My setup: Audience running 6.45.9 on which I have connected, to the 3rd wlan, a laptop with an Intel AX200 wifi card (A) and a small PC running an Intel AC9560 wifi card (B). Wlan3 is setup with a channel width of 160MHz and both devices are connected as such to it - see the screen capture taken with both devices in the same room with router, aprox 1.5m from it, in direct linesight (roughly positioned in a triangle shape).
Both A&B, downloading from my OpenMediaVault server connecting via cable, reach ~80MB/s
Both A&B have latest drivers since I’ve read that at least AX Intel drivers should be kept up-to-date.
A downloading from B or B downloading from A - in either case the speed is ~30MB/s ← is this the best to expect while on wireless?
Agreed! And I installed 6.46.8 this morning, I also fiddled a bit with the settings: switched from country US to country RO, also changed the channel to 5500 Ceeeeeee - no changes. Defaulted the data rates - no change. Won’t get over 30 MB/s..
30MB/s is pretty normal for MT wifi, there is nothing to do with the setting, keep alive frame?? what the… it will only drain your devices battery by keep sending frame every second, Wi-Fi 6 even introduce a technology called “TWT” to save battery life but here we do the other way round?? MT wifi slow is due to driver, not setting, no magic setting here at all. Accept the speed or change to another vendor, there are a very famous says “why do you need such high speed for wireless devices??”, lol
The only way to improve speed is to play with the data rates, basic rates 12/24mbps and supported rates 12~54mbps, other settings seems fancy but do nothing with the speed, forget about it.
Oh well, 30 MB/s it is then.. I tried with default data rates, now I’m back @ 12/24 and 12-54. I gave a try to “keep alive frames” but speed was the same and mobile phones lost a lot of battery.
Hopefully MT does something with their wireless software - I read in the past that they were intending to release an alternative firmware with Qualcomm drivers.. what happened to this idea?
How many Virtual RF AP interfaces do you run?
Due to Beacons per VAP (sent at low modulation) they create overhead which will reduce the overall throughput further.
If you have no interference, and 160Mhz channel on the Intel Laptop working (which I believe it does), then something is wrong
with 240Mbps speed (30MBs).
Might be one of your PC’s can not read or write faster?
have you tried from A to B, A being ethernet wired to Audience with iperf running on both (only valid test method) ?
have you tried from B to A, B being ethernet wired to Audience with iperf running on both (only valid test method)?
if one clients “steals” all the airtime because of bad reception (= low modulation rate, long time to transmit) , then you reduce the
system bandwidth quit a lot. and 30MB on two clients = 60Mbps total tranfert over the air (=480Mbps).
Update:
I thought I give it as well a try as I have a new 160Mhz client at home and it burned my fingers to do some real
iperf3 testing.
Intel 160Mhz 2x2 client (11ac) , showing 1500+ bit rate in the Winbox + 160Mhz channel
MacBook 3x3 with also 1000+ bit rate in Winbox
AP: RB4011 Wifi (same Wifi device as in Audience) 160Mhz Ceeeeeee. (free spectrum).
Running both to an 1Gbps ethernet iperf server, I can not get more than 400Mbps combined throughput
(like 300Mbps/~40MBs on one device and 150Mbps/ ~20MBs on the other).
So seems indeed with 30MBps you are on what audience can give you for now over 2 clients simultaneously connected.
I would have expected more with that 160 MHz channel…
Note: running over ethernet I get the 900+ Mbps speeds so no system bottleneck…
I have the standard wlan1+2+3 and 1 virtual wlan4 linked to wlan1.
As I live in an apartment, there are lots of wireless networks, mostly on 2.4Ghz, rarely on wlan2 frequency range and basically not one on wlan3 frequency range.
A & B both are running Win10 on SSDs that do ~500 MB/s sequential read/write.
I’ve done now another round of tests:
File transfer with both wired:
A copying from B: 90 MB/s (720 Mbps)
B copying from A: 110 MB/s (~900 Mbps)
iperf with both wired:
A server B client: 740 Mbps
B server A client: 940 Mbps
iperf with both wireless (160 Mhz, 2S-SGI both, pretty stable at over 1100 Mbps both A and B, Tx and Rx)
A server B client: 230 Mbps
B server A client: 230 Mbps
A & B were in the same room, some 1-2m away from the router and from each othert, sort of in a triangle positioning.
I must be having something wrong in the settings..
Why do you find it curious: 230mbit between two wireless clients on one radio is actually more or less the same as 470mbit from one of them to outside network.
To have the idea of maximum wireless performance test between two clients when one of them is on wireless and the second is wired.
So, B connected via WiFi is downloading from my eth-connected NAS with 80 MB/s which is in line with its reported Rx/Tx rate of 1300 Mbps.
Shouldn’t this speed happen between the 2 clients?
If they are both on the same wireless?
Of course not: each frame needs at least twice the airtime to be transmitted - from A to AP and then from AP to B.
So maximum you can get is 1/2 of what you have, when only one client is on wireless, and in reality even less, because of increased overhead.
