> The sidebands removed 1000 Hz from the carrier might be somewhat in > the clear but the problem is you would also be measuring the error of > the 1000 Hz source. Average them (upper and lower). - JimT