Posted: Sat Nov 26, 2022 22:10 Post subject: Slow wifi speed on Netgear 6700v3 after updating to r50927
Hi all,
Today I updated the firmware on my Netgear 6700v3 to the most recent version - r50927 (released on Nov 21, 2022).
However, I was surprised to find out that wifi speeds are very slow.
On 2.4G I get around 70-80MB\sec and on 5G around 250MB\sec.
I performed the test on two different laptops, on which I previously got around 700-800 MB\sec on my Netgear.
I would like to point out, that I tried playing with all the basic settings (such as channels etc), but to no avail. I also switched between "Router" mode and "Gateway" mode, but the results did not change (not that I could notice). I did not want to start changing the TX power because the devices were 0.5 meter from the router during the tests
The problem is, that I do not know from which firmware version the wifi issue started, because for approximately 2-3 months now, I've been using it as my main router rather than an AP (I rarely used wifi during that period, most of the time it was just turned off).
All I can say, is that around 2-3 months ago, when my Netgear was configured as a simple AP\Gateway, I did receive much higher wifi speeds..
What Am i missing? Is there any specific setting I need to change?
I just checked out speeds with my 6700v3 (Client Bridge) and they are running the same as my xr500 (Gateway). They are both running 70mbps on a 100mbps connection (some network congestion going on). Both are running r50927. You might have a configuration issue somewhere. I would recommend going back to the wifi set-up guides. _________________ Netgear XR500 - Gateway
R6700 v3 - Station Bridge
Joined: 18 Mar 2014 Posts: 12877 Location: Netherlands
Posted: Sun Nov 27, 2022 7:59 Post subject:
Probably a typo but no router can get 500 MB/s you probably mean Mb/s.
Measuring speed can be done via lot of ways, real throughput testing is done with two clients running iperf3.
If you want to test wireless performance you connect one client wired to the router and one client wirelessly and you run iperf between them.
If you want to test WAN<>LAN throughput you connect one client wired to the LAN side and one client wired to the WAN side etc.
So without knowing how and what you tested it is difficult to comment.
I have an R6400v2 which is practically the same router, I have done an extensive comparison about 6 months ago between stock, DDWRT and FreshTomato and there is little difference between these three.
(there are small differences but it is less then 5%).
Joined: 18 Mar 2014 Posts: 12877 Location: Netherlands
Posted: Sun Nov 27, 2022 10:21 Post subject:
I did a quick test with my R6400v1 (dual core 800 Mhz Arm A9, so a bit slower than your R6700v3) running build 50944 (alpha build) which shows expected throughput.
Note: the router is not doing anything else and this is a clean spectrum (no neighbours etc)
I erased the settings with "nvram erase", rebooted my router and reconfigured everything from scratch. However I still can not get above 250-270 Mb\sec (and no, I did not mean MB\sec of course ).
I performed the original tests on fast.com and speedtest.net (with no VPN). Yes, I understand that the results can be "noisy" and inaccurate with remote servers, however, wired speed testes on both sites show constantly above 750 Mb\sec.
In any event, I performed a test with iperf between a wired client and a wireless server (both on the LAN side, located 1 meter from the router), as it was suggested above.
With iperf I received the following results (5Ghz, mixed):
Changing various wifi settings hardly affected the above speeds (I would say that the Bandwidth range is 250 - 300 Mbits/sec).
Moreover, I wanted to exclude any issues with my main router, so I turned on the DHCP server on my Netgear and disconnected the main router, but it did not solve the problem. The speed remained around 250 - 300 Mbits/sec.
It is worth noting, that back then, a few months ago (with an older dd-wrt version) I got much higher speeds with no special settings (I did not change anything, save for setting a password and a MAC address filtering).
By the way, is it normal that I have so many errors? Please see the attached image below.
Although I already implemented some of the recommendations on the "Recommended Wireless settings" page, I guess I will have to carefully implement them all.
