Taffer,
I had the very same problem with mine while scanning 6X6 negs. It ran parallel to the long axis of the scanner.

I tried rebooting, power reset and many other things, but it went away only after I reconfigured the lid for reflective scanning. I went back to scanning negs with no problem for quite a while, and then it returned for a while. I don't think I ever cleaned the calibration area, and I didn't take the machine apart.
For a while I thought it might be a bad wire bundle from the lid, but I doubt it now because a bad wire would likely become a permanent problem quickly. And I don't think wire stress would replicate itself on so many units that have reported the problem (virtually everybody with a 3170, I think). I don't really know the source of the problem, so any of the advice you got so far is as good as any I can offer. Let us know if the problem comes back please.
Hi Tom,
definitely that was like the banding problem I was having. What I did was to follow both advices, the one on cleaning the calibration area AND do several full-frame scans with the 3170 tilted around 80 degrees on one side.
One thing worth mentioning is that the cleaning I gave was in some way even ridiculously strict, calibration area glass, translucid sections on the film holder and those two black squares on the bottom of the holder itself. Not sure if that did something, but certainly the scans just after that process showed no banding at all.
But as you say though, those full frame scans were on reflective mode, hmm...
Truth is it's a very annoying problem and really a pita to try to fix with the clone/healing stamp

@Mike: Thanks !
@Doug: Truly, thanks for those diagrams as well !
Oscar