NETGEAR R8000 Nighthawk X6 Review

[thiggins]

Please post here for discussion related to both First Look and Part 2 reviews.

 

[madmook]

NETGEAR told me the R8000 is using the new FCC rules for 5G Band-I. So Tx power has been maximized "as much as possible within FCC limits" (30 dBm total).

Does this mean the lower 5ghz channels are transmitting at the same power as the higher channels, and if so did your testing confirm that on the R8000?

 

[Ciris]

Thanks for the review Tim! Everything you said was pretty much spot on. I feel it's a great router but the price point is what would hold me back from recommending this to a friend or anyone else. Looking forward to the R7500

 

[Razor512]

Seems cool but I wonder why they would disable its ability to dynamically steer clients without adding an option to enable it.

I understand disabling it by default if it will improve compatibility, but it is just wrong to not even offer the option in the advanced wifi settings.

Netgear needs to come out with a new firmware update that adds more of the Xstream functions, and also allows you to adjust the transmit power of the radios.

For example, it should monitor for errors and then move clients to a different radio. Even in a home with all AC devices, this would be useful. for example, with my R7000, if my tablet gets within 2.5 feet or so of the router, the throughput drops until I turn down the routers transmit power. I would like to have more flexibility in how the 2 5GHz radios are used, for example If I do not have many devices, or all 802.11ac devices that can effectively be handled by 1 radio, then I may want to have the second radio set to run at like 100-200mw transmit power and then have it automatically steer my tablet to the low powered radio when I get too close to the router, and then push it back over to the high powered radio.

If I have a mixture of 802.11ac and N devices, then have both radios run at their normal transmit power, and then have all of the N devices on 1 radio and the AC devices on the other.

Overall the Xstream seems like it can be very useful, but it is just implemented poorly for now. (they should at least add a feature where it can detect incompatible devices and avoid steering them).

Does this mean the lower 5ghz channels are transmitting at the same power as the higher channels, and if so did your testing confirm that on the R8000?

From the RCC tests, the measured transmit power is pretty far below the 30dBm for the lower channels. Currently waiting for the FCC report on the upper channels, but I would hope that they are running much closer to the 30dBm for the upper channels.

https://fcc.io/PY3/14200264

Maximum Conducted Output
Power for 5GHz
802.11ac MCS0/Nss1 (VHT20): 25.99 dBm ;
802.11ac MCS0/Nss1 (VHT40): 25.19 dBm ;
802.11ac MCS0/Nss1 (VHT80): 17.85 dBm

 

[azazel1024]

It would be nice to have a lot more advanced configuration options for the band steering. Things like being able choose how you want to assign things. For example instead of breaking up n and ac, doing things like round robin assignment no matter the client type, or assigning all 1:1 clients to one band and 2:2+ to another, or based on link rate and so on.

Or maybe even select which clients can be dynamically reassigned between bands.

 

[Scooterit]

Wandering if the various transmitters are on the same channel or spread out?


Smiles across the wires,



Rogier

 

[thiggins]

Wandering if the various transmitters are on the same channel or spread out?

As the review stated, one 5 GHz radio covers only channels 36-48. The other only channels 149-161.

 

[sm00thpapa]

Cnet says this router is a bust and to save your money.

https://www.youtube.com/watch?v=mX88P0fLdQQ

 

[thiggins]

Cnet says this router is a bust and to save your money.
https://www.youtube.com/watch?v=mX88P0fLdQQ

Dong's reviews are usually better.

He said the 2.4 GHz band was flaky, but says that it passed his 48 hour stress test and 5 GHz didn't.

The example of not being able to use channels 36 and 40 doesn't make sense, since 11ac uses 80MHz bandwidth, which is four channels.

XStream has more potential than NETGEAR is exposing. It will be interesting to see what ASUS and other vendors do with it.

 

[sm00thpapa]

This tri band is a sales gimmick and can't believe so many people falling for it. Some don't even own any 5GHz devices and they are buying this router. LOL!

 

[thiggins]

This tri band is a sales gimmick and can't believe so many people falling for it. Some don't even own any 5GHz devices and they are buying this router. LOL!

The marketing is gimmicky using the high AC number.

But the router can provide real benefit as the experiments in the review showed.

If you have to have every new router as you stated in the ASUS thread, why not buy one and try it before you pass judgement?

 

[sm00thpapa]

The marketing is gimmicky using the high AC number.

But the router can provide real benefit as the experiments in the review showed.

