There appears to be less detail in the ED mesh on this puzzle, or you have changed the programme slightly. The threshold slider on this puzzle is best set to maximum, ie as far to the left as possible
The density quality will vary from protein to protein, as the resolution of the data gathered from the experimental methods isn't always the same.
I understand from Brow42 that the first posting of this puzzle didn't work because the ED data exceeded 128 Mb. Was there any data loss in re-issuing the puzzle with a smaller data file ?
The original puzzle was posted as an untrimmed case, where the entire density cell was included (including the extraneous bits of density from other copies).
We decided to repost the puzzle as the trimmed case, as the density is smaller. I believe what was lost was only the extra bits of density, and not any of the resolution.
The trimmed case isn't 100% blind, as we decided what to trim based on the actual solved structure - however, in these last 2 cases, the trimming could very feasibly be done without knowledge of the structure, so it's still a very cool result.
You can find secondary structure predictions from Jpred here:
http://fold.it/portal/node/993971