Discussion:
[ath9k-devel] unknown phase offset added to CSI measurement at 2.4 GHz?
Jeon
2016-06-17 07:08:06 UTC
Permalink
It is not directly related with ath9k, but with Intel 5300 NIC.

In https://github.com/dhalperi/linux-80211n-csitool-supplementary/issues/6,
On 2.4 GHz bands, it appears that each antenna experiences an unknown
phase shift by a multiple of π/2 (or 90 degrees). This is probably what
you're seeing?
If I recall correctly, this effect does not show up at 5 GHz(*). So if
you do your experiments in that band, you may not have this problem.
(*) Manikanta Kotaru from Stanford pointed out that this effect
disappears on 5 GHz, and pointed at Section 3 in Jon Gjengset, Graeme
McPhillips, and Kyle Jamieson's ArrayPhaser (@jonhoo's) paper as evidence.

I wonder whether it applies to all 2.4 GHz devices, which means it is not a
manufacturer-specific problem. Or, is it a manufacturer(Intel)-specific
problem, so thus ath9k does not suffer such problem? Or even worse, will
ath9k suffer such problem at both 2.4 GHz and 5 GHz?

Regards,
Jeon.

Loading...