Is the clock frequency of ADC5400EVL 1GHZ?

Hi,

You are correct that there will be some loss through the transformer for a sampling clock of 1000Mhz, but since we usually are using a signal generator for to source the clock we have the option to crank up the launch amplitude of the clock from the signal generator more than we might otherwise.  We also use a narrow-band band-pass filter on the clock and analog input signals to filter out as much phase noise from the signal generator as possible and the filters often have from 3dB to 6dB of loss themselves at the center of the passband, so some more loss through the input circuit is not that much of an issue for the EVM.   But since the signal generators and bandpass filters *are* single ended, we do need to do the single ended to differential conversion on the EVM and a transformer or balun is cleanest for that.

Your application may or may not need the single ended to differential conversion depending on your clock source.  If you have a fixed sampling frequency such as the max 1000MHz that the ADC is rated at, you could use the same balun as is used on the analog input of this EVM - it has a much higher bandwidth than the transformer that was used on the clock input.  Keep in mind that a balun does not provide AC coupling like a transformer does so additional AC coupling would be needed after the balun.  We often design the EVM for flexibility, and make the six pad landings for the magnetics to accept either a balun or transformer from a varitey of vendors depending on the need or application, and that is why this EVM has AC coupling caps after the clock transformer - in case we were to remove the transformer nd install a balun instead.  We would be covered for AC coupling either way.  And if you look at the layout for this board, the pad landings for this clock transformer allow for two different sizes of component to be soldered down - the smaller balun or the somewhat larger transformer.

Regards,

Richard P.