However, so far, my impression is that the recommended settings supposed to help you going from 90% of the max bandwidth to 95% of the max bandwidth, while it seems that I have some fundamental issue here..
I guess I'm missing something basic here. I will follow the instructions on recommended settings pages, but do you have an idea of any specific setting? Something I should start with?
Just to update you, that yesterday I updated the firmware to ver. R50963 (released on Nov. 28 ), erased again everything with “nvram erase”, rebooted the router and configured it again. But unfortunately, the speed did not change.
As to the channels, when I leave it on “auto” mode, the router selects constantly channels 52 + 54 (5Ghz). Frankly, I'm quite surprised, because according to my tests, channel 52 is one of the most used channels by my neighbors.. So I performed my tests on other channels, but I did not see any noticeable improvement.Currently I use channels 149 + 155, AC / N Mixed (5Ghz).
However, I do have some good news I’m just not sure what exactly helped, but it seems that adjusting the antennas on the router, did slightly improve the speed.
In addition, I tried applying your settings (from the PDF you have uploaded), including the CTF + FA, and this time I performed the tests according to the recommendation I found on the forum, with the following parameters:
I tried also with P8 but the results remained generally the same.
I’ve attached the results below. But it is worth noting that these are the best results I've got so far. After I kept playing with various settings, I again get much lower results :-/ currently I can not get these results again.
In addition, I applied almost all the recommendations in the Basic Wireless Settings & Advanced Wireless Settings pages, but nothing helped (I'm not saying "all" only because I know from experimenting earlier that some of the recommendations do not help in my case).
I would like to point out, that in order to set the “CTF + FA” setting, I had to turn on WAN (I selected a static IP), although I want to use the router as an AP only, so I disabled WAN (nothing is connected to the WAN port). Does it help in AP mode also (when WAN is not used)?
However, after changing the settings (and adjusting the antennas on the router), I noticed, for the first time, that the rate on the info page reached 866.5 Mbit/s. I'm not sure how this figure affects the actual speed I get, but I would say that this is a progress until now, it never exceeded 400 Mbit/s.
I just do not know what else can I do, save for trying to install an older firmware version and running the same tests.
you should test with iperf3
and you can leave out the -b and -w bullshit.
the -b option is only needed for UDP measurements and not for TCP
(default 1 Mbit/sec for UDP, unlimited for TCP)
iperf3 -s
iperf3 -c IP -P 1 -t 60
iperf3 -c IP -P 2 -t 60
iperf3 -c IP -P 4 -t 60
iperf3 -c IP -P 8 -t 60
iperf3 -Rc IP -P 1 -t 60
iperf3 -Rc IP -P 2 -t 60
iperf3 -Rc IP -P 4 -t 60
iperf3 -Rc IP -P 8 -t 60
good wlan drivers don't care how many streams you use, they will achieve the same throughput with one stream
Quote:
As to the channels, when I leave it on “auto” mode, the router selects constantly channels 52 + 54 (5Ghz). Frankly,
sounds like 40Mhz channel width and not 80mhz
Quote:
Today I updated the firmware on my Netgear 6700v3 to the most recent version - r50927 (released on Nov 21, 2022).
However, I was surprised to find out that wifi speeds are very slow.
On 2.4G I get around 70-80MB\sec and on 5G around 250MB\sec.
I performed the test on two different laptops, on which I previously got around 700-800 MB\sec on my Netgear.
I just do not know what else can I do, save for trying to install an older firmware version and running the same tests.
Nothing, according to your last Iperf measurement you reached 468Mbit, it can't get much faster.
For 80Mhz 2x2 ~600Mbit is the absolute maximum.
For 700-800Mbit you would need a 3x3 client and those are very very rare - I know of only a few MacBook Pros that have 3 antennas.
Otherwise, all common clients have a maximum of 2 antennas.
and for full diagnosis you should post a complete screenshot of your wifi settings and the wifi status page.
