New Orbi Setup now no disconnects on iPad and other devices

[andykerslake]

I have just installed a new orbi setup with router and 1 x satellite.

I struggled for 2 days trying to work out why some devices iPad 2 and an Android Phone were disconnecting on a very regular basis.

Eventually I reduced the 2.4ghz band transmit power to 25%, this has fixed all of the disconnects. 5.0ghz is set to 100%. Also currently beam forming and mu-mimo are turned off.

Previously I was using an Asus router with dual band wifi, but the bands had different SSIDs. I think the fact that the Orbi has a single SSID is causing some devices to disconnect due to the competing signals.

I am very familiar with wifi systems having installed simple access points and enterprise wide systems at our customers, and my reason for using Orbi was to look at a possible new solution for our customers.

[DarrenM]

Hello andykerslake

Thanks for sharing these tips on what is working for you.

DarrenM

[Slyke]

Did that fixed your problem? I am experiencing the same problem as you described, sudden and regular disconnection on some of the ipad at home. Other devices seem to work with no problem but I ADS are suddenly disconnected (I have even the latest version)

Thank for the feedback you can provide me 

[andykerslake]

Yes it did fix all the disconnects. The problem only affected Apple devices.

I have not tried turning back on the 2.4Ghz as the current setup works, and I do not need that band.

Andy

[Sevi2]

Just bought unit from Best Buy while on sale and got it up and running ok.

Wired connection was over 100Mps while wireless was a paultry 2MPS.

Went into my laptop's wireless adapter and switched from "Prefer 5Ghz" to "No preference" and BAM!  120MPS on wireless too.

I have to admit, as a new customer i am really impressed with the system, router on one corner of the house and satellite in the hallway second floor, have signal everywhere in the house.

I am following the threads with the Internet drops very closely while under the 15 days return policy.

If i were to take this back, i guess i would look into Velop from Linksys.

[Slyke]

The hardware is, without a doubt, the best I have ever owned. I tray many (many!) different hardware and is the first that can send my 100MB around my house and through concrete walls dropping only a few MBs on its way..... In my opinion, it is worth every penny.....

My only frustration has been the sudden internet disconnection on the iPads in my house... After reading Andy suggestion, I was finally able (yesterday) to drop the power of the 2.4 network to 25%. I tested the connection on my devices (fearing it would suffer from the reduction in some corners of the house) but it I still have very good speed...

So Far no connection drops but in need more days to confirm the change fixed the problem..... I'll keep testing it over the weekend and report back my results....

[andykerslake]

I still found that even at 25% I was still getting disconnects on iPad, iPhone and Mac, so I ended up switching off the 2.4ghz and since have had zero disconnect issues.

I think the problems lie with the way wifi is implemented on Apple devices, they do not quite follow the standards, and seem to be unable to handle a single SSID that works on 2.4Ghz and 5.0Ghz.if there are two SSIDs one for each then they are fine.

[marshymallow]

WIFI Drops are not happening only to iOS but Android devices too.

The positive is that I'm into the 3rd day of testing the new Beta FW and so far I've yet to encounter a single WIFI drop.

[JMU1998]

Hopefully beta addresses this because even with 2.4GHz completely turned off still have disconnect issues frequently 

[Slyke]

Yes... As Andy mentioned on his last comment, even reducing the 2.4ghz signal to 25% does not solve the connection issue with IOS (and Android) devices. Change my setup to 25% and I am still facing sudden signal drops.... 😞

Have you had any issues using the new beta firmware? Is it stable? 

Thanks

[tallackn]

A little bit technical now.

Your endpoint (iPad/iPhone/etc) makes roaming decisions for a number of reasons.  These include the signal getting weak (as you walk away from an AP) and packet retransmissions (often because you are walking away from the AP), and a number of other reasons.

Once your endpoint makes the decision to roam, the way it chooses where to roam to depends on what protocols your AP support (and your endpoint).

Today we have 802.11k, which amongst other things, provides a "neighbour list" to your endpoint, telling it what other access points are around and what channels they are on, and how strong those access points see your endpoint, and how busy they are, and lots of other things.

