AW: OsmocomBB Layer1 progress

Andreas.Eversberg Andreas.Eversberg at versatel.de
Mon Jul 5 15:06:59 UTC 2010


> JFYI: The AGC control in L1 is now active.  At least it didn't make
things
> worse for me.  Maybe people who have had problems acquiring weak cells
> should give it a try again.

> I suggest to add the L1CTL_DM_REL_REQ into the state machine, I
suppose
> it should avoid that problem.


hi,

i will check this issue about releasing and establishment of dedicated
mode. also i will do the power control thing from L3.

i got stuck while transcoding RACH request / confirm from L1CTL into
RSL_CHAN_RQD/RSL_CHAN_CONF. the goal is to have pure RSL between layer 3
and layer 2. the biggest problem are things like power measurement, sync
requests and other L1CTL messages, all which are not defined for RSL
SAP. i don't like the idea about converting everything from L1CTL to RSL
and parsing it and vice versa. this take allot of time to implement,
also we would need to add several new messages and information elements.
my suggestion is to leave L1CTL as it is.

currently i think about how to control tx_power and TA (timing
advance/delay) from layer 3. it is not possible to set them in every
layer 3 message, because this message may be splitted or repeated or
layer 2 may send responses without interaction of layer 3. therefore we
must store the tx_power and TA at layer 2, and it must be sent with
every frame then. since we don't send any idle frame to layer 1 (this is
actually generated there), we can only give tx_power and TA at frames
with content. the values must be stored at L1 for idle frames. therefore
it think it would be much easier to use something like L1CTL_POWER_REQ
to set the current tx_power and a L1CTL_POWER_CONF to get the actual
(closest) power set. (also includes the TA). these values are stored and
for every frame sent afterwards until L1CTL_POWER_REQ is sent with
different values.

regards,

andreas





More information about the baseband-devel mailing list