20/40 MHz coexistence

[stevech]

As you may know, 40MHz mode in 2.4GHz is recommended by many people only for those who live in a rural setting. The 40MHz mode means your WiFi is using 2/3 of the entire 2.4GHz band, and you compete with neighbors' WiFi for air time.

 

[sfx2000]

As you may know, 40MHz mode in 2.4GHz is recommended by many people only for those who live in a rural setting. The 40MHz mode means your WiFi is using 2/3 of the entire 2.4GHz band, and you compete with neighbors' WiFi for air time.

I never recommend, nor use, wide channels in 2.4GHz - even for rural applications - remember that there's about a two to three dB loss going wide on Tx power for both the client and AP, and this has a negative impact on overall range, and performance at a given distance from the AP...

in my testing, it's shown to be basically a net-zero for the clients that could most benefit from wide channels, which is the low-end single stream 802.11n single band adapters..

 

[sfx2000]

And there's a lot of reasons for this - 20MHz was not an arbitrary number that IEEE just made up - there's sound physics in the analog/RF domain that indicated that was about as wide as one could go in 2.4GHz in a short range wireless scenario...

Much like 1.25MHz in IS95 CDMA for 850Mhz, interesting story here...

 

[stevech]

Agree.. but the WiFi Alliance seemingly pushed for 40MHz mode, which leads to a 6dB disadvantage over 20MHz mode due to the noise*bandwidth product (laws of physics). But worse, the wider banwidth on the receiver side invites far more interference from other 20MHz systems.

I thought that because of this, WiFi Alliance members agreed to a policy to disable 40MHz mode if any 20MHz SSIDs were detected by the router. Voluntary, not a spec. from the Alliance. And with a user chosen override of the policy.

 

[sfx2000]

Agree.. but the WiFi Alliance seemingly pushed for 40MHz mode, which leads to a 6dB disadvantage over 20MHz mode due to the noise*bandwidth product (laws of physics). But worse, the wider banwidth on the receiver side invites far more interference from other 20MHz systems.

Gahhh - going back to the WiFi wars of 802.11n

One has to consider the latter end of 802.11g where many vendors were pushing proprietary extensions outside of the spec...

Three camps - MIMO (Airgo), FrameAggregation (Broadcom), and WideChannel (Atheros), along with many other great contributions - 802.11 TGn allowed free and open contributions - and this is what took so long for 802.11n to get approved... and while I was working on 802.16, I had a chance to sit in on the discussions of 802.11 TGn while it was in development - and yes, it was ugly...

WiFi alliance, as an outside business partner with a commercial interest, needed a spec to ship... but WiFi Alliance needed IEEE 802.11 WG to finalize a spec...

That's why we have a very, erm, flexible spec with 802.11n and so many options... many which will never actually be deployed as they're not workable in the real world...

 

[thiggins]

20/40 MHz coexistence has been mandatory part of 802.11 since the N changes. References are in this article

http://www.smallnetbuilder.com/wireless/wireless-features/31743-bye-bye-40-mhz-mode-in-24-ghz-part-1

Wi-Fi Alliance Certification checks for this operation in OOB (Out of Box, i.e. default) mode. There is however, some loophole that allows manufacturers to include 20/40 Coexistence DISABLES. NETGEAR most notably does this.

I have asked how/why it is ok to disable 20/40 coexistence and never got a satisfactory answer.

 

[Razor512]

I wonder, for 4 stream routers, is it possible to use 60MHz of the 2.4GHz band, if not, then will they ever get around to allowing routers to simply take advantage of all 11 channels at the same time?

I feel that some routers offer the option to disable it (uncrippling the performance) in order to not do false advertising. Unless you are in a rural area, it is hard to fine an area where your router will automatically choose to use a 40MHz channel width, this means that for most people, they will never enjoy the full speed of their router.

Is it legal to market market a router as doing 450 Mbps on the 2.4GHz band if it would never actually achieve that PHY rate due to artificial restrictions?

