Posted: Wed Aug 27, 2014 19:08 Post subject: VHT 80+80 MHz
I've been doing some reading (which, for those who know me, is dangerous). Trying to learn more about VHT 80+80 MHz channel widths and which channels get used for secondary channels, etc. Came across this rather technical deep-dive into how 802.11 works with the AC standard.
So, basically, I did what I've read most people do when they first setup their R7000 for wireless. I cranked all of the settings up, picked an open channel after doing a quick survey, and called it a day. This included picking VHT 80+80 MHz. After all, if I don't pick the highest setting, I'm not getting the full potential from my hardware right? ...right?
After reading that article, I'm left with a few questions (initially). As I understand it right now, I'm using 160 MHz channel width. However, I also chose channel 149 (I've gone back and forth between 149 lower lower and 161 upper upper to compare interference with neighbors, etc.) Here are the actual questions:
- Without getting into DFS, etc. in the US, I would think it wouldn't be possible to choose 80+80 MHz and 149/161 because there aren't enough channels in the 5.734-5.825 GHz range to support it (149-165). It would appear the only possible range for 80+80 MHz is channels 36-64 (5.17-5.33 GHz). Is DD-WRT/the Broadcom driver just realizing I'm dumb and scaling things back automatically for me, etc.? See this chart to understand what I'm getting at better (usable channels in black with gray proposed by FCC):
- It seems like there's lots of factors that play into this, but it appears that with three spatial streams (is that basically # of antennas?), you hit the maximum speed (advertised by Netgear) of 1.3 Gbps at 80 MHz. At 160 MHz, a three-stream device falls back to MC8 because MC9 isn't supported, so you get a lower value of 2.8 Gbps. I guess the actual question is does the R7000 really even support 160 MHz or does Netgear just dumb down the marketed speed because they push the use of the higher channels?
See the chart under the section called "Comparison of 802.11ac Data Rates to Other 802.11 PHYs" at the first link above to see one example of what I found suggesting this.
- The ultimate question out of all of this is how are my current settings working? Should I change to 80 MHz or whatever the lower setting is called (I'm about to VPN in and check just to sate my curiosity)?
Edit: So I connected just to check the available settings. There is no option for just 80 MHz? There's 20, 20+20 (40), and 80+80 (160)... Sigh. More confusion, unless 80+80 does as I suspect and drops down to 80 if not enough room for channels, etc. This seems to go along with the proper behavior of trying 80 MHz, 40 MHz, and 20 MHz to avoid interference with neighboring networks.
Also, even the first 160 MHz range overlaps with DFS channels (52-64). Haven't found anything willing to discuss contiguous vs. non-contiguous 80+80 MHz channel widths yet. Something like 36+40+44+48+149+153+157+161 would be tempting. This was another good page that I found though.
Finally, I don't think the loss in transmit power (50 mW/250 mW, down from 1 W at top four) in the lower channels is worth the potential increase in speed. On the other hand, no one uses them (or their range is very limited) so the odds of avoiding interference in an apartment building is good!
I never got why the label says 80+80, since that implies 160MHz. NONE of the routers on the market are capable of 160MHz, not even the new RT87U--they all use 80. Anyway, I just assumed it's a labeling error and never thought much of it. _________________ Buffalo WZR-1750DHP: 34311
TRENDnet TEW-673GRU: 34311
TRENDnet TEW-811DRU: 33986
So just to follow up on this, I've had some hands-on time messing with this now.
First of all, at least for the 5 GHz devices I have, DFS/TPC support is spotty. Some worked with some settings while the network disappeared completely for others. It also didn't help that there appears to be a delay (by design) when it tries to make sure the channel is safe to use before allowing any connections. I tried various combinations of the regulations options at the top for 802.11h and d. I also tried different mitigation factors. I understand that to mean how much it can tell the power to dial down if requested, but I don't understand what the actual values of 2, 3, or 4 mean without some context. Is that dB, percentage, etc.? I've seen it suggested that a factor of 8 is a common value to use (dropping by 9 dB or dividing in half three times so 3-3-3 dB or 2 divided by 2 divided by 2). In the US, it is required to drop at least 6 dB or to always run as a low as 27 dB (thus getting around the dynamic power change requirement).
Anyways, I'm done with those shenanigans, so I've basically ruled out using DFS channels at this point (I never did see it use 160 MHz wide channels). This limits me to the four lower channels and four upper channels. The power is supposed to drop from 1000 mW to 50 mW for the lower ("indoor") channels, and let me tell you, the range is pretty pathetic as a result. I have a one bedroom apartment, and my signal dropped to 50% going two rooms over. It usually gets 80% on the opposite end of the apartment from the router in the higher channel range. Overall, I saw about a 10% drop in signal quality for all devices on the Info page. I didn't mess with manually increasing the power, as I figure they have their reasons for picking the things they did when setting the rules, etc.
So, in effect, I'm stuck with the top four channels with three other networks, two of which feel like they're right on top of me (quite literally I suspect). At 80 MHz, I'm hogging all four channels, while each of them have picked two channels to squat on with 40 MHz each (one of which came along after I had mine up and running... sigh). At this point, I'd rather put up with mild interference and get the AC speeds than to drop down to 40 or even 20 and still get interference from my lovely neighbors. I'm sure it is only a matter of time though before this gets worse as more people upgrade their networks. What is really getting bad is the local ISPs are just giving everybody wireless networks. They're all starting to overlap now on the 2.4 GHz side (I counted about 15 networks last night).
As for there being no 160 MHz devices, I figured at least the R8000 would, right? It looks like it is advertising the kind of speeds that I saw for 160 MHz networks, though they may be doing what I was saying about the non-contiguous channels and combining the lower and upper channels to get around not using DFS channels like they do in stock on the R7000. And that's just what we need, networks that basically interfere with every channel possible out of the box. The fact that they turn on 40 MHz for 2.4 GHz on the R7000 is pretty atrocious as it is.
I repeat, there are no 160MHz devices at this time.
160MHz is a Wave 2 feature. The R8000 isn't a Wave 2 device. It's a device that clobbers together two separate Wave 1 radios on 5GHz. These are two separate, distinct, independent 80MHz radios.
The only Wave 2 device on the market, the RT-AC87U, does not do 160MHz, either.
AFAIK, the label should say VHT 80 or VHT 40+40, and that VHT 80+80 is likely just a bad label.
You'll also notice that there's no way to specify where you want that other 80MHz. The channel selection works by selecting the primary 20MHz within the 80MHz band that you want, and then telling it where in that 80 your 20 is. So primary channel 36 with lower lower means that you are using the 80MHz channel centered around 42 (i.e., 36 through 48), with the primary 20 in the lower 20 of the lower 40 (which is channel 36). And 44 with upper lower means that you're using the same 80MHz centered around 42, but now your primary is in the upper 40, lower 20, which is 44.
You'll notice that this selection scheme only works for 80MHz channels. For 160MHz, you'll need an additional parameter, esp. since 160MHz also allows for the other 80 to be disjoint.
So the UI doesn't allow for 160. The chipsets most definitely do not support 160. There is no 160. Period. _________________ Buffalo WZR-1750DHP: 34311
TRENDnet TEW-673GRU: 34311
TRENDnet TEW-811DRU: 33986
Yeah, that's certainly been what I've been seeing, though it is nice to hear agreement. I've been reading wave 2 AC devices aren't really a thing until 2015, but I really wanted to believe that Netgear wouldn't just slap two 80 MHz interfaces together and call it a day. Sigh.