So if your endpoint supports 11k, and your access points support 11k, then your roaming decision is very easy and intelligent.

Without this the method for roaming is quite unreliable.  What happens is your endpoint "probes" all of the channels in the band (frequency) it is on, and then chooses to attempt to roam to the AP with the strongest signal.  If it does not get an adequate response on that band it will scan the other band.

This is bad for many reasons, some of them being how long it takes to scan across the band (the 5GHz band has more than 12 non-overlapping channels in most regions) and some endpoints not considering a better AP on another band because it found a suitable AP on the current band.  There are many other bad reasons why roaming without 11k support is problematic.

In short, 11k support is required.  Most modern endpoints (iPads/iPhones newer than 2 years and with current iOS versions) will support 11k.  Most enterprise AP's will support 11k (because as you can imagine there are LOTS of AP's in the enterprise, and reliable roaming is important).

The question begs, do our Orbi's support 802.11k?

I know the AC3000's do, as of a firmware update last year!  They even use the 11k to "band steer" your client back to the band that is best, plus lots of other things.

So, if you have the AC3000's, and you have the current firmware, ensure your endpoint has 11k support and all of your roaming problems will go away.

Thank you Netgear for another WONDERFUL product!!!  🙂

[Slyke]

Mmm I think I followed your technical explanation but I am still experiencing sudden loss of connection on my iPads (not on my laptop or any other device)

I checked my Router info and I understand I have everything up to date: 

Hardware Version	RBR50
Firmware Version	V1.10.1.2
GUI Language Version	V1.0.0.284
Operation Mode	Router

On the other end, I have different iPad / iPhone models including the latest iPad Version (mine) running IOS 10.3.1 (believe the 10.3.2 is out) hence why do you think, if all my devices are new and updated, based on your technical explanation I still experiencing connection issues?

Thanks

[tallackn]

It becomes a process of elimiation.  Quite a few things can be ruled out though the answers to some questions.  Let's try this.

Firstly, let's identify two different "deliverables" we are addressing.  One is "performance" and one is "availability".  

We consider availability issues to be those that result in your wireless disconnecting.  That means your endpoint actually thinks it has disconnected from the wireless network and either rejoins itself after a few seconds or more or you need to rejoin yourself.

We consider performance issues to be those that result in you getting poor throughput or no throughput while you are still associated.  That means your endpoint still thinks it is associated to your wireless network.

So, with these terms firmly understood, let's have you anser some questions to the best of your knowledge.  Better to say "not sure" than to give inaccurate or incomplete answers.

1: You have more than one endpoint in the same room experiencing the same performance and/or availabilty issues at the same time.

2: You have one or more endpoints experiencing availability issues when you are not moving around the house.

3: You have one or more endpoints experiencing performance issues when you are in the same room as the access point.

4: You have a single endpoint experincing performance and/or availabilty issues in the same room as another endpoint at the same time that is not experiencing the same issue.

5: You have one or more endpoints that experiences an availablity and/or performance issue that can be reproduced on demand.

Answer those questions to the best of your abilities and we'll see what the possible causes are.

PS: Apologies for my wording, I am a wireless engineer by profession and sometimes find it hard to speak to regular people about the technology.  So please do understand that it is my problem that things are worded the way they are, not your problem in understanding.  🙂

[Slyke]

Hi 

Took me some time to gather all the info you request me (thank you for helping me with this!). I'll answer in Caps bellow your questions

1: You have more than one endpoint in the same room experiencing the same performance and/or availabilty issues at the same time.

I USUALLY HAVE MY LAPTOP, PHONE, AND IPAD TOGHEtHER IN THE SAME ROOM. IN SOME INSTANCES ONE OF MY SONS ARE AROUND WITH THEIR DEVICES. THE DISCONNECTIONS ONLY HAPPEN ON OUR IPADS DEVICES. I DO NOT EXPERIENCE ANY CONNECTION ISSUE WTH MY LAPTOP OR MY SON LAPTOP ON THE SAME ROOM. I CANT TELL YOU FOR SURE ABOUT MY PHONE, IT SEEMS NOT TO HAVE THE SAME PROBLEMS

2: You have one or more endpoints experiencing availability issues when you are not moving around the house.

