Hyperkloosterman sums revisited

We return to some past studies of hyperkloosterman sums ([9,10]) via $p$-adic cohomology with an aim to improve earlier results. In particular, we work here with Dwork's $\theta_\infty$-splitting function and a better choice of basis for cohomology. To a large extent, we are guided to this choice of basis by our recent work on the $p$-integrality of coefficients of $A$-hypergeometric series[3]. In the earlier work, congruence estimates were limited to $p>n+2$. We are here able to remove all characteristic restrictions from earlier results.