transmission line loss

Hi all,
I have a question about transmission line loss simulation using W-Element in HSPICE. I was doing a simulation with length sweep to evaluate the length for 2.5Gbps PCI Express trace layout. First, I extracted a short segment of different pair trace model(e.g. 500mil) using HSPICE field solver and declared a length parameter to define the trace length. Then I instantiated this trace model with different length parameter assigned. But I didn't see a obvious loss even the trace is about 30 inches long. With 30 inches long trace, at receiver, the eye(the eye diagram for differential value between D+ and D-) height is about 0.39357v , and with 20 inches length, the value is 0.40431v.
My question:
1. Is my method right for length sweep or is there any good approach to reach this purpose ?
2. Why is the transmission line loss so little with 30 inches length versus 20 inches length?
I ever read some document introduces there is about 0.25~0.35db loss per inch for microstrip line

Thanks in advance!
Gio

ayagbasan,

thanks, but why the eye height is still large when transmitting on 30 inches length. I constructed the transmissionline with two segments of microstrip, one segment of stripline and two vias model.

*via subcircuit:
.subckt vias p1 p2
Cp1 p1 0 0.25p
L1 p1 p2 1nH
Cp2 p2 0 0.25p
ends vias

*microstrip line W-Element
.MODEL ms_500mil W MODELTYPE=RLGC, N=2
+ Lo = 3.459889e-007
+ 4.680634e-008 3.459889e-007
+ Co = 9.958592e-011
+ -5.080216e-012 9.958592e-011
+ Ro = 4.919915e+000
+ 0.000000e+000 4.919915e+000
+ Go = 0.000000e+000
+ 0.000000e+000 0.000000e+000
+ Rs = 1.538128e-003
+ 2.310559e-004 1.538677e-003
+ Gd = 6.257168e-012
+ -3.191994e-013 6.257168e-012

*stripline W-Element
.MODEL sl_500mil W MODELTYPE=RLGC, N=2
+ Lo = 3.578891e-007
+ 3.883481e-008 3.578891e-007
+ Co = 1.415685e-010
+ -1.536170e-011 1.415685e-010
+ Ro = 9.789055e+000
+ 0.000000e+000 9.789055e+000
+ Go = 0.000000e+000
+ 0.000000e+000 0.000000e+000
+ Rs = 2.005708e-003
+ 2.355725e-004 2.005927e-003
+ Gd = 1.779002e-011
+ -1.930408e-012 1.779002e-011

Is there anything wrong?
Thanks