R7800 SmallNetBuilder Review

[sfx2000]

General rule of thumb is to reference the 20MHz primary channel - as not all 11ac network run 80MHz - many in the corp and hospitality areas run 40MHz channels to conserve spectrum and leverage into their already existing frequency plan developed for 802.11n..

 

[pege63]

Yes thats true, but i was talking about SOHO routers for home users.
Most small office/home office routers will reference an 802.11ac channel by it's primary 20 MHz channel instead of the correct 802.11ac channel number.
Here is an example:

This gives 1, 160MHz, 1, 80MHz, 3, 40MHz, 1, 20MHz or 4, 40MHz options for you in US (last one 140-144) on the 5GHz.
Or (3, "80MHz" and 1, "160MHz"), ( 2, "160MHz and 1, "80MHz"), (5, "80MHz"), (10, "40MHz") for you in US.

The primary 20 MHz channel can be any 20 MHz channel that falls under a wider channel above it. For example; possible Primary 20 MHz channels for 80 MHz Center channel 42 are:

· 36
· 40
· 44
· 48

Any one of those channels can be designated as the primary, you can see this here in this post.
But the Non-overlapping P channel is 36 for 80MHz.

So to show the right NEW convention CH, you have to use the OLD convention CH at 20MHz width in the GUI for 40/80/160MHz width in the present FW.

So CH 42 for the 80MHz in the NEW convention is the same as CH 36-48 i the OLD
inconvention and the P non-overlapping CH is 36, in this example. (ch 36=36-48)

So in reallity it should just show the right P 20 MHz channels on each 20/40/80/160 MHz band and not the full range of channels from 36-144, when you choose a channel.

Wide 80MHz and 160MHz channels improve throughput but only when full channel bandwidth is free from interfering transmissions.

You designate primary 20 MHz channels so that it results in non-overlapping 40 MHz channels (36, 44, 52, 60, 100, 108, 116, 124, 132, 140, 149, 157). If youre in the US you cant use 40 MHz channels 118 and 126 "TDWR restrictions", so this results in 10 non-overlapping channels. If you're in the UK/EU you cant use 40 MHz channels 151 and 159 "Band C licensing", so this also results in 10 non-overlapping channels.

You designate primary 20 MHz channels that result in non-overlapping 80 MHz channels (36, 52, 100, 116, 132, 149). If youre in the US you cant use 80 MHz channel 122 "TDWR restrictions", so this results in 5 non-overlapping channels. If you're in the UK/EU you cant use 80 MHz channels 155 "Band C licensing", so this also results in 5 non-overlapping channels.

Also to know that its a 4x4 11ac 80MHz, 2x2 160/80+80MHz router.

Sources; revolutionwifi.net, metageek.com, 802.11ac: A Survival Guide: Wi-Fi at Gigabit and Beyond 1st Edition by Matthew S. Gast.

802.11ac Channel Planning
802.11ac Channel Widths are Flashy but not Efficient
Safely using 80 MHz channels

 

[bodean]

If one is using true VHT160 (not 80+80), one is bound to run into DFS space, at least in North America... because 52-64 is squarely in that range...

So my best guess is that the UI consider the VHT80 plan, where the "channel" is the 20MHz control channel for A/N/AC compat, running 80MHz free, and the 52-64 is conditional if DFS passes...

View attachment 8077

So, if I am reading this correct, if you run channel 149 for 5Ghz, you will not achieve 160MHZ? So you have to either run a channel in the 30's or 40's, or a DFS channel?

 

[sfx2000]

So, if I am reading this correct, if you run channel 149 for 5Ghz, you will not achieve 160MHZ? So you have to either run a channel in the 30's or 40's, or a DFS channel?

Some might be able to do 80+80 and bridge the gap - but in NA, there doesn't seem to be a way at the moment...

 

[MartinW]

On a completely different subject. Please be aware that the WAN 2 LAN performance chart is nowhere near the reality. With the stock F/W i get about 610 Mbit and with DD-WRT 830Mbit. This is not near the 941 stated in the test and not good enough for my 1Gbit fibre !

Good to know if you have a fast connection.

 

[xeqw]

On a completely different subject. Please be aware that the WAN 2 LAN performance chart is nowhere near the reality. With the stock F/W i get about 610 Mbit and with DD-WRT 830Mbit. This is not near the 941 stated in the test and not good enough for my 1Gbit fibre !

Good to know if you have a fast connection.

That is odd, DD-WRT has no hw nat acceleration, how come it is faster than stock?

 

[thiggins]

How fast do you think a gigabit Ethernet TCP/IP connection goes?

What does your ISP ToS spec for speed?

 

[MartinW]

How fast do you think a gigabit Ethernet TCP/IP connection goes?

What does your ISP ToS spec for speed?

I know for a fact that if I connect with a DIR 879 to the same line i Get 960Mbit (at night), With my Edgerouter X I get 930Mbit. SO there is more to have than Netgear gives me, my R7800 is now tranformed into my most expensive AP ever.

 

[Voxel]

I know for a fact that if I connect with a DIR 879 to the same line i Get 960Mbit (at night), With my Edgerouter X I get 930Mbit. SO there is more to have than Netgear gives me, my R7800 is now tranformed into my most expensive AP ever.

I have read opposite feedback re: stock vs DD-WRT speed with gigabit line, DD-WRT forum:

http://www.dd-wrt.com/phpBB2/viewtopic.php?t=289788&postdays=0&postorder=asc&start=690

See user raymod2 feedback:

R7800 with stock firmware cabled: 949/946 Mbps (down/up)
R7800 with DD-WRT cabled: 543/562 Mbps

(The speed on DD-WRT was highly erratic but I never saw it higher than 745/824).

