Joe Isabella wrote: > Sounds to me like adding a regulated bias circuit (read: CHEAP) would solve its problems. Checking the limits of linear operation is an important test - likely to be lots less than the 'rated' output even with improved bias supplies. Steve