Now the 30% is not a true number, only some values changed and they changed by different amounts (fb, Fc, Ft, etc...). I am sure they will have a hard time arguing against you!
Well, the girder truss deflection they came up with is too close for comfort (for me), but I just re-calced it again and it looks like the deflection is barely within Code.
I have lot of other very serious doubts, however...I just needed to make a list of reasons why I am rejecting their repair, and I had serious doubts as to the accuracy of their truss software.
The truss software is in question? That is a major issue. I would mention to you that most the proprietary truss packages assume a slight spring connection at the plates which makes their models nearly impossible to replicate by hand or in RISA/Other.