Asus USB-AC56 turboQAM

[RogerSC]

Got an Asus USB-AC56 adapter a while ago. Have had some problems with it on Windows 8.1, didn't fully connect on reboot most times. Once it's connected it's fine, always connects at 867Mbps on 5GHz. when it full connects. So I've been watching for a newer driver for Windows 8.1 to get rid of the annoying problem of a "limited" connection at reboot.

Just the other day, Asus posted a new driver for Windows 8.1, which has fixed the "limited" connection at reboot time. Now the adapter always is connected at 867Mbps.

The release note for the driver mentions that turboQAM has been added. So I figured I'd try it out and see if it did anything. Without turboQAM, the wireless adapter connects at 300Mbps on 2.4GHz. with 40MHz. channel width, 2 streams. With turboQAM enabled in dd-wrt on the R7000, it connects at 400Mbps., about what one would expect on a per-stream basis So this would seem to add about 1/3 (33%) to the connect speed, interesting. Not having seen turboQAM actually working before, I had my doubts about it.

Just for fun, tried it with 20MHz. channel width on 2.4GHz. The adapter connected at 144Mbps. Then enabled turboQAM, and it ramped all the way up to 173Mbps *smile*. Less interesting, but it surprised me a bit to see that, for some reason.

I haven't measured actual throughput with this adapter yet, but I'm impressed that the much marketed turboQAM actually does something, even if only raising the connection speed. I don't use this adapter on 2.4GHz., only on 5GHz. where I can use wireless-AC, so turboQAM doesn't matter to me, just interesting.

 

[stevech]

turboQAM requires ASUS-specific hardware on both ends, right? It's proprietary.

We had that movement many years ago - early days of 11n - proprietary modes and least common denominator speeds for cross-vendor interoperability. I thought it had gone away.

 

[RogerSC]

The turboQAM support in the dd-wrt firmware for the Netgear R7000 seems to work fine with the Asus wireless adapter. So seems as though turboQAM has become an "industry standard" of a sort.

 

[RMerlin]

turboQAM requires ASUS-specific hardware on both ends, right? It's proprietary.

Broadcom proprietary, so most likely other routers based on Broadcom's Northstar platform (such as the R7000) would also be compatible.

 

[stevech]

Remember Atheros' Super-G ?
802.11g
http://en.wikipedia.org/wiki/Super_G_(wireless_networking)

and even turbo-g
http://www.amazon.com/dp/B00009ETKX/?tag=snbforums-20

 

[RogerSC]

Remember Atheros' Super-G ?
802.11g
http://en.wikipedia.org/wiki/Super_G_(wireless_networking)

and even turbo-g
http://www.amazon.com/dp/B00009ETKX/?tag=snbforums-20

I remember the Linksys WRT54GS, which required a Linksys ethernet network card to get that smokin' 108Mbps instead of normal wireless-g 54Mbps. That takes me back. While I had a WRT54GS router, I never had it together to get the 108Mbps going *smile*.

And I forgot that turboQAM was a Broadcom thing, never thought that I'd actually see it here at my place *smile*. I was surprised that the Asus USB-AC56 supports it, since it has a Realtek chipset in it.

 

[thiggins]

The turboQAM support in the dd-wrt firmware for the Netgear R7000 seems to work fine with the Asus wireless adapter. So seems as though turboQAM has become an "industry standard" of a sort.

Yes, it has. Pretty much all new AC MAC/PHYs support it.

Remember, you need a 3x3 adapter and 40 MHz bandwidth mode to get 600 Mbps. It is primarily a marketing ploy.

 

[azazel1024]

turboQAM requires ASUS-specific hardware on both ends, right? It's proprietary.

We had that movement many years ago - early days of 11n - proprietary modes and least common denominator speeds for cross-vendor interoperability. I thought it had gone away.

From everything I have seen, no. It only requires another router that has turboQAM, which is any AC1900 router right now. I suspect there will be some single and 2 stream 11ac routers that also add turboQAM too.

TurboQAM does have the advantage that it seems like it is LIKELY, though not a certainty, that it will eventually be a very common feature on 11ac chipsets from all/most manufacturers. This is unlike the past with some 11g and 11n manufacturer specific "speed up" features.

 

[azazel1024]

Yes, it has. Pretty much all new AC MAC/PHYs support it.

Remember, you need a 3x3 adapter and 40 MHz bandwidth mode to get 600 Mbps. It is primarily a marketing ploy.

ish.

I say ish, because it will add roughly a 30% increase in throughput with a good 2.4GHz connection no matter how many streams or bandwidth you add due to the increase character set encoding.

I've seen very little testing of various 256QAM 2.4GHz setups, but in theory if it is working properly, it should show a fair good increase in actual throughput. The two chief improvements in 11ac on 5GHz is the increase in bandwidth from 40MHz to 80MHz allowable (and 160MHz later), but also the increase from 64QAM to 256QAM. The change from 40 to 80MHz should only increase the net yield from around 150Mbps per stream to around 325Mbps per stream...the rest of the increase to 433Mbps per stream that 11ac supports is from the increase in the character encoding from 64QAM to 256QAM. Yeah...that is only a rather low interferance and high strength connection, but it is what increases the net ideal yield by an additional 33%.

So, same with 2.4GHz...on a good 2.4GHz connection, the change should increase the net yield by roughly 33%.

