Intel 7260 2x2 802.11ac question regarding beamforming

Hi all,

I've searched this forum in order to check if this had been asked before or if there's a thread for this NIC already, but it doesn't seem to be the case.

Does anybody know if the 802.11ac Intel 7260 supports beamforming? I know that it is officially part of the 802.11ac standard, but considering that the NIC is from 2013(?) and that following the official ac-spec does not seem to be very common among NIC and router vendors, I'm wondering whether it does.

The background is that I've ordered a new laptop that comes with this NIC built in and I'd like to know what to expect...

Thanks!

The Intel 7260 is kind of tricky because Intel used that number on three products:
Intel® Wireless-N 7260
[URL='http://ark.intel.com/products/75440/Intel-Dual-Band-Wireless-N-7260']Intel® Dual Band Wireless-N 7260[/URL]
[URL='http://ark.intel.com/products/75440/Intel-Dual-Band-Wireless-N-7260'][URL='http://ark.intel.com/products/75439/Intel-Dual-Band-Wireless-AC-7260']Intel® Dual Band Wireless-AC 7260[/URL][/URL]

Looking at the data sheet for the AC 7260 shows no sign of beamforming.
But the Wi-Fi Certification (if I have the right one) shows certification for
"Tx SU beamformee". This means it supports beamforming.

thiggins said:

The Intel 7260 is kind of tricky because Intel used that number on three products:
Intel® Wireless-N 7260
Intel® Dual Band Wireless-N 7260
Intel® Dual Band Wireless-AC 7260

Looking at the data sheet for the AC 7260 shows no sign of beamforming.
But the Wi-Fi Certification (if I have the right one) shows certification for
"Tx SU beamformee". This means it supports beamforming.

Click to expand...

Wow, thanks a lot, I wouldn't have figured out how to retrieve that information myself!

Very good news, so I'm happy that it will take full advantage of my router's capabilities

Don't expect miracles from beamforming. It doesn't increase range. It mainly provides a moderate throughput boost under mid-level signal conditions.

thiggins said:

Don't expect miracles from beamforming. It doesn't increase range. It mainly provides a moderate throughput boost under mid-level signal conditions.

Click to expand...

Yeah, it's more a matter of principle in my case (basically from a nerd point of view ). If I have a device that offers a certain functionality (in this case Beamforming), I like the fact that the client that I'm using can actually take advantage of it. Otherwise it'd feel a bit like a waste, despite the tiny difference it'd actually make when it comes to real-world performance.

Beamforming, done well, has helped cellular radio systems... mostly at cell sites near stadiums where the subscriber density goes way up at times.
It was an option in IEEE 802.16 mobile WiMax, and I worked with Lucent's very early implementation. WiMax gave way to LTE before WiMax got it to work.
Done right, it's too expensive for consumer and SMB class WiFi.

If you have a newer enough laptop to support it, get the 7265ac. It is a fair amount faster than the older 7260ac. Downside is, they ONLY make it with m.2 connectors, they don't make mini PCIe connectors. So generally only stuff from the last ~1 or maybe 2 years can support the newer Intel 802.11ac card.

Or I am sure you could get a desktop m.2 to PCIe adapter.

PS Unless I have missed something, support for beam forming is REQUIRED in the 802.11ac spec. That does not mean that the device has to perform beam forming itself, but it does have to support the management frames that allow the base station (if the base station has it) to perform explicit beam forming.

Looks like they make an mini PCIe adapter for the 7265ac card

http://www.amazon.com/dp/B00JGR3ULO/?tag=snbforums-20

Will the speed difference apply to using just N mode?

No. the 7260ac already maxes out 802.11n spectral and protocol efficiency. I mean, it might be a fraction faster at medium to long range, but it certainly can't be any faster at short range.

Intel claims (I have seen it in several spots), and the little anecdotal evidence I have seen backs up their claims, that the 7265ac is 20% faster than the 7260ac in actual performance. In a lot of testing I have done with the 7260ac, in 11ac mode you are looking at an absolute maximum speed of about 61MB/sec down and 48MB/sec up at close range and a good base station. That works out to 56% and 44% efficiency respectively. The 7265ac appears to hit closer to 72MB/sec and 60MB/sec respectively, which is a little less and a little more than a 20% gain (respectively), or about 66% and 55% efficiency. 802.11ac IIRC has a theoretical maximum efficiency of about 84%, where as 802.11n was/is about 76% efficiency. A difference in their error correction mechanisms, 802.11n uses FEC (Forward Error Correction) and 802.11ac uses LPEC (Low Parity Error Correction). One uses roughly 24% of the coding for error correction and the other approximately 16%.

