Inconsistency with Orbi Speed

Ok, some more information that very likely exonerates the Orbi. As I mentioned above, after exchanging for a new system and observing the same issue, I returned it for a refund on Friday. This morning on somewhat of a whim, after learning that the Linksys Velop indeed did have DHCP reservation (it's just hidden away in the app), I decided that despite it's reported simplistic feature set, I could make it work with my networking setup. So, even though I no longer have a direct interest here, I wanted to relate some new information for the sake of posterity, accuracy, and potentially helping others.

Once the Velop was up and running, I discovered that it too was exhibiting the same issue... good speed sometimes with random inexplicable slowdowns when connected through the satellite (initially I just had 2 of the 3 Velop units in operation). While two presumably defective Orbi systems in a row was arguably too coincidental, a totally different product also showing the same problem meant there had to be something else.

While I still don't know exactly what that "something else" is, I have a strong suspicion. In the 2nd floor closet in which my networking gear is housed, there are those white-coated wire "ClosetMaid" shelves. Now, I had previously had concerns about this, and had mounted my R7800 as high in the closet as I could to get it away from these shelves (antennas are a couple of feet above the top shelf), though it did not seem to be affected. With the Orbi, I could not wall mount it, though I was able to place it on a board that extended past the edge of the shelf, so that the signal down to the direction of the satellite would not have to pass through the shelves. It appeared to be fine, as I got good speed hardwired through the satellite, and got good speed when the client was connected to the router's WiFi, even if I was downstairs and in a position where the signal DID have to pass through the wire shelves.

Because the Velop can do an ethernet backhaul, and I had another node, I placed one in my upstairs office (which didn't really absolutely need one, but I have ethernet run here from the wiring closet).  Once it was up and running wirelessly, it too gave iffy speeds. Connecting it to my hardwired network predictably boosted speeds on this node to pretty much the same as I was getting through the main unit, and, as I hoped would occur, the troublesome downstairs node switched its wireless link away from the main unit in favor of this additional hardwired node (which was in a good position for reception). The speed problems when connected to the downstairs node vanished.

The only explanation that I can come up with is that the close proximity of the metal in the closet (even though it was not in the path of the signal) was causing it to reflect around a bit, and this was confusing the router (which interpreted it as interference, and slowed the connection down as a result). But what doesn't make sense to me is, if router placement was indeed the issue, why would this only occur when the client was wirelessly connected to the satellite, and not when a client was hardwired to the satellite?

I don't know. But in my Orbi troubleshooting, and with the Velop, though I varied the satellite's position, the placement of the main router was constant, which although I thought I had taken sufficient steps to confirm was not an issue, this was apparently revealed to be a mistake. Though this was seemingly proven by the addition of another hardwired Velop in the upstairs office that the downstairs node connected reliably through (taking the main unit in the closet out of the satellite node equation), I suspect I would have seen the same positive result had I gotten the Orbi router out of and away from the closet for a test, which in hindsight I could have done by running some temporary network and cable modem wiring.