Main clock voltage levels for Cyclone III

We haven't solved the problem yet, but we found that due to manufacturing defect our Cyclone III was not soldered to the board. We confirmed this by using a microscope to inspect the edges of the device where we could see that the BGA balls were not touching the pads under them. It is possible that the bottom end of the clock signal will be closer to GND once it is actually connected to the FPGA, but we won't know until we manage to get another board made.