I only see two problems:
1. As I add more and more RGB devices, it means buying another PI and another dongle (Ether or USB) for each area. That's going to add up some.
2. The biggie. The 4 remote areas in my yard are all much further away from the wireless router than they are from the main house router, through several walls. Just to get the existing Pi to work, I had to buy high gain external antennas, and the existing Pi is much closer (right at the house wall) than the remote areas. Maybe I can buy an outdoor WiFi access point, but those aren't exactly cheap, either. Hmm.
First of all, if you were to connect the SSC directly to the PixelNet output, where is the power that the SSC both needs and sends to the strings going to come from. Another issue is wiring compatibility. The Pixelnet output is transmitting 4 universes of data, each one on a separate pair. The SSC expects and is wired for PixelNet on one pair and the other three pairs are combined for power input. If you were to plug a SSC into the PixelNet output, while the first pair would be ok, you would be shorting out the other three pairs which could damage the EDT. You could use a 4 port hub between the Zeus PixelNet output and the SSC. While this would require you to use the first PixelNet universe for your control, it would prevent the shorting out of the other three pairs and supply power to the SSC.
So the short answer is, "I'm missing something"?

The ETD is an E1.31 device so there is no reason that the PI cannot drive it from its Ethernet port.
Unless of course the Ethernet port is busy with the outdoor AP I just bought. <sigh>
You really don't even need a PI 2, a PI B+ would more than do the trick and for the price, how could you go wrong.
You ask that of an engineer? We can inflate the cost of project using only baking soda and vinegar into the stratosphere. >.d9
Seroiusly, thanks for the very thorough and clear explanations.