Perhaps the scaling of your initial SCA iteration isn’t a problem, but you can read my posts on Can anyone help with Scaling issue(maybe, including the link in one of those posts, with respect to difficulties from using SCA. Note that your initial linearization point might be crucial.
If you encountered infeasibility on the first SCA iteration, follow the advice at Debugging infeasible models - YALMIP, all of which except for section 1 also applies to CVX.