As I'm ruling LightJet 430 by myself,

From my point I can tell that EXACT reproduction is plain impossible because of differences in pigments, for example. From other point customer can get VERY exiting result and device itself can use different materials at their full range - mean densities, color puriness, resolution and other.
And, of coarse, materials itself are different - for example, in field of spectral sensitivity of pigments and matching that to exposuring lasers wavelength - which leads to different achievable color puriness, etc.
In field of precision such kind of devices are unbeatable - linearisation charts let us control and compensate development at least ten times more precise then Kodak charts, as to say. Working grey balance on any material deviating not worse than 0.012 D on any channel on any field of chart - in real life. Device exposuring resolution overexeeds resolutions of most materials themselves.

So, if you can't get straight exellent result - there are problem not in devise or digits, but most probably in scanner operator.

Good luck!

P.S. Fuji Velvia can achieve density like 3.95 D, as I can recall. On other hand, Kodak Duratrans can achieve like 3.3 - 3.4 D under standard development/exposuring on such devises. And, as I can recall, there are no materials with more than 3.4 MaxD for expose. So, you can't EXACTLY reproduce your transparencies - if they are using full range of Fuji Velvia. But you definitively can get VERY good approximation and exiting result.