sorry for my late reply !! i didn’t see this thread. I hope Deep K is still around ? The problem is interesting.
In my opinion, your parameters should be alright, if each parameter looks like it is around the default value. The only why the HO is not taking place might be because the PBGT HO is not allowed :
what is the frequency band of the serving and target cells ?
what is the cell type / layer of serving and target ?
other possibility : the target cells are congested (check the QoS Stats : incoming HO congestion). It seems highly improbable if that happen on 50 cells ! 🙂
finally, still about the traffic : maybe the target cell is in “HIGH” load, while the serving cell is in “LOW” load. That might delay the execution of the PBGT HO by 10dB.
Ok, i think that’s about it. All other problems might be due to software issues : are the cells in same BSC, same LAC ?
To pursue the investigation, I propose you to do something during drive-tests : try to make HO from cell A to cell B (drive slowly, in order to give time to the BSC to make the averaging) : at what rxlev difference does it happen ?
Then, drive in reverse gear (:)) from cell B to cell A. When is the HO occuring ?
Ho… check also your PBGT equation. It might take other parameters into account, such as the MS TX POWER MAX, the BTS TX POWER MAX, or some PING PONG FEATURE. As said earlier, check the impact of traffic load, it might add some extra-margins to the HO_MARGIN.
Good luck, and again, sorry for the late reply.