ALL CONNECTION DROPS SEEMS TO BE WHEN CONNECTED TO THE SATELLITE. MY HOUSE HAS TWO FLOORS, UPSTAIRS IS WHERE I HAVE THE MAIN HUB AND DOWNSTAIRS THE SATELLITE. NO MATTER WHERE I LOCATED (DISTANCE TO THE SATELITE) THE IPADS HAVE SUDDEN DISCONNECTIONS. 

3: You have one or more endpoints experiencing performance issues when you are in the same room as the access point.

YES. 

4: You have a single endpoint experincing performance and/or availabilty issues in the same room as another endpoint at the same time that is not experiencing the same issue.

YES (MI IPAD GETS DISCONECCTED, MY LAPTOP KEEPS A STEADY CONNECTION)

5: You have one or more endpoints that experiences an availablity and/or performance issue that can be reproduced on demand.

IF I UNDERSTAND THE QUESTION CORRECTLY, THE ANSWER IS NO. THE SIGNAL DROPS ON IOS DEVICES OCCUR UNEXPECTEDLY.  ALL iPADS IN THE HOUSE HAVE THE SAME PROBLEM (AND I HAVE TO BEAR MY SON MADDNES WHEN THAT SUDDEN DISCONACTION HAPPENS IN THE MIDDLE OF ONE OF THIS CLASH OF CLANS BATTLES... 😉 

Thanks again for your help and let me know if you need any other information

[tallackn]

Now number 2 is quite telling.  If you never experience the connection drops with endpoints associated to the router access point but you do experience them with endpoints associated with the satellite access point then we could be seeing a case where the backhaul link between the two acces points is being affected.  On some devices where the uplink (backhaul) goes down the client endpoint network is shutdown as well.  This is referred to as uplink tracking.  This is done so that your endpoints do not remain associated to an access point that has nowhere to route the traffic, and gives your endpoints the opportunity to connect to another access point (perhaps the router one in your case) to remain connected to the network (at a lower speed as it is likely further away).

But in your case, if this is happening then you may find that some of your endpoints are simply not in range of the router access point when the satellite loses its backhaul.  That is why you may see that some endpoints reconnect (or do not lose connection at all) while on the satellite but others do.  You can test this by intentionally turning off the satellite and mapping out the coverage edge of the router.  If the endpoints that experience outages are outside of the range of the router then you have likely identified a possible cause.  Also your laptop will have a better antenna inside it so will likely have the range to reach the router when the satellite fails (as per your answer to question 4).

Why is your satellite losing its backhaul?  There are a number of reasons, so we will address them from most liekly to least likely.

The first reason is the possibillity that the backhaul frequency is experiencing some interference.  These units backhaul on the top end of the 5GHz spectrum (UNII-3 channels, 149 upwards).  These channels are traditionally the least noisy in urban environments, and are far enough away from the lower UNII-1 channels that the units use to provide connectivity to your endpoints.

You can mitigate this problem by increasing the signal strength of the backhaul.  In the case of these units, the only real way you can do this is moving the units closer together, which will let them hear each other "louder" which hopefully will let them talk over any interference they are experiencing.  Read on for more info about this.

The second reason is that your units may be too far apart or have too many materials between them that are weakening the signal.  The golden rule of how far you can place them apart is the same distance your endpoint can remain connected to your router on.  So you can validate this by turning off your satelite and connecting to the router access point with your laptop client and then moving away from it until you start seeing the signal stregth indicator go down.  Do keep in mind you do not want to go to the very edge of your coverage as this will result in a signficant decrease in network speed.  Move away until you start seeing the coverage bars go down and then that is a good location for your satellite.  Remember, best to place the units up high on a shelf somewhere, so test with your laptop where you intend to put the unit.

This may expose the reality that the network does not really extend that far away when connecting indoors.  But then you need to think about the signal path.  If you can place it in room x where the path between the two units pass through 3 internal (read: plasterboard or wood) walls and floors or in room y where the path passes through an external brick wall or a concrete floor, then the first room will be better.

