you're not going to meet that cable impedance with a shorter cable
This comment may be misleading. The characteristic impedance of a cable is independent of its length. Short is the same as long. The differences are in attenuation. Cables can be made distortionless (something like LG = RC, I forget). But probably typically are not. They all have attenuation, so the longer the cable, the more signal you lose. This is more significant at higher frequencies, hence a rounding of leading edge transients.
The problem with short cables is one of reflections. If the terminations are imperfect, some of the signal gets reflected back. Try and follow me here, a fast edge is sent out from driver into cable. The cable has a propogation rate of around 2ns per foot. If the risetime if 5ns, then there is enough time for the first part of that edge to reflect back into the cable, reflect back off the driver (assuming imperfect 75 ohm terminations and cabling) and then back to the receiver. The result is some glitches showing up on the transition itself. These squigglies can cause timing errors in the receiver. Now imagine the same thing with a longer cable. The extra time it takes for the glitch to propogate round trip delays the impact such that it occurs after the transition, not during. You only have the first incidence to worry about.
Hence, the answer is to try and get really good 75 ohm terminations, connections, and cabling. Just like dougigs suggested. The second benefit is to add a little length to the cable. Don't worry, the drivers can handle a hundred feet no problem. Then again, that leads to another issue, that of the HF attenuation...
The HAGDAC itself is designed to be a pretty good 75 ohm source.
jh
