BLER performance in different Channel Model i.e. EPA, EVA and ETU are giving weird results.

I am presently working on the effect of Different velocity on the LTE's BLER. I used the reference of Matlab communication toolbox's example on LTE i.e. http://www.mathworks.com/help/comm/examples/lte-phy-downlink-with-spatial-multiplexing.html. I am getting weird results for EVA 70Hz, ETU 300Hz. I tried using Matlab's definition for Excess Tap delay's and Relative power and also tried defining them by myself. But the result is very unbelievable for e.g. the value of BLER for CQI = 1 (QPSK and CRate = 0.101499) is "1" . For CQI 1, PDA 0Hz SNR -1_0____1____2_____3__3.25___3.5__3.75____4 BLER 1___1_0.95__0.65__0.275_0.175_0.125_0.075_0.05 PDA 5Hz 0_2____1____3_____4__5___5.5__6 BLER 0.95_0.725_0.55_0.325_0.15_0.1_0 EVA 70Hz 1____2__ 3____4 0.225 0.2 0.11 0.08 for eva 70Hz and other channel the result does not come as expected; the increase in doppler shift must increase the BLER so, why is it so wrong?

If anyone have run through similar problem and found out thw solution please let us know. Thanks