Omnia Turris Wifi throughput very slow 2.4 and 5 Ghz

Hallo Omnia Community,

i have a Omnia Turris 2GB ram ( CZ11NIC20 bought in April 2019) everything on the Hardware and Software side is stock running Turris 3.11.

Now that i have a upgraded my ISP contract with 400 mbit/s download and 20 mbit/s upload, I tested the Wifi and Lan speed throughput.

Here are the test results (

Compex WLE200N2 N WIFI Card 2.4Ghz 1meter distance = 50-80 Mbit/s

Compex WLE900VX AC WIFI Card 5 Ghz 1 meter distance = 120-200 Mbit/s

LAN= constant 400-440 Mbit/s !!!

Are these horrible Wifi throughput test results normal for the Omnia Turris? Is this the WifiGate some people wrote about that Turris is trying to censor?
I already considered upgrading the Wifi Cards with MikroTik Cards and MMCX pigtails with 8 dbi dualband antennas. But I read in this forum that they have with this upgraded setup low throughput asweel. So everything is still stock!

My question is are there any software tweaks that can increase the WIFI throughput?
Is there a Hardware bottleneck with mpcie 1x lanes or the CPU that cause low WIFI throughput?
Can the community post their WIFI throughput Test-results? Would realy appreciate it.

I am really frustrated because this Router was not cheap and it perfomes not good with WIFI.
I considering selling this Open Source router and buying a proprietary Router and then installing OpenWRT onto.

I hope the community can help.

Best regards,

1 Like

I have your own device (in fact mine is even more recent: CZ11NIC23 revision), and similar results with WiFi. I have an optical fiber at home with speeds of 900 Mb/s in download and 100 Mb/s in upload. The Turris Omnia goes to about 850 Mb/s in download with LAN, but on WiFi it is around 350 Mb/s for 5GHz and 25 Mb/s on 2.4GHz practically with the device attached to the router (about 50 cm away ). By now I understood that it is a concrete and effective fact, even if at the beginning I was upset and it seemed a bit crazy. On the forum I read about some passionate fans who come at crazy speeds with WiFi (like super Gigabit), but I look at the facts. And so it is. It’s not bad, but it’s not even perfection. However WiFi in general is a sore point in any network. Perhaps only with dedicated wireless access points connected to the router via LAN, one can hope to have something better. Or maybe WiFi that supports ax specifications, but clients should also support them. I have performed tests with very recent devices and all have full support for the 801.11 ac specifications, so I assume that it is the maximum that it can achieve. With the modem/router of my ISP (a very ugly and anonymous Technicolor) I reach 750 Mb/s on 5GHz, but it has all the disadvantages of a horrible and closed firmware.
The Omnia CPU shouldn’t be a problem, unless you’re using Pakon or Ludus or all those other wonders for which you buy a Turris product. In fact, everything related to data and packet traffic control leads to saturation of the two CPU cores during an intense data movement.

1 Like

Very unfortunate and disappointing to hear that I am not the only one with bad wifi. Your WiFi with 5Ghz and 350 Mbits/s is pretty good compared to mine but for AC standard very low. Your 2.4 Ghz seems to be pretty low too for the N standard.

I seems that Wifi on Omnia Turris has its hardware or software bottlenecks with WiFi. Very unfortunate.

I already considered buying a MOX but I will ditch that idea. Maybe I will build my own low power X86 Router or maybe a ASUS AX Router.

I hope the community can provide more WiFi throughput test data. Can anyone provide thoughput test data with the MikroTik wifi cards?

1 Like

I make test with my router Turris Omnia 2GB - fw: Turris OS - 3.11.10 (OpenWrt omnia 15.05 r47055). Notebook with WiFi - Intel AC 8260 - distance 5m from router.
Test with
2,4GHz - Ch 13 - width 40MHz => 120Mbps/113Mbps
5GHz - Ch 44 - width 80MHz => 310Mbps/195Mbps
Lan cable => 316MBps/195Mbps
for 5GHz upload is probably possible take better speed, but my internet connection is limited.

As far as I can tell the TO (hardware revision CZ11NIC20) provides PCIe 2.0 (being downward compatible) with 4 Ports x1 (links). Thus, max. net througput that can be expected does not exceed 500 Mbyte/s on the 1x link in each direction, less the overhead for the transport protocol, respectively 250 Mbyte/s on the 1x link in each direction for PCIe1.x.

Wlan throughput test results are likely only meaningful for the environment wherein the test is conducted or if for comparision purposes in reproducible lab settings.

Is basically obsolete and perhaps not the best stepping stone to conduct Wlan throughput tests. Better perhaps TOS5.x that is based on the upcoming OpenWrt 19.07 release or even Master. Ther are lots of improvements in TOS5.x | Master compared to TOS3.x

Please note, that this is often not ideal, wifi will not scale linearly with distance, but once you are “too close” achievable rates will often decrease again. Now this is not the main effect you are noting, and even placing the station and router a bit farther apart is not likely going double the wifi throughput, but still worth considering. Also at 100cm you definitely should prefer a wired connection as it will leave more “wifi to share” for stations that can not be easily wired-up.

Fair enough, but note that 500 MBps = 500*8 = 4000 Mbps, wich should be enough for 802.11ac’s maximum 3466.8 Mbps on-the-air gross data rates (not all of this is transmitted via the PCIe bus).

1 Like

What kind of client are you using? Have you done a wifi survey of the surrounding networks, especitially on the 2.4GHz band?
My RTROMO1 1GB pre-order Omnia on TOS 4.0.2 easily gives me 620+ Mbit on download and 500+ Mbit upload on a 1Gbit/500Mbit fiber plan, with a 802.11ac 3x3 client.