Firmware Version 1.10.1.2

Also Computer Engineer with 17 years in the field.. I tried there beta firmware and again discconecting all of my devices.. I wish they would post something fast not a good look for future sales.. if it wasnt for the range i would just keep my FIOS Gateway

st_shaw

1) The backhaul drops, by definition, are a problem.  It is literally the equivalent of unplugging and replugging an Ethernet cable every few seconds.  You may not notice on TCP, due to liberal socket timeouts, but with fire-and-forget UDP (nee streaming and VoIP), you will absolutely see disruption.

2) I can't say a thing about your scanning apps, but yes, the backhaul is standard wifi with no SSID broadcast.  With a master's in computer engineering, I'll take it as read you know how to telnet:

root@RBR50:/# iwconfig ath2
ath2      IEEE 802.11ac  ESSID:"NETGEAR_ORBI_hidden85"
          Mode:Master  Frequency:5.785 GHz  Access Point: [redacted]
          Bit Rate:1.7333 Gb/s   Tx-Power:30 dBm
          RTS thr=64 B   Fragment thr:off
          Encryption key:[redacted] [2]   Security mode:open
          Power Management:off
          Link Quality=0/94  Signal level=-95 dBm  Noise level=-95 dBm
          Rx invalid nwid:9172  Rx invalid crypt:0  Rx invalid frag:0
          Tx excessive retries:0  Invalid misc:0   Missed beacon:0

It is not the same situation as Sonos, whose non-802.11 interference with 5GHz channel overlap is well-documented.

3) We actually know precisely how Netgear uses backhaul - I documented it extensively shortly after the product's release on this very forum.  We also later determined and documented how 2.4GHz backhaul fallback functions.

You may want to do a bit of academic research on how wireless communications function before coming to erroneous conclusions that total signal loss doesn't cause a problem.  I suggest you start with a more meaningful test that uses UDP or custom packet framing, as TCP by design will not show loss in very short bursts.

Rodney

Orbi uses the one of the 5.0 Ghz channels as the backhaul.

It’s also not a true mesh system like a Ruckus but we are hopeful they will fix this.

It works more as a point to node system. So if you have more then one satellite to want the base in the center of them.

A true mesh system it would not make any difference where the base is.

st_shaw, I am a little confused by your credentials because your post made no sense. 

"The backhaul very rarely shows up on any of them (scanners) during normal operation.  This would indicate that whatever protocol Netgear is using for the backhaul is not standard WiFi and does not reliably appear on WiFi scanners."  - Why would it indicate that? I'd like to learn more about this proprietary protocol that randomly drops off of scanners.

"I have a screenshot where you can see the backhaul on channel 157 and it disappears from the graph." - Of course you do, that is the backhaul dropping problem. 

"So, definitely the backhaul was still working even though it wasn't showing in WiFi Explorer." - I guess we can go ahead and mark all of those threads as solved because the backhaul is definitely supposed to drop, right?

"If somone could collect data that directly correlates the timing of the backhaul signal changes with network outages and returns to service, then that would be more meaningful." You have 28 years of debugging experience, I can think of no one better to run that experiment.

Here is what I am talking about. The image upload function seems to be gone, so here is a link to a screenshot on google drive.

Backhaul Graph

In the top section of the screenshot you can see the Orbi's hidden 2.4 and 5 GHz signals on WiFi Explorer.  These are highlighted and appear on channels 1 and 157.  The hidden Orbi signals appear briefly then go away.  They were last seen at 9:55 AM as shown in the upper right.  At the bottom of the screen is a window with a UDP iperf3 session conducted just before the scan was stopped at 9:58 AM.

The iperf session was between my MacBook and a Raspberry Pi that is hardwired to the router. The MacBook was connected via WiFi to the satellite.  The AP associated to the MacBook is indicated by WiFi Explorer with bold formatting.

The iperf session was started after the Orbi backhaul signal dropped from the scanner, yet 233 MBytes and 29K packets were transferred with 0% packet loss. This example shows the backhaul working while the signal is not appearing on a WiFi scanner app.  This is why I say I'm not convinced the signal dropping from a scanner app means the backhaul is not working.

Maybe there is something strange with my setup or something wrong with my logic, but it seems pretty simple.