After dedicated mode MLB activation why Low BW layer TA samples: near reduced and far increased?

Hi Experts,
After dedicated mode MLB activation why Low BW layer near TA samples have reduced while far samples have increased?
It impacted on increased of QPSK usage and degraded multiple KPIs including CQI.
Traffic/users are shifted to higher BW layer, Thanks

This is because MLB is done with measurements.
So far away users in low BW layer do not measure higher BW layer this si why are not subject to MLB. Only close users are subject of MLB.

Right, but this was not the case in Idle MLB?
Can we improve this scenario?

I don’t think so.
As long as you achieved your target (equal PRB utilization on both layers) it should not be a problem.

No target is not achieved aswell as most of users on Low BW layer shifted to Higher BW which was already high utilized.
Also I kept MLB OFF on lower layer.

Then you need to review your settings for MLB.
Thresholds for source and target cell, timers etc. usually MLB works well.
There are plenty of params to control it.

Using these settings for initial tuning and keeping InterFreqMlbSwitch OFF for L900.

You should make load evalute threshold from 30 sec to 5 seconds.

Do you mean for both layers?

Just for higher layer
What’s your prbloadcalcmethod? Should be prb_usage.
Also mlbtriggermode shoudl be prb_only.

Yes prbloadcalcmethod is prb_usage and mlbtriggermode is prb_only.

What’s the bandwidth for high layer and for low layer?

Higer is 1800 (15 MHz) and Lower is 900 (5 MHz).

There’s no way you can decongest 1800 MHz (75 PRBs) by shifting traffic to 900 MHz (25 PRBs).
Because 25 PRBs will get easily saturated and MLB will not work anymore.
What you can do change InterfreqMLBthd from 65% to 50% for 1800 Layer (75 PRBs).
This way you will force MLB to work almost all the time.

Thanks, I got your point @RFSpecialist, and well explained.
But 50% InterfreqMLBthd will balance out the layers, but as far as from measurement point off view near by users will waive off thats why CQI degraded and 64QAM reduced.

This is on 1800 Layer?

No, that is L900 layers which as per your explanation nearby users/64QAM shifted to L1800 and far off users/QPSK users remained in L900 thats why individual layer L900 KPIs seems degraded.

When you load balancing by forced handovers (e.g. MLB), such statisticall changes are bound to occur.

One method to marginally avoid this issue is, use A5 based measurement of target frequency.

And set A5 threshold such that, very good radio condition user will not shift to target carrier.

You should then deactivate MLB on L900 layer so users from L900 will never be shifted to L1800 layer.

Yes, I already kept InterFreqMlbSwitch OFF for L900.
Shall I keep PrbMlbSynchronizedUE OFF too for L900?