[Banana BPI-R4] Wifi7 status

It seems you have wrong configuration. It was mentioned in this thread few times basic configuration - e.g. here

thnx sooo much for pointing me in right direction!!!
finally mld QCA9274 5/6GHz works not just on openwrt but also on debian/ubuntu

With the new commits in mt76 snapshot, which seem to include that patch, they do not correct the power as the ImmortalWrt patch did.

And what's worse, with the new commits, the ImmortalWrt patch has stopped working for me...

Sadly only 6Ghz band seems to be affected/improved with yesterdays mt76 tx drivers.

As long as the situation continues I’ll continue to utilize this patch:
https://github.com/openwrt/mt76/commit/8088940ffd047a6d282a95af829562e8038f5b2d.patch

Negative side effect is the increased heat dissipation of the ic’s involved (+10 to +20 C).

Some time before the mt76 update, I had with 6GHz a high upload speed, quite the same as the download speed, about 1.3Gbit/s (measured with iperf3). Now I have a way lower upload speed, about 800Mbit/s or so.

I am writing again to confirm that the ImmortalWrt patch does work after the latest commits in mt76. Without it, even despite the latest commits, my empty EEPROM BE14 still did not work.

But, again, its performance is far inferior to an MT6000. Next week I will receive RG178 pigtails and a plastic casing that I will have to drill holes in so that I can attach them to the antennas.

1 Like

I found out one of the culprits: my laptop antennas. When the lid is open, the antennas are inside the lid, the upload speed is way faster, sometimes even faster than download speed. But if I close the lid the upload speed goes way down.

We had today a big commit with updates for the mt76 driver, `ed66c2dfa304e0c1054023d2cb9f5d892a7a75c6`, but it seems to keep the fundamental problem about the TX power being showed incorrectly, and the real TX power is also weak.

1 Like

Build a new image based on this new driver and had significant lower heat generation (I assume based on lower power consumtion) of the BE14 card for the 2,4 GHz and 5 GHz module (6 GHz was already in regular temp window).
Temperature of all three radios is now between 45-51° C / 113-124° F.
WiFi Range is kept same (not good :wink: ) ~10/12m for 5/6 GHz and 15/20m for 2,4 Ghz with external antennas.

Now my two BE14 cards (1x 1st batch with filled eeprom and 1x 2nd batch with missing eeprom parameter) performing equal based on temperature/power consumtion with OpenWRT Snapshot builds. MLO also kept working as intended.

This BE14 eeprom/tx_power patch was kept included in the mentioned latest build:

https://github.com/openwrt/mt76/commit/8088940ffd047a6d282a95af829562e8038f5b2d.patch

1 Like

Hi!

Are you saying that with that patch, your faulty BE14 achieves performance parity with your BE14 with the correct EEPROM?

I will test this patch. The other one I was using didn’t work because the way the code was rewritten.

Please remember that appearing the correct TX power is different than achieving the correct TX power. I sit literally at the side of the router and without any patches I get a distance as you mentioned, about 10/12m, even I’m in the same room just a single meter apart.

1 Like

absolutely, yes

1 Like

I can confirm that. But I still had to use the forementioned patch: https://github.com/openwrt/mt76/commit/8088940ffd047a6d282a95af829562e8038f5b2d.patch

2 Likes

Thank you!

I had read that this patch pushed the chip to its maximum power and caused it to overheat.

Until now, I had been using the immortalwrt patch because I came across a message on the Banana Pi forums that suggested it:

But I don't have a BE14 with a correct EEPROM to compare.

1 Like

IIRC, I had a problem with this patch since it sets the values but never commit it them so they can actually being used. The code for this “commit” part was changed, so I think I will try this patch again.

1 Like

I tested this patch, and it seems to be working. The TX power levels are correctly showed in luci, and I can see the correct signal strength in my device (I use the Wi-Fi Analyzer app in my Android to measure that).

2 Likes

Hi. I'm a bit new to openwrt, but i have be14 with zeroed eeprom.
How do you apply this patch? Compiling openwrt is a bit to much for me.

I installed immortalwrt(last version till 08.02.2026) and still have really bad 2.5g and 5g. 6g is great as far as i can tell.

I installed rg178 pigtails. And there is a slight improvement: 5g is now a bit faster than it was, but still like 300mbit/s slower than 5g on my tp-link

off main topic…

As you probably noticed MT7927 is gettting more and more available for buying.

Offcially there is no linux drivers but some approach was already made to working it under linux.

image

Hi guys, i switched back from a MTK custom firmware to the 25.12.0 release, it kind of works, except for the 6Ghz channel: i can connect to it with no issue, but no data is coming through (tried multiple frequencies) ( icmp, udp, tcp: nothing). Is it something that some of you saw also ?

I suggest to unplug power for 2 minutes, plug back and check again.

If that does not work, please share wireless config.

1 Like