Joined: 18 Mar 2014 Posts: 12877 Location: Netherlands
Posted: Wed Nov 30, 2022 9:22 Post subject:
It is as @ho1Aetoo said, you should test with iperf3 (like i am doing and when properly setup and having CTF (+FA for WAN) enabled you should get well over 400 Mb/s and with a clean spectrum (I do not have any TX/RX errors as I live in a detached house) and under optimal circumstances (my router is in the same room as my laptop with Intel AX200 Card) you should get close to 500 Mb/s
I also did setup as a WAP (CTF enabled, WAN port not used):
Quote:
A secondary router connected wired LAN<>LAN on the same subnet as the primary router.
On Basic Setup page:
• WAN disabled
• DHCP server Disabled (=off and NOT set as Forwarder!)
• Local IP address in subnet of primary router but outside DHCP scope, make sure the used IP address is unique on your network you cannot have duplicates.
You can run udhcpc to give the WAP a static lease but because you can it doesn't mean you should
• Gateway and Local DNS pointing to primary router
• DNSMasq enabled
• Router kept in the default Gateway mode (the wiki says Router mode but do not do that, either it does not matter (this case) or break things)
• Connect LAN <> LAN (do not use the WAN port unless you really need that extra port, for most routers traffic still must use the CPU so performance is lacklustre )
• I do not change the Firewall settings although you do not want a firewall, the Firewall is automatically disabled as there is no WAN, but it does not hurt to follow the wiki and Disable the Firewall anyway.
You have to add the following rule to the firewall in order to get internet access from clients attached to the VAP.
In the web-interface of the router (the WAP): Administration/Commands save Firewall:
#Always necessary (alternatively set static route on main router and NAT traffic from VAP/Bridge out via WAN):
iptables -t nat -I POSTROUTING -o br0 -j SNAT --to $(nvram get lan_ipaddr)
This gives me even better throughput of about 550 Mb/s, I did a file copy test with Windows (one large 2 GB file) and that reported a sustained throughput of 54 MB/s which is in line with what you would expect.
If you see Rx/Tx error your spectrum is not as clean as mine that could be part of your problem, other problem could be your client and clients WLAN etc. But seeing your iperf measurements you are close to what is possible with 80 MHz 2 stream client.
About using auto channel, that is not always the best choice and is not recommended, I use Wifi Analyzer on my phone (not so much for my 5 GHz but for my 2.4 Ghz)
you should test with iperf3
and you can leave out the -b and -w bullshit.
the -b option is only needed for UDP measurements and not for TCP
(default 1 Mbit/sec for UDP, unlimited for TCP)
iperf3 -s
iperf3 -c IP -P 1 -t 60
iperf3 -c IP -P 2 -t 60
iperf3 -c IP -P 4 -t 60
iperf3 -c IP -P 8 -t 60
iperf3 -Rc IP -P 1 -t 60
iperf3 -Rc IP -P 2 -t 60
iperf3 -Rc IP -P 4 -t 60
iperf3 -Rc IP -P 8 -t 60
I performed the tests according to your instructions. Please see some of the results below (it does not allow me to upload more images). It is worth noting that with -Rc I got slower speeds, so I did not bother uploading the results. If I understand correctly, there is no much room for any improvement?
ho1Aetoo wrote:
For 80Mhz 2x2 ~600Mbit is the absolute maximum.
For 700-800Mbit you would need a 3x3 client and those are very very rare - I know of only a few MacBook Pros that have 3 antennas.
Otherwise, all common clients have a maximum of 2 antennas.
Noted. I wanted to do a speed test with iPad 12 pro, but it did not even find the wifi network (5Ghz, AC/N Mixed, 80Mhz channel width). Only 3-5 year old laptops can find it
ho1Aetoo wrote:
and for full diagnosis you should post a complete screenshot of your wifi settings and the wifi status page.
If you think that there is anything I can change to improve the results, I will definitely upload screenshots of my settings. Thanks!
