wifi boosters , mesh, powerline kits : getting better

Forums › LAN (ethernet/Wifi/routers/Bluetooth) › wifi boosters , mesh, powerline kits : getting better

1101

3122 posts

Uber Geek

#225410 17-Nov-2017 13:56

Hi guys.
Whats the current state of wifi coverage bosting tech, to remedy poor coverage ?

**** please dont say run cables. If I could run cables everywhere this would be a non issue
With Many newish homes, running cables & cutting holes in walls etc just isnt an option

Ive installed a few powerline kits at various sites , them are OK, but far from perfect

Have boosters & repeaters become any better ? Ive never had much success with those in the past.

I see some ISPs now offering ~mesh~ kits . What the real world experience with those ? If its a house/building with real
wifi degradation through walls, will mesh will still struggle ?

There a company thats been advertising on the radio : guarenteed better wifi
Looking at there site, they are relying on running network cables: so thats a fail, in many homes it cant be be done without ugly trunking runing up walls
and across ceilings
That also claimed "capable of transmitting data at a rate of 1300Mbps" . Is that just marketing BS ?

Is handoff between access points/mesh nodes etc still an issue , or is that more or less solved now.
Ive seen mesh systems where Ithings would stay connected to the weaker more distant node, requiring the ithings wifi to be cycled off/on .

People just expect wifi that allways works, everywhere . Trying to tell them its not realistic seems like you're making excuses

1 | 2

228 posts

Master Geek

#1903233 17-Nov-2017 14:11

The trick to residential mesh is not to exceed one hop. Think of it as a short point to point bridge. If you want to have two or three hops then you have an issue.

The 1300 Mbps is the raw speed. Note that it is a speed not a throughput and it is raw. To convert it to a TCP/IP speed used 40% for legacy radios, 50% for 802.11n and 60% for 802.11ac as a baseline. If you want a throughput use iPerf.

Aredwood

3885 posts

Uber Geek

#1903502 17-Nov-2017 22:33

I'll bite. Does the house have phone wiring that isn't used? As it can sometimes be repurposed as data wiring. Even the old telecom homelan twisted 2 pair cable was made to cat5 standard, so it can be used for 100baseT Ethernet.

And buy commercial grade Wi-Fi access points, instead of using an all in 1 device. I highly recommend the Xclaim Xi3.

PhantomNVD

2619 posts

Uber Geek
Inactive user

#1903509 17-Nov-2017 23:10

And doesn’t every house have a roof space too? Most of these also have ceilings, and in my (now 2 year old) new house I just ran the Cat6 cables loose in the ceiling from the garage star cabinet to the furthest end of the house and dropped it into a cupboard where the AP now serves this end of the (28m long steel frame) house quite happily..

deadlyllama

1264 posts

Uber Geek

Trusted

#1903540 18-Nov-2017 08:11

Crowdie:

The trick to residential mesh is not to exceed one hop. Think of it as a short point to point bridge. If you want to have two or three hops then you have an issue.

The 1300 Mbps is the raw speed. Note that it is a speed not a throughput and it is raw. To convert it to a TCP/IP speed used 40% for legacy radios, 50% for 802.11n and 60% for 802.11ac as a baseline. If you want a throughput use iPerf.

It's worse than that. 802.11n and later use MIMO which means they have a number of spatial streams or chains. 802.11n streams are "150Mbps" (so yeah, then multiply by 50% as you say). 802.11ac streams are 433Mbps. Manufacturers take the biggest numbers they can, add them all together, round -- usually up, and slap that number on the box.

1300Mbps is probably 867Mbps (two stream 11ac, 5ghz, so poor building penetration) plus 450Mbps (three stream 11n, 2.4GHz), rounded down to 1300.

1200Mbps is probably 867Mbps (two stream 11ac, 5ghz, so poor building penetration) plus 300Mbps (three stream 11n, 2.4GHz), rounded up to 1200.

No individual device is going to get more than the 11ac bandwidth (with your 60% discount factor, so ~500Mbps), and cheaper devices without 5GHz radios will all be on the 2.4GHz side.

