ok… but exactly, what are you looking at ? can you share some values ?
with a CRH = 14, you should see a huge issue in GPRS mobility (a lot of TBF drops)
you should also see less SDCCH LAU, and a decrease in CSSR, because MS are stuck in their serving cells, without being able to move to (reselect) a better neighbour cell.
With a ACCMIN = 110, the effect should be even more pronounced.
The Ericsson implementation of the CRH is fine, I can see it in the many DT I’ve done, no worry about that aspect.
Let’s do some testings ? select two cells, one from each side of a LAC border.
Put CRH = 0dB on both cells
Put CRO = 10dB on cell A, observe that trafic increases in Cell A
Put CRH = 20dB on cell B, observe that trafic increases (a little) on cell B, and decreases (a little) on cell A. Also, you should not there is less SDCCH LAU in cell A.
Now there is another possibility that the border between cell A and B is very “poor”, meaning that RXLEV dips below -110dBm, in which case the MS will not be using the CRH anymore.
CRH works well if RXLEV at borders are “CRH” dB higher than ACCMIN.
Let me know how it goes…