Bug #1833
closed
UPLINK MCS gets reduced from MCS9->MCS6 when good radio condition exists
Added by arvind.sirsikar over 7 years ago.
Updated about 2 years ago.
Description
During EGPRS UL testing with osmo-trx setup, it is seen that MCS gets reduced very fast even when radio condition is good. Same issue is present for GPRS testing as well.
This condition exists only with osmo-trx setup.
We have modified osmo-pcu.cfg as below
cs link-quality-ranges cs1 6 cs2 5 8 cs3 7 13 cs4 12
TO
cs link-quality-ranges cs1 0 cs2 0 8 cs3 0 13 cs4 0
and validated in integration setup. further investigation needs to be done on how measurement parameter is conveyed between TRX and PCU apart from the usual DATA-IND message
- Status changed from New to Stalled
- Status changed from Stalled to New
- Assignee set to 4368
- Assignee changed from 4368 to msuraev
- Assignee changed from msuraev to lynxis
IIUC this issue was expected until recently when we started providing and fixing (osmo-pcu.git e13cdc503e2faf06699aedd4e84c69ad42916755) measurement values from osmo-trx through osmo-bts-trx into osmo-pcu. fixeria what do you think?
- Status changed from New to Feedback
IIUC this issue was expected until recently when we started providing and fixing
measurement values from osmo-trx through osmo-bts-trx into osmo-pcu.
fixeria what do you think?
ACK. This should not happen anymore, but we need a TTCN-3 test case to be sure.
arvind.sirsikar, could you please test with the current master and update the issue? Thanks!
- Related to Bug #3395: Uplink CS/MCS control is broken osmo-pcu is used with osmo-bts-trx/osmo-trx added
BTW, Fairwaves noticed that OsmoTRX computes significantly lower C/I values for EDGE bursts (e.g. 1 dB for 8-PSK vs 8 dB for GMSK). There was a patch somewhere, I'll send it to Gerrit as soon as I find it.
- Assignee deleted (
lynxis)
- Related to Bug #4373: OsmoTRX calculates relatively lower C/I values for EDGE/8-PSK bursts (compared to GMSK) added
- Assignee set to arvind.sirsikar
We have some TTCN3 tests passing already for that, as explained in #3395.
Improvements regarding C/I in osmo-trx are still WIP in #4373.
arvind.sirsikar please can you test again with current master osmo-trx + osmo-bts-trx + osmo-pcu and see if you still see any issues?
- Status changed from Feedback to Resolved
All related tickets are closed, the problem of low C/I for 8-PSK bursts has been fixed.
No feedback during the past 2 years, marking as resolved.
Also available in: Atom
PDF