What about in areas were the 2.4GHz band is already screwed? e.g., if I launch inssider V2, the counter quickly goes up to over 250 access points (95% of which are on the 2.4GHz band), 20/40 coexistence is not going to do much good here, the channels are already congested like crazy.

Beyond that, there are still many single stream WiFi radios in mobile devices (some allow for 40MHz mode), without 40MHz channel width, you will get poor performance (unable to take advantage of even the 3rd world country style of broadband that we have in the US).

 

[L&LD]

I wonder, for 4 stream routers, is it possible to use 60MHz of the 2.4GHz band, if not, then will they ever get around to allowing routers to simply take advantage of all 11 channels at the same time?

I feel that some routers offer the option to disable it (uncrippling the performance) in order to not do false advertising. Unless you are in a rural area, it is hard to fine an area where your router will automatically choose to use a 40MHz channel width, this means that for most people, they will never enjoy the full speed of their router.

Is it legal to market market a router as doing 450 Mbps on the 2.4GHz band if it would never actually achieve that PHY rate due to artificial restrictions?

What about in areas were the 2.4GHz band is already screwed? e.g., if I launch inssider V2, the counter quickly goes up to over 250 access points (95% of which are on the 2.4GHz band), 20/40 coexistence is not going to do much good here, the channels are already congested like crazy.

Beyond that, there are still many single stream WiFi radios in mobile devices (some allow for 40MHz mode), without 40MHz channel width, you will get poor performance (unable to take advantage of even the 3rd world country style of broadband that we have in the US).

It is legal. They don't promise those speeds. They state their hardware can attain them.

I advocate that my customers read 'ad speak' when comparing products with their 'specs'.

That is why I always recommend to test, in their own environment, the top two or three choices they are considering. Theory and practice do not always come together in the best interest of the consumer.

 

[stevech]

Most anything is legal (regulatory compliance), in the unlicensed band at 2.4GHz, as long as
a) you don't exceed the max radiated power for a given antenna per your region's regulatory authority
b) you don't "spill" power outside the band edges per the regulations
c) In 2.4GHz there are no regulations I know of about transmitter duty cycle, occupied bandwidth (in-band), etc., as there are in sub-GHz bands.

As to whether your transmissions are good, efficient, prudent, etc., at least the US FCC doesn't care. They may send a warning letter if lots and lots of written formal complaints are filed about who is creating intentional interference and name/address of offender. That's about it in the unlicensed bands.

Note too that licensed radio services share the 2.4GHz band, e.g., Radio Amateurs (hams) operating under their license. They can, but don't, radiate lots of power, legally in this shared band.

 

[Razor512]

It is legal. They don't promise those speeds. They state their hardware can attain them.

I advocate that my customers read 'ad speak' when comparing products with their 'specs'.

That is why I always recommend to test, in their own environment, the top two or three choices they are considering. Theory and practice do not always come together in the best interest of the consumer.

For the legality, I meant it more as if it is legal if the product cannot attain the advertised PHY rate due to a manufacture level crippling. For example, if nvidia tried to advertise their GTX 970 as having 2048 CUDA cores instead of 1664, it would be considered false advertising, even though the GTX 970 and the 980 have the same GPU, but some of the cores, ROPS, L2 cache sections, and crossbar ports are disabled. Just because the components are there, does not mean they can advertise the product as having them if they cannot be used by the user.


Another example, is the AMD Readeon R9 290, it has the same GPU as the 290x, but they never advertise it as having the same specs, even though the hardware is physically there, and with a bios hack, you can literally unlock the disabled cores and turn it into a 290x (since not all of the 290s are GPUs that were binned lower. The fact that some people can unlock their card does not mean that AMD can suddenly advertise the R9 290 as having better specs even though those who were able to unlock their cards, were able to get an instant 7-10% performance boost.

For a router, if you cannot cannot disable 20/40 coexistence, then it means that the company will be advertising functionality that probably 99% of the population would be unable to use even if they have the hardware to use it.

It would be like a company selling you a 3000 square foot, 4 bedroom house but 3 of the bedrooms are filled from floor to ceiling with cement. At that point, can they really advertise it as a 4 bedroom house?

