[srsran-users] RACH/TA Oddities

[Dan Dole-Muinos]

Hello all,

I have been experimenting with the limitations of the RA system and have a
question I hope you can give me some insight on.

(For background, I have the EPC, UE and eNB running through a GR/ZMQ
broker, as described in the guide on your site.)

I have removed the parts of the system that follow the RA procedure, namely
the portion of the code that sends MSG3 as a result of RAR reception. This
is in order to prevent the attach procedure from proceeding, and maximize
the number of RACHs sent post-cell search and sync, while trying to keep
within the prebuilt structures of the code. The resulting behavior is that
RACHs are sent until the count hits the "max transmission" value (10 by
default), before the MAC throws an error (though nothing is done as the
relevant timers are still running).

I attempted to increase the UE's "max transmissions" limitation when
sending RACHS, only to find that after the initial 10-12 RACHs are sent,
and accurate TAs received, that the TA values start to incrementally
increase in value erroneously. I have introduced an artificial propagation
delay into the broker, so for the first 10-12 RACHs I receive an accurate
TA of 9, then after those 10-12 the received TA increases to 10, 13, 15,
20, and so on. I never saw an accurate TA again after this point, without
killing and restarting all relevant programs.

Furthermore, when the attach failure is found as a result of a timer
expiring and the RA procedure is restarted from the initial cell search and
sync up through the RACH sending, the TA values received are all erroneous.
As I began to guess that this could be due to the UE falling out of sync
with the eNB, this came as quite a shock.

What could account for this behavior? I have tried a number of different
ways of resetting the MAC and other relevant systems between RACHs, to no
avail. At the moment I'm investigating whether this could be an issue on
the eNB side, though I deem this unlikely as the eNB is unmodified from the
production code. I plan to test sending RACH to a real eNB today to see if
I get similar behavior--whatever my results, I'll be sure to report back.

Thanks for your time in reading this, and I look forward to hearing back
from anyone who can help!

Very best,
Dan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.srsran.com/pipermail/srsran-users/attachments/20220223/02c93635/attachment-0001.htm>