Bad connection to IOS devices

Many of us are facing same issue and is mentioned in many threads.  I have been able to mitigate this to an extent by rolling back to 1.4.0.16 with a factory reset. PLEASE bring this to Engineering attention this same issue started in 1.4.0.18 and persist in 1.4.0.34 as well along with random drops in wifi, some are patiently waiting new Firmware while others have given up to try competitors products sadly, Orbi is a very powerful machine just needs a stable Firmware.

Just bought Orbi and using it for a week or so. 

I'm facing the same problems with disconnects. My iPad Mini 2 connected in 10 hours 97 times as I wrote here already. Same is happening with other iOS devices. Also sometimes my iPhone 7 just looses the connection and falls back to cellular for a minute and reconnects. This is really a big issue as this can cause problems for people not having big data plans. 

Another important thing is that the logs from Orbi does not show to which network (2.4 or 5 Ghz) and to router/satellite the device was connected. 

Please get this issue fixed asap!!

Not speaking for Netgear, but as someone who was in this game for many years, I can tell you that it's not a simple issue.

Every vendor - Netgear, Asus, Linksys, D-Link, you name it - uses off-the-shelf SoC (System-on-a-Chip) parts for the core, which includes the wireless radios.  The drivers for said radios are provided by the SoC vendors, which are typically in unmodifiable binary-only form (ostensibly to prevent reverse-engineering and intellectual property theft, though that's a very dubious excuse).

When an issue like this is discovered, it means a few important things:

- Any other vendor using that SoC is likely facing the same issue (and indeed, diconnect issues abound, with eero, Velop, and others who used off-the-shelf chips)

- Even after the vendor can reproduce the problem reliably and consistently (not as easy as it sounds, given the convoluted nature of wireless and the myriad types of client devices and chips and drivers!), they are almost always entirely powerless to mitigate it.  The only thing they can do then is escalate to the SoC manufacturer, hope they listen and acknowledge the issue, and that they have enough leverage with them to force a driver-level fix or SoC microcode update to correct it (because if it's a chip-level issue, field repairs obviously aren't going to happen).  This generally requires lots of discussion, negotiation, and threats of losing future business before anything material comes of it, which can take months to years depending on the situation.

Yes, it's a crappy place to be, for all companies, but it's the way it is.  If any company wanted to produce their own custom SoC, they could, and then they'd have full control and latitude to provide their own fixes...and having to manage foundry partners, small yields, a tremendously larger engineering support staff, much slower time-to-market, and the costs will be passed along accordingly.  Would you pay $1000 for Orbi, and wait for 2018 for it to come to market, being bypassed by every other major player in the industry?  No?  Me, either.  That's why all the major vendors do more or less the same thing and face more or less the same problems.

There's also the issue of a very small number of SoC makers to choose from - it's an incredibly niche, low-volume market as a whole (mobile devices excluded), which is another reason you'll find common issues across devices from different manufacturers...typically, the SoC defect spans all chips and platforms for a specific type (for example, AC3200).

In the past, the 'fix' was to admit defeat that the SoC maker is never going to acknowledge the (very obvious) issue, much less correct it in-market, and lo and behold, a couple of years later the next iteration of hardware will come out with an updated SoC that offers more speed and - more importantly - fixes to past connectivity issues, forcing anyone still using the older platform to upgrade just for the stability.  As crappy and profit-driven as that may sound, the vendor's hands are tied...if the chip doesn't work as it should, what should they do?  Switch makers and retool their entire surrounding design in the middle of a production run?  Bow out of the market and offer refunds to everyone who purchased the product, probably going out of business in the process?

As if all this weren't bad enough, the mobile SoC manufacturers (mostly cell phones and tablets, though some laptops as well) - especially those that don't rhyme with Pintel - are even worse about adhering to (often poorly written and thought-out) wireless standards, which compounds the problem (see: Apple).

We all went through this with the transition from 802.11b to 802.11g, then from 802.11g to 802.11n.  802.11ac has been somewhat smoother, but hardly flawless...and it's a vicious cycle that will never end until product manufacturers have much more control over the SoCs they use.  Given it's been this way for over a decade, I don't anticipate it getting better any time soon.

I'm not defending Netgear (or any other product manufacturer), just trying to offer a bit of an "insider" view on what really goes on in industrial wireless product engineering and the awful-but-unavoidable downstream impact it can have on consumers.  Given Netgear's target audience and sensitivity to price points, they really have no viable option other than to go with the same SoCs everyone else uses and the same growing pains that result.

Welcome to wireless, kids!  Ain't it fun?  =)

Rodney