 

[stevech]

Well, consider today's digital oscilloscopes... Most are sold with extra-cost upgrade options. But the product is sold with most in place but disabled. Some are firmware only functions that come alive after you pay and get a license key. Well, I can sort of see that.
But it's common that they sell a front-end bandwidth increase (say, from 100MHz to 200MHz). That's recurring cost hardware that is in the price but sits unused until/if you pay to enable it. So if you don't purchase the upgrade, it's gravy for them since the recovered all their recurring cost of sales in the original purchase minus options. If you do purchase the enabler, it's more gravy.

Where will this all go? Look to Tesla cars?

When did you say to your boss: If you paid me more, I'd be more valuable.

 

[L&LD]

For the legality, I meant it more as if it is legal if the product cannot attain the advertised PHY rate due to a manufacture level crippling. For example, if nvidia tried to advertise their GTX 970 as having 2048 CUDA cores instead of 1664, it would be considered false advertising, even though the GTX 970 and the 980 have the same GPU, but some of the cores, ROPS, L2 cache sections, and crossbar ports are disabled. Just because the components are there, does not mean they can advertise the product as having them if they cannot be used by the user.


Another example, is the AMD Readeon R9 290, it has the same GPU as the 290x, but they never advertise it as having the same specs, even though the hardware is physically there, and with a bios hack, you can literally unlock the disabled cores and turn it into a 290x (since not all of the 290s are GPUs that were binned lower. The fact that some people can unlock their card does not mean that AMD can suddenly advertise the R9 290 as having better specs even though those who were able to unlock their cards, were able to get an instant 7-10% performance boost.

For a router, if you cannot cannot disable 20/40 coexistence, then it means that the company will be advertising functionality that probably 99% of the population would be unable to use even if they have the hardware to use it.

It would be like a company selling you a 3000 square foot, 4 bedroom house but 3 of the bedrooms are filled from floor to ceiling with cement. At that point, can they really advertise it as a 4 bedroom house?

I think you're getting your analogies all mixed up.

The (router) hardware is capable of the speeds they advertise.

The (environment) the user may have is not guaranteed (ever).

A realtor selling a 3000SqFt home just needs to ensure the home is close to that spec. The number of 'bedrooms' is not (ever) guaranteed (you can knock a wall down to make a bigger room or add a wall and have an extra room, buyers choice).

A router doesn't need a certain amount of hardware to 'route'. Or a certain level of WiFi chips to be good at 'WiFi'. (Other than the bare minimum, of course). What it needs is the firmware to enable the hardware to achieve what is promised. And that is also never advertised (or guaranteed).

With your gpu examples, it is like saying why doesn't every other N900 class router perform like the RT-N66U? A lot of them are similar, if not identical hardware. The secret is in the 'sauce'.

 

[Razor512]

The main thing that I am getting at is can they advertise a feature or specification the vast majority of users are unable to use it due to intentional software crippling?

For example, in my extremely congested WiFi environment, I still see a speed improvement with forcing the router to use 40MHz mode.

For the house analogy, the idea is can ha house be advertised as a 3000SqFt, 4 bedroom home be advertised as such, if the company selling the house decided to fill 3 of the rooms (and about 1000SqFt)of space with cement (thus you are unable to use that space), is the potential existence of that space enough to not be false advertising?

The issue is not a router performing like other competing ones, it is the router not reaching its own full potential in its given environment due to software crippling below the advertised specs.

 

[pete y testing]

For example, in my extremely congested WiFi environment, I still see a speed improvement with forcing the router to use 40MHz mode.

yes but screw everyone else in the process

it is the router not reaching its own full potential in its given environment due to software crippling below the advertised specs.

incorrect its the router complying by the set standard in congested areas , the router is fully capable of the advertised sync speeds if the correct conditions apply , if however its in a congested area the software if forced to reduce to 20mhz and whats called neighbour friendly mode

by forcing 40mhz if the router has forced co existence you are just sticking your finger up at the neighbours and saying i will do what i want no matter the consequences and who it effects , but hell who cares about them anyway

 

[L&LD]

The main thing that I am getting at is can they advertise a feature or specification the vast majority of users are unable to use it due to intentional software crippling?