In conclusion, marketers will be first up against the wall when the revolution comes.

Crowdie

228 posts

Master Geek

#1903778 18-Nov-2017 17:20

deadlyllama:

It's worse than that. 802.11n and later use MIMO which means they have a number of spatial streams or chains.

If is the spatial streams that counters multi-path and allows the higher data rates. Without them we are back to the 54 Mbps days.

deadlyllama:

Manufacturers take the biggest numbers they can, add them all together, round -- usually up, and slap that number on the box.

What the manufacturers are quoting is the total capacity of the access point. For residential deployments this commonly doesn't make sense but for enterprise deployments it is actually quote important for the channel plan and (application) capacity analysis.

deadlyllama:

1300Mbps is probably 867Mbps (two stream 11ac, 5ghz, so poor building penetration) plus 450Mbps (three stream 11n, 2.4GHz), rounded down to 1300.

1300 Mbps is VHT MCS 9 with 400ns GI and three spatial streams. It requires a minimum RSSI of -51 dBm and a minimum SNR of 37 dB.

deadlyllama

1264 posts

Uber Geek

Trusted

#1903856 18-Nov-2017 20:55

Crowdie:

If is the spatial streams that counters multi-path and allows the higher data rates. Without them we are back to the 54 Mbps days.

They help, but better encodings help too. Single-stream 11n can still achieve a max advertised rate of 150Mbps. T

Crowdie:

What the manufacturers are quoting is the total capacity of the access point. For residential deployments this commonly doesn't make sense but for enterprise deployments it is actually quote important for the channel plan and (application) capacity analysis.

...that's still going to have to be discounted by a fair bit, right? Hidden node, interference, boring TCP/application level issues, etc are going to reduce your user-visible performance, even in the aggregate, to well below that 1300MBps figure.

Crowdie:

1300 Mbps is VHT MCS 9 with 400ns GI and three spatial streams. It requires a minimum RSSI of -51 dBm and a minimum SNR of 37 dB.

I stand corrected. And wish I had more devices capable of three spatial streams. The original poster was referring to a company advertising wifi installs "capable of transmitting at 1300Mbps" which is true in a technical sense but extremely difficult to achieve in anything approaching real-world circumstances. I wonder if their APs have 1Gbps network ports...

Crowdie

228 posts

Master Geek

#1903878 18-Nov-2017 22:06

deadlyllama:

They help, but better encodings help too. Single-stream 11n can still achieve a max advertised rate of 150Mbps.

1x1:1 802.11n antennas still have multi-cast issues which is why 2x2:2 was the original minimum 802.11n specification.

If you only want a 150 Mbps data rate you would be better to have a 2x2:1 system than a 1x1:1 system. The addition antenna should result in higher transmit and/or receive sensitivities so should consistently maintain a higher MCS rate. With a 1x1:1 system you need to consistently maintain the highest MCS rate, which is unlikely in a residential deployment.

802.11n also includes some performance enhancing functionality, such as Space-Time Block Coding, which only occurs when the number of antennas exceeds the spatial streams.

deadlyllama:

...that's still going to have to be discounted by a fair bit, right? Hidden node, interference, boring TCP/application level issues, etc are going to reduce your user-visible performance, even in the aggregate, to well below that 1300MBps figure.

Hidden node is not as much of an issue as it used to be. I normally see it when the radio transmit power is too high and/or the data rates are too low so clients are reluctant to roam. Consumer and SOHO access points still have issues with hidden node as they provide little to no roaming assistance unlike enterprise vendors like Aruba and Cisco.

In residential areas co-channel contention (wait states caused by channel contention), adjacent channel interference (why did the telcos ship routers that can select 2.4 GHz channels other than 1, 6 or 11?), poor access point placement (commonly where the ONT is) and an inadequate number of radios for the required coverage area are your biggest issues.

deadlyllama:

The original poster was referring to a company advertising wifi installs "capable of transmitting at 1300Mbps" which is true in a technical sense but extremely difficult to achieve in anything approaching real-world circumstances. I wonder if their APs have 1Gbps network ports...

