From: Jordi Murgo (jordi_at_lleida.com)
Date: 2002-04-08 11:09:19 UTC
Vladimir Ivaschenko wrote:
>Based on the formulas from wvlan_cs and post from Jodi Morgo I have made a >patch against 2002-04-05 to retrieve signal, noise & quality statistics. >
>Please find it attached. > >The reason I decided to use formulas from wvlan_cs as they seem to be more >correct to me. Quality value for prism2 never went above 92 in my tests >and wvlan_cs formulas seem to compensate for that. I also changed maximum >for quality (max_qual.qual) to 0x5C (92) as per wvlan_cs driver. > >In any case original formulas as posted by Jodi Morgo are available (I >commented them out). >
I don't remember where, I have read about Intersil relation from COMMSQUALITY RID and dBms:
Level dBm = COMMSQUALITY_RID_SIGNAL - 100
You can see in the attached table, this relation isn't ok.
My first proposed relation is based on Intersil one, with scale adjustment, and isn't far from orinoco's statistics:
Level dBm = ( COMMSQUALITY_RID_SIGNAL * 100 / 255 ) - 100
My second proposed relation (see attachment) is based on Intersil formula, with adjustment based on QUALITY indicator from COMSQUALITY_RID:
Level dBm = COMMSQUALITY_RID_SIGNAL - COMMSQUALITY_RID_QUALITY - 100
The levels in dBm are similar to the first proporsal, and not far from orinoco/hermes levels.
Maybe Jouni have some info about, under the NDA they have signed with Intersil ... :)
Salut, <jordi />
| Distance from AP: | 70cm | 1floor | 2floor | 3floor | NOAP | |
| Hermes iwconfig dBm: | -38 | -53 | -78 | -87 | -102 | |
| COMMSQUALITY RID P2_SIGNAL: | 150 | 104 | 28 | 19 | 27 | |
| COMMSQUALITY RID P2_SILENCE: | 0 | 0 | 0 | 0 | 27 | |
| COMMSQUALITY RID P2_QUAL: | 92 | 55 | 0 | 0 | 0 | |
| P2-SIGNAL-P2_QUAL-100 dBm: | -42 | -51 | -72 | -81 | -73 | |
| P2_SIGNAL*100/255-100 dBm: | -41 | -59 | -89 | -93 | -89 | |
| P2_SIGNAL-100 dBm: | 50 | 4 | -72 | -81 | -73 | |
| MAX(MIN(P2_SIGNAL,0x8a),0x2f)-0x95 dBm: | -11 | -45 | -102 | -102 | -102 |