@ozzy It all depends on the slew rate of your clock. For typical transport (25-30ns) you need at least 1.5m or less than 0.5m, but the latter should include all connections (inside of the clock and DAC), so safer would be <0.25m for the cable. I would go with safer choice: >1.5m. 1m might be OK if your clock transitions are faster (likely), but you have to experiment. Whole point is that reflection on impedance boundary at the end (if any) should miss (be late) midpoint of transition that originated it. It is almost impossible to measure characteristic impedance and very difficult to measure slew rate, because any capacitance (scope probe) affects it. You have to experiment.