If you have to have every new router as you stated in the ASUS thread, why not buy one and try it before you pass judgement?

I plan on getting it eventually but not right now. Asus seems to be the leader in router development right now. And since Merlin is supporting Asus routers that makes it even sweeter.

I bought the Netgear R6050 to test out and with in 12 hours I sent it back for a full refund.

 

[azazel1024]

My only concern is looking at the concurrent client numbers, those numbers seem very low, especially when you are segregating between ac and n clients on the two bands. Maybe some firmware improvements needed or something?

With my simple N600 router, if I connect two clients to 5GHz, a 1:1 and a 2:2 and hammer it with concurrent downloads I get in the range of 13-18MB/sec to the 2:2 client and 4-6MB/sec to the 1:1 client, which is about 60% of the max bandwidth to each I get when it is only the one client connected hitting the router at max, but it also ends up being around 80-90%% of the total bandwidth I get if I just have the laptop connected and hitting it at max (200Mbps/25MB/sec net yield)

With multiple clients, it still seems like each band is falling apart with the R8000.

With my AC router with the same clients, but in 1:1n and 2:2ac mode this time, I get around 70% of max ac client speed and about 50% of max n client speed for a total around 80% of what the router could do for just the laptop (about 330Mbps combined bandwidth).

I'll grant that is only with 2 clients and not the 3 (4?) tested here, so things might fall apart a lot more with a bunch of clients hitting either router, but at least on the surface it seems like the R8000 has an issue with multiple clients on a radio and/or multiple disimilar speed clients on a radio.

 

[thiggins]

I agree that the bandwidth sharing on each radio, particularly between the AC clients was disappointing.

Bandwidth sharing fairness is very dependent on specific client mix and router radio management.

I suspect that XStream (and end-product vendor implementations) will be a work in progress for awhile.

 

[RMerlin]

I agree that the bandwidth sharing on each radio, particularly between the AC clients was disappointing.

Bandwidth sharing fairness is very dependent on specific client mix and router radio management.

I suspect that XStream (and end-product vendor implementations) will be a work in progress for awhile.

Browsing through the R8000 source code, it seems that Netgear isn't using the stock bsd (Band Steering Daemon) provided by Broadcom but their own modified one. Not sure if this is temporary, or their idea of improving upon Broadcom's own offering.

If any R8000 owner feels like tinkering, see if the filesystem includes an exe called "bsd". That'd be the BCM daemon - Netgear is running something called gbsd instead, and they also modify something at the wireless level.

If you can see the bsd daemon, try this over telnet (assuming Netgear has a telnet interface - I don't know if it's the case):

1) Kill the gbsd daemon
2) Run the following commands:

wl -i eth2 bssload 1
wl -i eth3 bssload 1
/path/to/bsd

To revert back to normal, simply reboot your router. None of the above commands make any permanent change.

 

[thiggins]

Browsing through the R8000 source code, it seems that Netgear isn't using the stock bsd (Band Steering Daemon) provided by Broadcom but their own modified one. Not sure if this is temporary, or their idea of improving upon Broadcom's own offering.

I noted in the review that NETGEAR has exposed a subset of XStream in its implementation. So this should not be a surprise.

 

[thiggins]

PC World Review

PC World review just out. Has more performance data than CNET review. But he didn't test mixed device scenario.
http://www.pcworld.com/article/2456...view-the-best-router-for-a-crowded-house.html

I have no idea how he gets 400+ Mbps from a 2x2 AC adapter.....

 

[sm00thpapa]

Are 867Mbps AC adapters 2x2?

 

[azazel1024]

PC World review just out. Has more performance data than CNET review. But he didn't test mixed device scenario.
http://www.pcworld.com/article/2456...view-the-best-router-for-a-crowded-house.html

I have no idea how he gets 400+ Mbps from a 2x2 AC adapter.....

Dunno. I am getting around 420-440Mbps with my Intel 7260ac to my AC1750 router for same room performance. That is only Tx from the router, Rx it is more in the range of 370-390Mbps. That is only testing with 5GHz set to channel 36. I haven't tried setting it to the upper channels.

 

[thiggins]

Dunno. I am getting around 420-440Mbps with my Intel 7260ac to my AC1750 router for same room performance. That is only Tx from the router, Rx it is more in the range of 370-390Mbps. That is only testing with 5GHz set to channel 36. I haven't tried setting it to the upper channels.

How are you measuring throughput?