The third cause could be a unit problem.  This is unlikely due to the fact that you are on the latest firmware and the problem is intermittent.  But sometimes, just like a computer, the units can have a bit of a software problem.  You can double check this by firmware defaulting both units (lilttle hole with a paperclip) and setting them up fresh again.  Only do this if your backhaul problem resoltuion steps do not fix your issue.

What I would like to see is that indicator ring atop the units flash some sort of indicator color regarding backhaul status.  If you saw it flashing red when you have network problems (for example) you know your backhaul has gone down and you need to move away from the satellite into coverage of the router.  Netgear, can we have this with a firmware update please?  Perhaps it is there now and we just have not noticed.

I love this product!  It truely is a masterpiece of consumer engineering.  Well done Netgear!

[Retired_Member]

I can almost guarantee that it is a hardware problem causing the backhaul to drop that is causing all of these issues. Some devices handle a brief drop differently than others which is why it appears that nothing happened.

I would recommend you install the Wi-Fi Analyzer app and you can watch your backhaul. There is a 2.4 GHz and a 5 GHz backhaul and they both come from the router. The backhaul will have no SSID. Since the router drops the backhaul, satellite placement doesn't matter. 

I really like your troubleshooting process @tallackn, they should hire you! Do you have a stable backhaul if you check? Here is my backhaul:

[Bing-Stroller]

BobertSaget, for some reason images are not loading.

I use WiFi Explorer and load AP Timecourses. You can switch between 2.4Ghz and 5Ghz and then select an individual SSID.

As BobertSaget says the backhaul ones have no SSID.

The results are eye opening.

[tallackn]

All things considered, except due to a hardware fault or a firmware error, the unit should not be contributing to backhaul failure.  Not intermittently anyway.

Going to get really technical now, sorry for those who have trouble following.

Let's assume that the backhaul is a standard WDS link, with the router being the root and the satellite being the client.  Let's also assume the total power budget of the link to remain the same (no change in obsticals or antennas), and let's also assume a minimum SNR of 20dB.

So, the WDS operates at a symbol rate that is optimised for the RSSI, nice and stable, well established.  It may indeed get some additional spatial streams up at reasonable symbol rates, but for the sake of troubleshooting let's assume it stays on a single stream.

So now you have an established link, at least 20dB above the noise floor required for the symbol rate and no changes in physical environment.  Why would the link fail?

There is a small chance that there is noice introduced, and considering we are likely operating on an 80MHz wide channel using omidirectional antennas that is quite likely.  But even if such a thing happened the unit would drop the symbol rate to bring the RSSI back to the minimum reliable level and/or narrow the channel width so that it could move away from the noise (assuming the noise is not right across the UNII-3 band), thus maintaining the link.

If we were operating right at the edge of our coverage area that might not be possible, so that could be the cause of the link failing.

If the hardware were fauly it is unlikely it would be faulty in a way that would result in intermittent failures.  And if the firmware were faulty we'd see a whole lot more people experiencing this problem.

There are other more sinister causes of course.  You could have someone spoofing disassociation frames, but I believe there is management frame protection enabled by default on the backhaul radio (maybe even client radios).  And any kind of RF denial of service would manifest for the endpoints associated to the router unit as well as the backhaul.

I'd rule out the range possiblity probelm first.  Alot of people place their units right at the edge of their coverage areas leaving little room for the radios to adapt to changing RF conditions.

My units are placed about 20m apart with about 8dB of loss from internal walls, so I am getting a great experience (backhaul datarate seems to be around 400Mbps of client payload plus management overhead) and am quite confident that there is plenty of margin to work with for RF problems.  Placement is 90% of the key to success.  🙂

[Slyke]

My nightmare still continue....What I do not get the analysis is why only ios devices are suffering the sudden disconnections. If the were a problem of communication between the hub and the satellite won't all devices be affected equally? The access point is located on the room above the satellite. when measuring speed, the satellite is able to keep amazing internet speeds hence I take it as a sign that the distance between the two devices is acceptable. Moreover, all other devices on my house, even the one on the far reaching places keep a good connection except iPads....