Generally higher capacity access points use multiple 1 Gig interfaces in LAG or proprietary standards to exceed 1 Gbps on CAT 6 cable, such as Cisco's mGig.

Sharesies

Trade NZ and US shares and funds with Sharesies (affiliate link).

mdav056

609 posts

Ultimate Geek

Subscriber

#1903905 19-Nov-2017 07:37

To answer the original question, my answer is Netgear Orbi -- a pair of these solved all my problems in my long, 2-level house (wood and brick), no handovers and very very fast (I get the same 100/20 over wifi as over Ethernet). Easy to install, worked for 6 months with nary a problem.

gml

Crowdie

228 posts

Master Geek

#1903976 19-Nov-2017 11:43

mdav056:

To answer the original question, my answer is Netgear Orbi -- a pair of these solved all my problems in my long, 2-level house (wood and brick), no handovers and very very fast (I get the same 100/20 over wifi as over Ethernet). Easy to install, worked for 6 months with nary a problem.

Most residential wireless routers have omni-directional antennas that propagate signal horizontally out in a spherical pattern but differ in their vertical signal propagation. The higher the gain, measured in dBi, of the omni-directional antenna the more horizontal signal propagation and the less vertical signal propagation. Therefore, if you have a single level dwelling a 5 or 6 dBi gain omni-directional antenna would be a better choice than a 2.14 dBi gain omni-directional antenna. The 2.14 dBi gain omni-directional antenna will propagate more signal down into the ground and up in the ceiling cavity than a 5 or 6 dBi gain omni-directional antenna. For a two level dwelling you might want to start with a residential router with a 2.14 dBi gain omni-directional antenna on the ground floor on one side of the floor with an access point with a 2.14 dBi gain omni-directional antenna on the upper floor on the other side of the floor. This is a general theory and the size, shape and construction of a dwelling has a serious affect on your radio placements.

If you are worried about the signal covering the required areas remember that, all things being equal, an external antenna will provide greater signal coverage than an internal antenna.

Lastly, I haven't seen a standard residential router that can out perform a quality access point yet so, as many people have already advised in other threads, the best solution from a technical point of view is nearly always going to be disabling the radio(s) in the residential routers and deploying access points with Ethernet cables back to the residential router. The router acts as the router and the DHCP server. The access points handle all the wireless requests.

Crowdie

228 posts

Master Geek

#1904168 19-Nov-2017 18:20

This is quite a good "what is inside the Orbi" review at https://www.smallnetbuilder.com/wireless/wireless-reviews/33028-netgear-orbi-reviewed

vulcannz

436 posts

Ultimate Geek
Inactive user

#1904280 20-Nov-2017 07:03

Crowdie:

Generally higher capacity access points use multiple 1 Gig interfaces in LAG or proprietary standards to exceed 1 Gbps on CAT 6 cable, such as Cisco's mGig.

2.5Gbps and 5Gbps have been standardized, I run 2.5Gbps on cat5e no problem.

phrozenpenguin

842 posts

Ultimate Geek

#1904376 20-Nov-2017 10:12

vulcannz:

I run 2.5Gbps on cat5e no problem.

What gear do you use for this?

vulcannz

436 posts

Ultimate Geek
Inactive user

#1904378 20-Nov-2017 10:14

phrozenpenguin:

vulcannz:

I run 2.5Gbps on cat5e no problem.

What gear do you use for this?

Sonicwall with 802.11ac wave2 mu-mimo, mostly it seems to help with contention. Kids rooms are not wired so they use wifi - for gaming they are not seeing the lag issues.

deadlyllama

1264 posts

Uber Geek

Trusted

#1904385 20-Nov-2017 10:37

Does the MU-MIMO need MU-MIMO capable user devices or is all the magic in the access point?

vulcannz

436 posts

Ultimate Geek
Inactive user

#1904410 20-Nov-2017 11:41

deadlyllama:

Does the MU-MIMO need MU-MIMO capable user devices or is all the magic in the access point?

good explanation here: https://www.computerworld.com/article/3072469/wireless-networking/13-things-you-need-to-know-about-mu-mimo-wi-fi.html

1 | 2