I didn't see the error with your first integral and I had to stop evaluation of the last one as it seems to run "forever" and was eating up all system ressources.
The first integral gave a solution but basically the integral wasn't calculated, also "simplify" didn't help.
Mathcad interestingly enough can solve the corresponding indefinte integral quickly and without pain. Do you really need a symbolic solution for symbolic t1 and t2? Then you could use I(t2)-I(t1) where I() is the solution of the indefinte integral.
As Mathcad can solve without limits it should be able to do so for the definite integral as well. Oftne it helps to add additional assumptions (that a variable is not complex but pure real, or that a variable is always greater than 0, etc.) In case of your problem Mathcad is willing to solve the definite integral if you make assumtions about the order of l and r. I could solve for l>r, l=r and l<r respectively.
See attached file. In case you use a different MC version which throws the errror you reported, I attach a pdf of the file as well.