I love the idea (only sending Luma over to a BW set). There are some issues though. The PCE Luma signal is made up from a 5bit value and 5bit DAC, which isn't enough to represent the full range of luma for 512 colors.
What should normally look like this as all luma values for the 512 color set:

Actually looks like this:

So some colors (BW "shades") will be completely the same if you just use Luma from the VCE.
It's one of the primary attributes that makes the "VCE" palette different than the RGB equivalent. The other attributes being the entire compressed Blue saturation range, and the roll-off for upper/peak saturation on reds/greens/yellows (it gives you a couple more usable colors that otherwise almost invisible in RGB space).
FYSA: The other thing I can see in your pic, is that the colors that would be near peak saturation (near max red, near max blue, etc) have the chroma signal leaking through on the Luma line. It shows up as XOR checker boarding (which is easy to see in BW than color signals). I have the same turbonanza, and it's there as well (I confirmed it on a scope). I did verify that this is not coming from the base Luma out pin on the VCE. I didn't get around to finding out where the leakage is coming from (I did have a cable from amazon that made it much worse than the source pin at the turbonanza connector).