Datacolor User to User Support Group.

After a year of using MIS inks with my R1800 I just returned to the
Epson OEM and started making profiles for it.

The first profile is Harman Fb Al.  As previous postings here
recommended the 729 patch target for the R1800 given its unique inkset
I did that plus the extended grays.  It was also recommended that I
print the targets with gloss optimizer (glop) off even if I intend to
do my actual prints with glop on (and I do).

A print with glop off using the default profile with my Spyder3Print
was a little on the yellow side, especially in the highlights.  While
annoying, a small bump toward blue and magenta in the S3 software
fixed this.  

Here's where things get confusing- for reference I printed out a 21
step grayscale patch using the profile that I was happy with.  I then
printed a second 21 step grayscale ramp with gloss optimizer on.  It
had a distinctly yellowish appearance.  Real world images (a portrait
with a dead neutral white backdrop) exhibited the same issues- this
isn't a measurement artifact- the prints themselves look quite different.

All images were printed with the same (proper) driver settings and
saturation intent as recommended.

Here's an excel comparison of the two ramps:
http://www.jingai.com/phototests/Spyder3%20glop%20no%20glop.xls

Based on this, should I be printing the S3 targets using gloss on if I
intend to use glop with my actual prints?

For the record, my glop cart doesn't appear to be contaminated, etc as
I used QTR to print out a calibration strip of all my ink channels and
it is clear.  100% glop measures about the same as paper white except
it's (very) slightly greener.