Replaced CDP w/ Transport; Now, Soundstage Changes Dramatically w/each CD. Why?

Here's something I've never encountered. I've borrowed  a Heed DT Transport and am loving the increased resolution, compared to my Jolida JD100 cdp BUT, now the sound-stage expands and contracts, depending upon whichever CD is being played.The effect is very evident-- not subtle. This is really weird. 

System: Heed Transport -> Aqua La Voce II DAC -> Wells Majestic Integrated -> Silverline SR17.5 Monitors.
I don't even know where to begin. . . any suggestions? ? ? 

BTW, replacing the Blue Jeans BNC cable with an Empirical Audio BNC cable didn't change anything. 
StuartK, the suggestions to replace your BNC cable are on to something.

It's a crucial link. And unfortunately the state of the art SPDIF cables are expensive.

When I upgraded to the Black Cat Tron, I couldn't believe the difference it made. It was like I was only getting 60 percent of the DACs performance with the previous cable.
macdude: you may be right, but when I replaced the Heed transport with a Simaudio 260DT, the extreme variations in sound stage and overly (to my ears) analytical/forward sound I'd been experiencing vanished. All I can conclude is that the Heed is in a class, along with some DACs, engineered for maxiumum resolution. For me, the added detail it provided, although enjoyable in and of itself, compromised the system's overall musicality to such a degree that I no longer enjoyed listening. The previous owner had mated it with tube power and I can see how that would make sense. I plan to keep the Moon. 
The two problems I see are (1) scattered laser light getting into the photodetector and (2) the CDs tend to be out of round and thin and flimsy, so when spinning at high speed, they flutter, vibrate and wobble. If the transport was vertical the problem with vibration/flutter wouldn’t be so bad since gravity exacerbates the problem when the disc is horizontal. Unfortunately the Reed Solomon error detection/correction codes do not (rpt not) correct all errors AND the laser tracking servo mechanism cannot keep up with all the motion of the disc. The laser and the data are both nanoscale so it doesn’t take much relative motion to throw things off. Therefore, being able to control scattered light and CD vibration should be a high priority, since there is no way to recover losses of sound quality from either one. I.e., by the time the data gets to the DAC it’s too late!