Lower Signal Strength After Switching from GL.iNet Stock Firmware

Hardware Questions and Recommendations
1

Using two Flint 2 routers. One setup as main router with Wi-Fi and the second as an AP (LAN and Guest).

I kept records of signal strength in multiple locations in my house when using the stock GL.iNet firmware. I used Wi-Fi Scanner from LizardSystems to measure Strength in dBm.

After a few weeks, I installed the OpenWrt firmware directly from the OpenWrt.org website (OpenWrt 24.10.2 r28739-d9340319c6).

Measuring the Strength in exact locations, the Strength now is about 6 to 10 dBm weaker using the OpenWrt firmware.

For example, at my work desk, the Strength went from -58 dBm to -64 dBm.

I'm using the same Wi-Fi config in both setups: Country = US, Maximum transmit power = driver default, 160 MHz width, channels 52 (router) and 100 (AP)

Is anyone else seeing the same behavior? Any thoughts on how I can improve my signal strength using the OpenWrt firmware?

Thanks.

2

Try wifiman from ubiquiti and make coverage maps, usually it tskes relocating AP one wavelenght off like 10cm

besides channels 36-64 are same just like 100-128

Btw 64-58 is SIX , not 8-10

1 Like
3

Stock firmware uses proprietary Mediatek drivers. OpenWrt uses open source mediatek drivers. Differences should be expected, but stability is very good right now with mt76 open source drivers.

4

6dbm is within beamforming not catching up….

5

Is it possible to install the proprietary Mediatek drivers?

6

Thanks for the correction. I updated the post.

7

Make a coverage map, singular measurement represents about nothing.

8

The easiest way would be to go back to stock. The proprietary stuff is still tied to OpenWrt 21.02 IIRC.

1 Like
9

What do you mean by this statement? Are you recommending I change my channels? Thanks.

10

See here: (channels 36-64 are SAME CHANNEL, wifi adapter/hostap automatically selects any for signalling in presence of other signal in channel)

Make a coverage map, then re-make it with other firmware. You can always use “auto” channel selections.

11

Will do. Using WiFiman from Ubiquiti?

12

Does v24.10.3 have updated drivers?

13

Compared to what ?

14

OpenWrt relies on wireless-regdb to define broadcast power limitations. The wireless-regdb format does not allow specifying different transmit power limitations for client and AP modes of operation. For example, an AP transmit power of 1 W (30 dBm) is allowed in the US on U-NII-1 channels, but client transmit power is limited to 250 mW (24 dBm). So, wireless-regdb enforces a transmit power of 23 dBm in both AP and client mode, because a) it cannot distinguish the different limits and so chooses the most conservative one and b) the wireless-regdb maintainers incorrectly set the US U-NII-1 client transmit power limit at 23 dBm instead of 24 dBm over a decade ago and haven’t corrected the error since. Losing 1 more dBm doesn’t sound like a lot, but it’s a log scale, so…

Most stock OEM firmware allows US AP transmit power of 30 dBm in U-NII-1, which is perfectly legal. So if you are using those 5 GHz channels, you will experience a rather large drop in signal strength after installing OpenWrt, which restricts US U-NII-1 transmit power level to only 23 dBm.

4 Likes
15

It takes kilowatts since a while.

16

Open source mediatek drivers in version 24.10.2.

17

What do you mean? Are you saying that OpenWrt is not limited to a max of 23 dBm?

18

Not really, but you can propose regdb change to normative 250mW as laid out in fcc rule.

19

Also discussed previously in the big thread:

I use channel 149 just in case.

2 Likes
20

Another weird thing I see with the open source drivers is a big difference in Rx and Tx rates.

15 feet nine of sight to the AP:

    Radio type             : 802.11ax
    Authentication         : WPA3-Personal  (H2E)
    Receive rate (Mbps)    : 287
    Transmit rate (Mbps)   : 2402
    Signal                 : 90%
    Rssi                   : -51

This is from Windows command: netsh wlan show interface