I really do not know what to do.... 

[ikjadoon]

At Slyke, for your devices that have stable connections, could they be connected to the router instead?

For me, my Satellite-wired clients died and this confirmed the Satellite was essentially out.

[tallackn]

Ok, we need to take great care here to ensure we are 100% that the following statements are true.

1: ONLY iOS devices are disconnected, never any non iOS devices.

2: When more than one iOS device is associated to the satellite at the same time a disconnection occurs, all of those iOS devices are disconnected.

3: When the iOS devices are disconnected from the satellite, all non-iOS devices remain connected.

Are these three statements true?  If so, it is possible that you have a problem that is not related to the satellite itself.

But, and this is very improtant, if just one of these statements is not 100% true then your problem will be with the satellite not with the endpoints.

[Retired_Member]

1. False
2. Not related to the problem
3. Not related to the problem

The backhaul connection from the router to the satellite is dropping which means every device connected to the satellite is technically dropping but is usually only noticed on iPads or sensitive devices like cameras and Wi-Fi calling. The problem is with the router and not the satellite and definitely not the end points.

[tallackn]

Apologies BobertSagnet, my reply was related to Slyke's last post.  It does not thread so well on here.  I am not sure how it would relate to a problem you may have.

What we are focusing upon with those questions from Slyke is the likelyhood that the iOS devices are not roaming to the router when the satellite drops the client network when its backhaul goes down.

Given the proximity of the router and the satellite there is likely coverage overlap for the client networks and it is possible that the non-iOS devices are seamlessly roaming to the router when the satellite fails.

That is where my questions were focused.

[Retired_Member]

Got it, sorry about that I was only focusing on the backhaul but I'm glad you guys are taking it a step further to see how devices respond to that differently.

Quick thought that might relate, do you notice the the same speed all the time whenever you run a Speedtest on any of your devices? The backhaul drops it's 5 and 2.4 GHz signal at different rates and I wonder if non iOS devices are better at switching from one to the other instead of switching completely to the router and noticeably dropping. I will go from 230 Mbps to 50 Mbps on my android with no explanation.

[tallackn]

Yeah, that is related to band steering.

When your endpoint is trying to associate to the SSID, the AP will supress the responses on 2.4GHz for a bit hoping the endpoint will give up and move to the 5GHz band, where it will answer and allow association.  If it never hears the endpoint on 5GHz then it will respond on 2.4GHz.  In this way endpoints that support both bands will be "steered" towards 5GHz.

The reason for this is 5GHz has a whole lot more channel space.  The AP can operate a 160MHz wide channel up in the UNII-1 channel space that is more often than not free of significant noise.  That way it can get multiple spatial streams at nice high symbol rates and get that nice fast speed you are doing.

But when your endpoint cannot associate to the 5GHz radio, perhaps because the signal is too weak (5HGz does not penetrate obsticals in the home as effectivly as 2.4GHz does) then you are going to end up on a 20MHz wide 802.11n channel on the 2.4GHz radio.  Lower symbol rates, likely single spatial stream (because the additional paths are too weak), hence your lower speeds.

Some endpoints do not bandsteer so well, so they may end up "sticking" to 2.4GHz after you move away from the AP (out of range of the 5GHz coverage) and then remaining on it when you walk close to the AP again.  You can see this is the case if you are in the same room as the AP and associated at lower speed, turn off the endpoint wireless adapter for a few seconds, turn it back on again, and you will likely reassociate to the 5GHz radio.

Some endpoints don't need to be band steered, and will always "preferr" 5GHz, always trying to use it first, and always trying to get back onto it at regular intervals.  This is true for iOS devices, so you may see these suffering less from the "stickiness" that other non-iOS endpoints do.

Honestly, I could prattle on for HOURS about this stuff if you let me.  You'd think doing this as my day job would satisify my appetite for it, but it does not.  I love this stuff.  And I love this Netgear kit.  Honestly, for a consumer product, this kit is amazing!