Basic Node performace testing ?

[thatSFguy]

Node quality is important as it can be the difference between a good working mesh and messages that never get answered. Over the last couple of years I've got quite a collection of nodes:

• TLora 1.6
• TDeck
• RAK4631
• Station G2 (best performer)
• Seeed Xiao
• Waveshare hat on RPI

When shopping for a node, I generally look at the specs to see what the expected "performance" should be. However, recently I've noticed very different RF performance between nodes, despite using the the same antenna or a similar antenna verified with a NANO VNA. E.g. I go out a relatives house, verified coverage with my mobile node and then attempted to place a "repeater" client node and to my surprise the node does not connect. So besides just using the App to compare SNR and RSSI from a test node to a given node, what other tests do you guys perform?

[SimbimChimbetov]

How do you Test Performance?

Here in Europe on 868Mhz I found out to test anything about Antennas except testing its swr is pretty useless.
The SNR and RSSI can fluctuate hard...
Every antenna, with every gain I tested on the same spot, to and from the same node can be better and worse then others at different times.

The only thing I watch on my Node Builds are
Good swr
Cables as short as possible
Adapter/connectors less as possible

[thatSFguy]

Up until now I would have agreed that. I generally test my antennas with a Nano VNA, minimize cable length or use high quality cable, like 1/2" superflex for longer base runs.

I recently got some Seeed Xiao nodes and promptly designed a magnetic roof node that looks like a shark fin. Shortest possible cable length, right? Well, I pulled up to a location that I know has coverage and nothing.... I came back later with TLora node that I just put on my dash and all traffic started going through the Tlora, despite being in the the vehicle.

I ended up doing some more testing, such as using the same antenna, in the same location, on the TLora and Xiao and the Xiao always performed worse. (Spoiler: it seems on the receive side)

I'm not doing this to rag on the Xiao node, it's tiny and there are use cases for it, and I thing it would be useful to have some way of categorizing the nodes depending on what one is looking to achieve. For example, if I'm establishing a new mesh, I would go straight to a Station G2, as it makes up for a lot of weaker nodes.

Cheers!