I'd like to see more testing across a number of chipsets/manufacturers/drivers at varying distances and test conditions to see how much it bears out in actual use, but a lot of chipsets can handle full 64QAM encoding down to around -70dBm signal strength. I know 256QAM requires an even stronger connection, but I'd bet a monkey it might be able to maintain it down to around -62 to -65dBm or so...which in practice might mean 2-3 hollow (2x4 and drywall) walls between you and the router and maybe 20ft of distance, at least with 2.4GHz.

Odds are excellent that 5GHz 11ac 80MHz is going to provide higher throughput than turboQAM 2.4GHz 40MHz will, by virtue of the doubled bandwidth, even at moderate distances...but until someone has done a lot of testing across a number of devices, I think it is unfair to call it a marketing gimmick or worthless.

 

[thiggins]

First, I meant you need a 3x3 adapter that supports 256 QAM and 40 MHz bandwidth mode in 2.4 GHz to get 600 Mbps link rate. 256QAM has been supported in 5 GHz for AC from the get-go.

In 20 MHz mode, a 3x3 client will get a maximum link rate of 288.9 Mbps with 256 QAM, which, as you noted, is 33% higher than the 216.7 Mbps shown in the MCS table (shown here) for 64 QAM.

I agree that with strong to medium signal strength, adapters that support 256QAM in 2.4 GHz will provide a bit more throughput, but I doubt it will be 33% more. But, as you note, testing will tell.

I still maintain AC1900 is a marketing ploy (I didn't say worthless) until there are more client devices that support 256QAM in 2.4 GHz. The main advantage of AC1900 routers at this point is their faster processors, which helps boost 5 GHz performance.

 

[azazel1024]

Can't quibble with that.

Though, I'd imagine a decent chipset probably would support roughly 33% faster speeds with 256QAM versus 64QAM. I've seen plenty of good chipsets in low noise environments which are able to push roughly 60% payload yield on 11n and 11ac, which is roughly the theoretical maximum once overhead (error correction and TCP/IP overhead) is taken in to account. I don't see a reason why a good 256QAM 2.4GHz supporting chipset wouldn't likely see the speed up promised.

Emphasis on good. Also there should be a bit inserted on a GOOD CONNECTION.

I have pretty ideal conditions in my home for 2.4GHz. I have no problems getting 170-180Mbps on my 300Mbps connection near my router on 2.4GHz. I'd probably see more like 210-230Mbps if that connection could use 256QAM and both ends had good chipsets.

All theoretical and MOST people aren't nearly as lucky as I am with low noise environments and lack of competiting networks near them.

My old network when it was in my townhouse with exact same router and network card had trouble pushing 130Mbps...but a lot of competing networks and different build on my townhouse.

Wide spread support for 256QAM is needed though before it is too much more than a marketing gimmick...but I think that will come with time. I could be 100% wrong, but I suspect it will slowly become a standard feature of 11ac chipsets over the next couple of years. I at least don't see a reason why it shouldn't. It shouldn't be a whole lot more than a firmware change in most cases so the "overhead" for adding it to 11ac chipsets should be little more than validation costs.

 

[thiggins]

Wide spread support for 256QAM is needed though before it is too much more than a marketing gimmick...but I think that will come with time. I could be 100% wrong, but I suspect it will slowly become a standard feature of 11ac chipsets over the next couple of years.

256QAM already is standard on AC MAC/PHY. AC1900 routers simply use two of them vs. an N MAC/PHY on the 2.4 GHz side in AC1750 routers.

The question is whether manfs will create 2x2 products with 2.4 GHz 256QAM. The ol' MCS table says the 2x2 40 MHz B/W 300 Mbps max link rate would become 400 Mbps with 256QAM.

So when we see AC1267 routers appear, then we'll know 2.4 GHz 256QAM is going mainstream!

 

[RogerSC]

256QAM already is standard on AC MAC/PHY. AC1900 routers simply use two of them vs. an N MAC/PHY on the 2.4 GHz side in AC1750 routers.

The question is whether manfs will create 2x2 products with 2.4 GHz 256QAM. The ol' MCS table says the 2x2 40 MHz B/W 300 Mbps max link rate would become 400 Mbps with 256QAM.

So when we see AC1267 routers appear, then we'll know 2.4 GHz 256QAM is going mainstream!

Yes, exactly what I noted above with the Asus USB-AC56 *smile*, 400Mbps was what I saw on this one.

Of course, for me, the 867Mbps on 5GHz. is much more interesting *smile*.

They could call it a wireless-AC1267 (or wireless-AC1300 if marketing holds true to form *smile*) adapter. They're missing a bet at this point by calling it "AC1200", not like those marketing guys, must be having an off day.

I'll be watching to see if they up the speed designation on this wireless adapter *smile*.

 

[sfx2000]

Rang up a friend of mine a couple of weeks back and got a half-hour of bench time in his lab on a Saturday morning...

Tried two boards, one was Broadcom based, the other was Marvell...

QAM256 in 2.4GHz provided 25 percent improvement on both chipsets on the connected PHY data rates on an Agilent test box - on a clean signal, but as soon as we started adding some AWGN in, both fell back to QAM64...

The improvement isn't trivial - and the modulation scheme is baked into the chipsets, at least with the current Broadcom and Marvell chips - just need to flip the bits in the driver, so it's basically free to implement on the AP side...

That being said - on the low-end single band router tier - don't be surprised to see single band products with N600 on the box -- it's just too tempting for the business guys not to do it, esp. now that Wave1 AC silicon has come down in price... (at a radio/chipset level, they're almost at parity now)

it's just driver settings in the chipset firmware.

Goes without saying though that the STA also has to support QAM256, otherwise they negotiate QAM64...