R7800 with stock firmware WiFi: 522/406 Mbps (down/up)
R7800 with DD-WRT WiFi: 146/142 Mbps

All tests were using speedtest.net with a server inside my city's fiber network. The speed tests are usually pretty consistent and I get around 950/950 when I connect my computer directly to the ONT (bypassing the router). ​

Voxel.

 

[RogerSC]

There are firmware settings that will disable NAT acceleration (CTF, or whatever they're using now). I'm thinking that you most likely are using a feature that disables NAT acceleration on the stock firmware, that's what it sounds like. Those used to be things like QoS or per-IP traffic statistics. There may be others as well. I would try leaving your stock firmware settings as close to the factory defaults as you can and see what speed that you see (other than turning on NAT acceleration if there's a setting for that). I don't use the stock firmware, nor do I have the kind of ISP speed where I'm pushing the router at all, so not currently something I've looked into.

I can't account for the speed that you're seeing with dd-wrt, though, no NAT acceleration in dd-wrt...if you did an accurate speed measurement, that must be due to the faster dual CPU's. Don't see what else could be getting you speed like that with dd-wrt.

 

[MartinW]

Thanks all for your suggestions. I'm also thinking that the hardware and/or software changed since intial testing and this would be the cause for the erratic behaivour.

I do know about HW assisted NAT, this is what i use on my ERX currently and it took me from 480->930 Mbit.

So this is my setup rigth now <fibre> -> <mediaconverter> -> <egderouter> -> R7800 and works well.

My only reservation would be if the testing done by this site can be trusted, do manufacturers have the habit of sending "golden samples" ?

 

[RogerSC]

Thanks all for your suggestions. I'm also thinking that the hardware and/or software changed since intial testing and this would be the cause for the erratic behaivour.

I do know about HW assisted NAT, this is what i use on my ERX currently and it took me from 480->930 Mbit.

So this is my setup rigth now <fibre> -> <mediaconverter> -> <egderouter> -> R7800 and works well.

My only reservation would be if the testing done by this site can be trusted, do manufacturers have the habit of sending "golden samples" ?

I've found the testing on this site to be about the best out there. In addition to test results, the firmware version being used for the testing is given, which is helpful. As well as thorough descriptions of the test environment and procedure. However, throughput and signal strength will vary with the environment, including how many and how strong and busy the networks in your neighborhood are, so many factors with wireless that it can be hard to extrapolate to your place. It can also be affected by your furniture, bookcases, filing cabinets, sofas, etc. *smile*. As well as how your walls and floors are constructed.

Wired connections are so much easier to predict than wireless *smile*.

 

[MartinW]

I've found the testing on this site to be about the best out there. In addition to test results, the firmware version being used for the testing is given, which is helpful. As well as thorough descriptions of the test environment and procedure. However, throughput and signal strength will vary with the environment, including how many and how strong and busy the networks in your neighborhood are, so many factors with wireless that it can be hard to extrapolate to your place. It can also be affected by your furniture, bookcases, filing cabinets, sofas, etc. *smile*. As well as how your walls and floors are constructed.

Wired connections are so much easier to predict than wireless *smile*.

Yes , I totaly agree. Although my measurement are done on a wired connection and I have a difference of more then 300 Mbit compared to the result given on this site. So my question is why this is since I can get higher speed with other manufactuers equipment. As an expample my DIR-879, also measured to 941 Mbit here, do give me 960Mbit so in this case everything is correct.

 

[pege63]

So this is my setup rigth now <fibre> -> <mediaconverter> -> <egderouter> -> R7800 and works well.

I also have the same setup as you BUT i cant get my edgerouter to work with the mediaconverter, so i use the ISP fiber router and the R7800 as AP.
Are you Swedish with Telia fiber? Can you PM your settings for the Edgerouter?

 

[sm00thpapa]

Running the latest firmware my R7800 was set back to factory defaults at 8:59 AM this morning on it's own. How on earth could that happen?


Sent from my iPad using Tapatalk

 

[RogerSC]

Running the latest firmware my R7800 was set back to factory defaults at 8:59 AM this morning on it's own. How on earth could that happen?


Sent from my iPad using Tapatalk

Good question...and a good reason to have a saved settings file. I think that's happened to me as well, not with the R7800, though. I've learned to always save my configuration settings once I'm happy with them for a particular version of firmware. Has saved me a lot of time.

 

[pege63]

I newer use old setting, may be thats why it resets who knows.

 

[ulaganath]

I have one big query on this router. I know the clients need to be mu-mimo.

How well this router handles say 3 su-mimo and 1 mu-mimo.

Basically mu-mimo will be fastest simumtaneous transfer.

Say out of 4 devices connected 3 su-mimo and 1 mu-mimo.

If its traditional router like x4 or R7000. irespective of type it will switch between all 4 devices and do the traffic request.

Will this router also does the same or splits the traffic between 3 su-mimo and 1 mu-mimo.

If so it has to switch the response to 3 su-mimo at the same time delivering mu-mimo . Am i right. Which could be faster than traditional su-mimo.

 

[thiggins]

ulaganath-
First, you need at least TWO MU-MIMO devices to see any benefit from MU-MIMO.
MU-MIMO routers handle mixes of SU and MU devices just fine.

The diagrams in this article may help you understand how MU-MIMO works.

 

[ulaganath]

ok thanks for the info. Does upgrading to this router is atleast future proof for upcoming mu-mimo clients and 4x4 usb dongles.