Behavior of H entry in association list

[tog]

This post isn't about a problem I'm having, I'm just curious about how things work.

I just put a new v3 board up today in an area where there are some neighboring APs/stations, although nothing overwhelmingly strong or directional aimed at me.

For the atheros interface that is in AP mode I see signals of -89, -91, -92 intermittently for the "H" entry. There are 6 associations on this and all have signals better than -70.

I'm wondering if the -92 and -89s it's displaying for the "Host Card" entry are 802.11 frames the driver heard and understood (but otherwise ignored I would expect) from other 802.11b/g devices in the area.

[lonnie]

Yes, those are signals it has heard but has not even tried to associate. It is the nature of the method, in that we take a snap shot of the state and display it. We thought of removing it but that woulkd require a special display function just got Host. Since it hurts nothing we left it.

[tog]

It's cool, I like it fine.

[ninedd]

Hi Lonnie / Tog;
I'm currious about this as well. Just trying to wrap my head around what exactly those numbers mean. On a test WAR we have up with two clients, if we ping one of the clients constandly, the H Host's Signal generally match's the Signal of the ping'd client. If we don't pass traffic, it fluctuates a bit, but generally noise is at -98 and the H signal is -95ish with no traffic going.

On another test AP, when we set it to Channel 1 with zero associations, the signal is at -76, while noise was at -98. If it's signals that it sees that are not associated with it... I guess I'm trying to wrap my brain around what the signal is showing and how this is different than noise?

BTW, I love the new display and this new info! I am finding the Q% a very good deal for the troublesome customers who have a -69 signal but complain of problems. We can now see some of them with great signals, but a Q% that bounce down to 50 sometimes. If can't tell us exactly what the problem is of course, but it's great info to let us know that there is a genuine problem. Thanx guys. :)

[tony]

The signal readings in the Host entry are more or less bogus, and does not mean anything useful. This number will change to reflect the signal from the last packet recieved in the system, weather it be a probe request or data from a client.

[ninedd]

OK, thanx. Actually, knowing that - that's perfect. :) I really like the new display, and seeing the Host's idle time, througput (in KB/s) and even these signal numbers are great. Thanx for the info. :)