Logic behind Relative tolerances in ode15s ?

Dear All,

I am using ODE15s to solve a system of DAE, there are times when the solution does not converge and ends with the error :
"Unable to meet integration tolerances without reducing the step size below the smallest value allowed (3.469447e-18) at time t. "
But the same system converges If I put decrease the relative tolerances to 1e-0236 randomly.

My question is that while changing the Relat. tol. to 1e-0236 , I am making it more difficult for the system to converge, then why doesn't it run on 1e-03? WHat are the basics of rel. tolerances and abs. tolerances ?

Apologies for the long script ,

Thanks, Tonu