Thanks for the advice. I agree that the problem is most likely with the grounded conductor. It was firmly attached to the neutral bar and was coated with Noalox, but I didn't look in the meter base yet. If anybody thinks of anything else or a good way to test, please add to the thread. Thanks.