mroczis/netmonster-core

Misidentification of B66

Closed this issue · 20 comments

So now after the update netmonster misidentify B66 as B3

Screenshot_20221213-102216.jpg

Screenshot_20221213-102133.jpg

@mroczis please check this

This is common issue with Samsung modems. From the screenshot I'm assuming you are using Pixel 7 (Pro) device.
Considering the correct EARFCN is 66811 and NetMonster Core returns 1275 we know that there's buffer overflow somewhere in Samsung's native code (the difference is 2^16 = 65536).

The question is how devs of the other app know that it's overflow and 65536 should be added. Please send a bug report from NetMonster app in order to examine raw data from the system.

In order to do it, go to app's settings, fast tap 5 times on Version and then follow instructions.

This issue is affecting my OnePlus 8 5G as well as my Pixel 7. I will you the send bug report from my OnePlus phone since I don't have the pixel right now.

@mroczis here is the report file:
report.txt

The question is how devs of the other app know that it's overflow and 65536 should be added.

The solution other apps are using is based on the fact that B3 is never used in North America. If the MCC indicates a country in North America and the EARFCN indicates B3 (or any other band that isn't used in North America such as B1, B8, B20 and B28) just assume the EARFCN has overflowed and add 65536.

The question is how devs of the other app know that it's overflow and 65536 should be added.

The solution other apps are using is based on the fact that B3 is never used in North America. If the MCC indicates a country in North America and the EARFCN indicates B3 (or any other band that isn't used in North America such as B1, B8, B20 and B28) just assume the EARFCN has overflowed and add 65536.

Is this going to be implemented to netmonster / netmonster-core?

@mroczis

Did some digging and here are the results. Overlapping occurs from band 66+. When it comes to current deployment of LTE networks, there might be following conflicts:

Band EARFCN Reported ARFCN Overlaps with Where
B66 66_436..67_335 900..1_799 B2, B3 Canada, Chile, Taiwan (B3), USA (B2)
B67 67_336..67_535 1_800..1_999 B3, B4 Switzerland (B3)
B71 68_586..68_935 3_050..3_399 B7 Canada, USA
B75 69_466..70_315 3_930..4_779 B9, B10, B11 Switzerland

Some rules can be hardcoded to NetMonster Core. However there are some limitations:

  • B66 + USA - thanks to overlap with B2 only range 1_200..1_799 can be used,
  • B66 + Taiwan - thanks to overlap with B3 only range 900..1_199 can be used,
  • B67 + Switzerland - thanks to overlap with B3 only range 1_950..1_999 can be used.

Considering that Bell runs B66 on 65_536 + 1275 = 66_811 we are safe in Canada and even for this EARFCN in the US.

Before I add this to Core, don't you know on which EARFCN do conflicting US, Swiss or Taiwanese networks run?

EDIT:

Canada uses pretty much the same bands as USA.

Meaning that we also have band 71 and b2

Beware, Canada Have band 7 however. We also have b41. @mroczis

I also got a screenshot if b66 misidentification on Verizon (roaming)
Screenshot_20230101-163707

in my first screenshot, the band 2 screenshot was correct. The only incorrect value was band 3.

Also Canada have band 7 so I you might want to modify the code

A little poke here, @mroczis

Screenshot_20230302-111121
Screenshot_20230302-111101

Band 7 is now being identified as B71!!! Canada HAVE band 7

Hello @mroczis

Could you take this in accordance?

Hi even I'm having the same issue for freedom mobile Canada the app misidentifying b7 as b71 even though freedom has b71 but it hasn't been rolled out yet and my phone doesn't even support lte b71 and other apps show b7 and even the API from phone states b7
Screenshot_20230314-135005
Screenshot_20230314-134944
Here is the cell tower and as u can see same

Yes, very blind move from the ghost
@mroczis

Hello @mroczis . I also wanted to bring this to your attention:

Screenshot_20230323-092634.jpg

Thanks, neighbouring cells were also fixed