Thank you!
iperf3 results P2 t60.png
Description:
Filesize:
10 KB
Viewed:
1665 Time(s)
iperf3 results P1 t60.png
Description:
Filesize:
68.73 KB
Viewed:
1665 Time(s)
iperf3 results -P8.png
Description:
Filesize:
62.25 KB
Viewed:
1665 Time(s)
Last edited by noob01 on Wed Nov 30, 2022 23:31; edited 3 times in total
It is as @ho1Aetoo said, you should test with iperf3 (like i am doing and when properly setup and having CTF (+FA for WAN) enabled you should get well over 400 Mb/s and with a clean spectrum (I do not have any TX/RX errors as I live in a detached house) and under optimal circumstances (my router is in the same room as my laptop with Intel AX200 Card) you should get close to 500 Mb/s
Based on your response, I decided to move the router to the living room and repeat the tests there. You can see the results in my previous response. I think that I those are the best results I have seen so far I believe that the wifi spectrum there is a bit cleaner. On fast.com I was able to get 450Mb/s (but only once..)
egc wrote:
I also did setup as a WAP (CTF enabled, WAN port not used):
A secondary router connected wired LAN<>LAN on the same subnet as the primary router.
Thanks, I implemented some of the settings.
By the way, does my ping times (wired->wireless) seem reasonable?
depends on which device you are pinging
if it is e.g. a smartphone with powersafe and in standby then the latencies are plausible.
a PC whose wifi card is not in powersafe and which is connected via 5Ghz should have latencies around 1-5ms.
edit:
by the way the fluctuating throughput in "iperf3 results P1 t60.png" indicates that the spectrum is not clean and the frequency is used by other devices
The ping test was performed between a plugged in laptop in “performance” mode (not a power saving mode) that was connected to the router via LAN to another plugged in laptop in “performance” mode that was wirelessly connected to the router.
I tried playing with diffident channels, but it seems that the results remain generally the same.
I performed another ping test, and received the following results:
**
PING 192.168.1.8 (192.168.1. 56(84) bytes of data.
64 bytes from 192.168.1.8: icmp_seq=1 ttl=64 time=92.0 ms
64 bytes from 192.168.1.8: icmp_seq=2 ttl=64 time=113 ms
64 bytes from 192.168.1.8: icmp_seq=3 ttl=64 time=135 ms
64 bytes from 192.168.1.8: icmp_seq=4 ttl=64 time=157 ms
64 bytes from 192.168.1.8: icmp_seq=5 ttl=64 time=180 ms
64 bytes from 192.168.1.8: icmp_seq=6 ttl=64 time=206 ms
64 bytes from 192.168.1.8: icmp_seq=7 ttl=64 time=23.3 ms
64 bytes from 192.168.1.8: icmp_seq=8 ttl=64 time=222 ms
64 bytes from 192.168.1.8: icmp_seq=9 ttl=64 time=65.1 ms
64 bytes from 192.168.1.8: icmp_seq=10 ttl=64 time=509 ms
64 bytes from 192.168.1.8: icmp_seq=11 ttl=64 time=214 ms
64 bytes from 192.168.1.8: icmp_seq=12 ttl=64 time=138 ms
64 bytes from 192.168.1.8: icmp_seq=13 ttl=64 time=157 ms
64 bytes from 192.168.1.8: icmp_seq=14 ttl=64 time=180 ms
64 bytes from 192.168.1.8: icmp_seq=15 ttl=64 time=1.48 ms
64 bytes from 192.168.1.8: icmp_seq=16 ttl=64 time=20.1 ms
It seems that I get better times pinging a server that is located in another country (from a laptop that is connected to the router via LAN) than pinging a laptop that is wirelessly connected to my router (the distance between the router and the laptop does not exceed 1 meter).
Do you have any ideas?
Do you think that the problem is that the wireless spectrum here is heavily “polluted” by my neighbors and that there is no point trying to find the "right" settings?
56(84) bytes of data.