[Prism54-users] Output power regulation - SMC2802W

[kaleda]

On Tr, 2004-01-07 at 03:21, Aurelien wrote:
> Hi,
> a patch was submited some time ago for this. 
It was me, who submitted the patch.
So I;ll try to exlain the issues.

> I took a look at it but did not merge it because there were some unresolved issues.
> How should we use this oid (OID_INL_OUTPUTPOWER) ? 
> 1) The unit is claimed to be 0.25 dBm, 
> and the max value is equivalent to 30dBm. 
> I don't think any card will emit at this power. 
> Legal value seems to be 20dBm. Or the power might 
> be mesured before antenna, at some point in the unit.
There's no doccumentation about max value. It was found only by
experiment, that to output power oid you can set max value 127 (0x7F)
As you know - the oid holds value in 0.25 dBms.
so 127 / 4 = 31.75
Of course its unreal value. We can limit it to max 20 dBms.
There is no problem 'bout that.

> 2) There is probably a value for automatic mode (127 ?).
Also no documentatation.
But one note: when I bring the card up - the default value in output
power oid is _always_ 127, so maybe this is auto mode.

> 3) How does it behaves with negative values ?
???
What do you mean negative values ??? How strangth can be negative ?
And as I already mentioned - you cannot set value bigger than 127, so
there's no negative value.

> 4) Does it really works as expected ? I'm not able to test it...
I also have no specific equipment to precise measure output power.
But simple experiment showed that it changes:
I set output power to max (127) and connect with client.
I set output power to much lesser value (e.g.70), client shows lower
connection quality.
Again set txpower to lower (lets say 10), client shows event lower
quality.
So I think it works.


> And last but not least, I don't think this feature will be usefull. 
> Max value is hardware (or firmware ?) coded, and by default card adjusts 
> power dynamically (allowing max value).
If you're using prism54 card in client mode - its useless, 'cause afaik 
cards sets output power according to AP output power ( correct me, if
I'm wrong).
But in AP mode it's very usefull.

> But if anyone find it usefull, feel free to work on it...
Ok, I'll try to submit another patch, to newer driver version
'cause since last time I patched it, many things changed in the driver.

Kaleda
(aka Antanas Gadeikis)