The 7260ac already hits 28.5MB/sec down and 27MB/sec up on 2.4GHz in a lot of testing I have done (both 2.4GHz and 5GHz forced in to 11n mode). That is 76% and 73% efficient respectively, so I can't see the 7265ac beating that.

*edit* To add, by looking at the numbers, even the newer 7265ac has a fair distance to go before 802.11ac is actually maxed out. By comparisons, through a wall and about 15ft of distance a wireless bridge used from an AC1900 range extender I tested to my AC1200 router managed 74MB/sec down and 69MB/sec up. I didn't test in the same room and closer, but I'd bet almost anything it would have hit over 80MB/sec if it had been. If you also look at 802.11n, it wasn't significantly better when it was first introduced. Early draft n adapters and routers typically hit in the high 50% range for efficiency and it took a good 6-8 years before they inched up to maxing out efficiency for the better adapters and routers.

Intel Wireless-AC 7260 is pure junk i have that adapter in Lenovo ThinkPad T440p... So many problems just check intel forum about it.. i would avoid it at all cost.

ilukeberry said:

Intel Wireless-AC 7260 is pure junk i have that adapter in Lenovo ThinkPad T440p... So many problems just check intel forum about it.. i would avoid it at all cost.

Click to expand...

It's just a matter of driver. Mine worked flawlessly for the 12-18 months I had it in my Asus laptop, and I have no problem either with the 7265AC in my current Thinkpad.

RMerlin said:

It's just a matter of driver. Mine worked flawlessly for the 12-18 months I had it in my Asus laptop, and I have no problem either with the 7265AC in my current Thinkpad.

Click to expand...

The Intel wireless NIC's work well, but tend to be very sensitive to drivers...

Haggster said:

Does anybody know if the 802.11ac Intel 7260 supports beamforming? I know that it is officially part of the 802.11ac standard, but considering that the NIC is from 2013(?) and that following the official ac-spec does not seem to be very common among NIC and router vendors, I'm wondering whether it does.

Click to expand...

Intel NIC's do support 802.11ac Tx SU Beamforming as a client (beamformee)...

I've seen wireless packet captures where it responds correctly if/when the Router/AP sends out the request (Beamforming in 11ac is initiated by the AP, and controlled by the AP, the clients can respond with information that the AP uses to build the beam forming matrix)

Because they hold such a volume lock on the client market, does anyone have a sign on when (or if) Intel will support mu-mimo? I can't find it anywhere in the Intel marketing machine so of course that certainly means "not now" and the 8260 is the upcoming product that if it doesn't have it, probably rules out options for a year.
My mind is thinking that if you don't have their chips support mu-mimo then the standard is pretty defunct? This may end up spawning a different thread. I felt it was the same chip family so I started here.

When i was testing wireless AC to get above 80% link rate i used this adapter. Driver wise there are some issues for laptops such as if your laptop sleeps/hibernates when connected to an AP it has issues connecting to that same AP again till you disable and reenable it. To get above 80% link rate use and packet losses you need a laptop with a fast CPU and a wired server also with a fast CPU and you would need to disable all the windows networking features or dont use windows to test it and than stuff the line full of wire to wireless transfers.

Dave Fey said:

Because they hold such a volume lock on the client market, does anyone have a sign on when (or if) Intel will support mu-mimo? I can't find it anywhere in the Intel marketing machine so of course that certainly means "not now" and the 8260 is the upcoming product that if it doesn't have it, probably rules out options for a year.

Click to expand...

Perhaps unlikely that Intel will support MU-MIMO on laptops as there's little use for it in the Laptop use case...

sfx2000 said:

Perhaps unlikely that Intel will support MU-MIMO on laptops as there's little use for it in the Laptop use case...

Click to expand...

Say some more about this, please. As I think you pointed out awhile ago, MU STA designs are moving to 2x2. So why not MU 2x2 in laptops?

Dave Fey said:

Because they hold such a volume lock on the client market, does anyone have a sign on when (or if) Intel will support mu-mimo?

Click to expand...

Do you have any market share data to back this up?

Fair enough. Not really. That's an opinion based on a lot of market hunting for laptops, which of course do not represent the entire body wifi clients. Thought there would be some SEC revealed info in filings on just what some of the makers believe their share is, but of course that is another area held tight to the vest. So nope I've got nothing worthy to present. Sorry

sfx2000 said:

Perhaps unlikely that Intel will support MU-MIMO on laptops as there's little use for it in the Laptop use case...

Click to expand...

My thought on this is that with Wifi "useage" pretty much being huge to the point of continuously hunting for more spectrum, wouldn't a standards based technology that helps better utilize existing spectrum be a desirable feature to promote? I get that those ideas that are proprietary and step outside of the standards body are not big wins but I was thinking that within the spec would be more likely to be jointly promoted.