When I made the sections for the image you’ve linked above, I did it at 40 um thickness, with the spinal cord placed horizontally in OCT as you have in your sketch. I tried cutting dorsal first and ventral first, and didn’t see a meaningful difference either way. In practice, what matters most for getting this sort of image is how flat you can get the cord to lie in OCT, and how much the cryo is behaving on that day.
Some tricks I used for this image: I pre-froze half of my mould with OCT, and then applied a very thin film of OCT that I immediately then placed the cord onto. This thin film would harden very quickly from residual coolness from the frozen layer, but remain flexible for long enough to allow me to straighten out the cord and ensure it lay flat. Once the thin layer had frozen in place, I then added in the rest of the OCT to fully enclose the cord.
These were directly mounted onto slides, not free-floating.