Diallel cross experiments for comparing test lines with a control line under the model of completely randomized designs are studied. A-optimal Type-S designs are derived when the number of test lines is two and three. A more efficient method is provided to find A-optimal Type-S designs when the number of test lines is between four and six.