Renormalization of a Tamm-Dancoff truncated light-front Hamiltonian for quantum electrodynamics (QED) is carried out to third order in the couplings. This is an application of the Tamm-Dancoff Hamiltonian renormalization with Bloch-Wilson methods to a gauge theory, an application slightly more realistic than the previous application to ${\ensuremath{\varphi}}^{4}$ field theory. We wanted to flush out any difficulties with the method with an eye toward its possible use with quantum chromodynamics (QCD). We find a noncohering coupling and we find a longitudinal divergence, both in the effective Hamiltonian. In spite of this, we derive Tamm-Dancoff equations. The net result is that the longitudinal divergence found does not imply divergences in an observable, and, therefore, does not ruin the predictive possibilities of the Hamiltonian, in our version of QED, and it removes from the observables any consequences of the noncohering coupling. This leads us to suspect that the Bloch-Wilson renormalization scheme, which requires use of massless constituents, is a valid scheme, despite the persistence of divergences when used on the light front. The observables are both scale independent and finite. It is then the resultant ability to generate renormalized Tamm-Dancoff equations, to any order in the couplings, that distinguishes this method from past Tamm-Dancoff analysis. This renormalization scheme promises successful application to Hamiltonian light-front QCD, with only additional labor being required.