User:Robinc

From Rt2x00Wiki

This page is for my testing and development notes regarding rt2x00

Random notes, testing and thoughts are here, i may not follow instructions when testing as i want to see what works and what dosn't as the average user dosn't follow instructions either. Also these tests can verify regressions and new bugs

Contents

My test setup

Toshiba Satelliite 2060CDS laptop 366Mhz 160M ram Debian unstable


Some notes on RSSI

After speaking to ivo I am doing some work on rssi and sensitivity calulations by looking at legacy code and doing some experiments.

With rt2500 Maximum collected RSSI seems to be 121 ( i have not quite observed 121 but got over 115 with touching antennas). I would assume if the antennas are physicaly connected 121 is what you would get. This seems to equate to 0dbm as the rssi to db offset is 121

dbm=rssi-121

Now averaging.

Ralink use a expodentaly decaying mean, this is very good at coping with environmental conditions changing unlike a normal mean and only uses a single storage location unlike a rolling mean. :-

Adv=Adv-(adv/8)+new_reading

the /8 can be efficienly done with a bit shift >>3

Becuause of the above code to get an actual value you must take adv and >>3 and the data passed to Asicbbptuning() is infact the real (not *8) value of adverage rssi.

So the first thing ralink do is convert rssi to dbm by taking off 121.


What i have noticed with rt2500 is that there is no tuning when not assicated, i think this is a bad idea because i started loosing beacons when the signal became very strong or very weak.


RT2500:-

BBP R17 is the Voltage Controlled Gain selector (VCG)

On The rt2500 when the chip initalises R17 is set to least default sensitivity which is 50, maximum sensitivity is set by code and is 64, but it appears that for very strong signals the lower level of gain has been reducde to 44 in rt2500

So this gives values for RT2500 BBP R17 (VCG) of Least sensitive 44 (decimal) Most sensitive 64 (decimal)

It is possible that more or less gain exists outside these ranges.

RT2400 :-

BBP Register for gain is R13

Seems max level is 0x20 for R13

Eeek can't find a minimum level presume its 0x00

Code does not look in as much detail as rt2500 does and levels



23-May-2007

Managed to get rt2500usb to kernel fault when unplugging the device, didn't record the backtrace however :-( and have not achieved same results.

Firmware location on debian unstable is /lib/firmware/

RT61pci hard faulted on an eject of card, partial backtrace points at rt61pci_txdone+0x26/0x58 [rt61pci], yep thats the EIP, stack context was :-

rt61pci_txdone+0x26/0x58 rt2x00pci_beacondone +0x2c/0x58 rt61pci_interrupt +0x9d/0xa6


22-May-2007

Testing rt2400 pci against office 802.11B AP and some nearby neighbours, network is 128bit wep encrypted

Seems to not enter associate if not kicked in some way when the interface is up

Switch to RT2500USB

Similar problems with associating as rt2400. Can manualy associate by setting key and essid after interface up but logs so associated but cannot set ipaddress and sucessfully ping router

Retrieved from "http://rt2x00.serialmonkey.com/wiki/index.php/User:Robinc"

This page has been accessed 1,258 times. This page was last modified 14:49, 3 June 2007. Content is available under GNU Free Documentation License 1.2.


SourceForge.net Logo
Find
Browse
Main Page
Community portal
Current events
Recent changes
Random page
Help
Donations
Edit
Edit this page
Editing help
This page
Discuss this page
Post a comment
Printable version
Context
Page history
What links here
Related changes
User contributions
My pages
Log in / create account
Special pages
New pages
File list
Statistics
Bug reports
More...