For example, in my extremely congested WiFi environment, I still see a speed improvement with forcing the router to use 40MHz mode.

For the house analogy, the idea is can ha house be advertised as a 3000SqFt, 4 bedroom home be advertised as such, if the company selling the house decided to fill 3 of the rooms (and about 1000SqFt)of space with cement (thus you are unable to use that space), is the potential existence of that space enough to not be false advertising?

The issue is not a router performing like other competing ones, it is the router not reaching its own full potential in its given environment due to software crippling below the advertised specs.

pete y testing has answered you already. The answer is still the same; no (not false advertising).

The analogy with the house doesn't make sense either. No builder will make a much bigger foundation, roof, walls, etc. and then spend even more money (on concrete) to fill 3 of the rooms with in a way to 'get the customer'.

What you are seeing as limitations for a single user (yourself) is really to protect the WiFi experience of the many who use and rely on wireless routers in general. If your specific and immediate environment allows you to enjoy the full specifications of the equipment you bought (and yes; the hardware will allow that where possible), then all is good. If however, you find yourself at the opposite end of that spectrum, it will adjust to allow all concerned (not just 'you') to have a minimal level of wifi enjoyment.

WiFi is a shared commodity. It isn't something you buy and have irrefutable rights to.

 

[Razor512]

I currently do not use 40MHz mode, since everything that needs good performance, is on the 5GHz band now. The 2.4gHz band here keeps getting worst since there are a large number of APs, and people are using it for everything, even when they don't need to.
It is increasingly common to see completely illogical network setups, e.g., someone has a desktop PC that is about 4-6 feet away from their router, and they are using WiFi to connect to it instead of Ethernet.

For the 20/40 coexistence, I see it more as a feature to get a good burst of performance when you need it. 40MHz mode does have its trade-off, for most routers, the transmit power often drops by upwards of 50% when in 40Mhz mode, thus reducing range, but for the range that it does cover well, you get a good speed boost.

 

[sfx2000]

For the 20/40 coexistence, I see it more as a feature to get a good burst of performance when you need it. 40MHz mode does have its trade-off, for most routers, the transmit power often drops by upwards of 50% when in 40Mhz mode, thus reducing range, but for the range that it does cover well, you get a good speed boost.

In real-world - more radios, more streams are much more important that wider channels if you have a mix of clients in 2.4GHz that cannot do wide channels, or in a mixed G/N environment...

The Tx power hit alone on 40MHz channels impacts the very devices that can ill afford to lose any power, mostly B/G/N capable tablets and mobiles, which are pretty low power as it is...

The constructive gain given by MIMO is about 3dB per spatial stream from the AP - even if a client is single stream, the AP sends the same streams across all radios... and each radio provides about 3dB of diversity gain on Rx...

Best possible case - these recent 4*4:4 AP's, where in narrow channels, they can go up to 290Mbps, and single stream clients will benefit more from that AP, than a 2*2:2 in Wide Channels at a given RSSI... actually any client will...

 

[stevech]

40MHz in 2.4GHz WiFi is .. like .. the 500Watt amps and 18 inch speakers in the car 100 feet from yours.

But if that amp'd car is rarely driven, well OK.

 

[sfx2000]

by forcing 40mhz if the router has forced co existence you are just sticking your finger up at the neighbours and saying i will do what i want no matter the consequences and who it effects , but hell who cares about them anyway

Yes - and for those that run multiple AP's (present company included) - park them all on the same channel in 2.4Ghz - only three channels out there, leave the other 2/3's for the neighbours...

 

[sfx2000]

40MHz in 2.4GHz WiFi is .. like .. the 500Watt amps and 18 inch speakers in the car 100 feet from yours.

But if that amp'd car is rarely driven, well OK.

I was just thinking... I'm using about 220Mhz of spectrum across two bands - and getting a collective 4758 Mbps of bandwidth..

Gotta love MIMO