The N termini of a multimer of SP-D still seems to me to be a “ring” rather than a center bundle. The ring in the case of this image (not mine) measures about 20 nm in diameter and each of the arms about 40 nm.  These measurements are relative to concensus measurements on SP-D (typically the dodecamers) and I am not at all convinced that most individuals use correct micron bar markers.  In fact, in images where fuzzyballs (SP-D multimers) appear along side of dodecamers, I wager that they are slightly more compact,  per this image from Hartshorn et al (Respiratory Research, 2008) keeping in mind that this particular SP-D is a mutant which they label as Met11, Ala160, their bar markers and image below. Diameter of a dodecamer is actually 130 nm according to their measure (which is in orange, at 100 nm).  There is no center gap in this particular multimer, however there are also many many arms and the center ring might actually be shown here dimensionally covering what would be a center dark area as seen in multimers with fewer arms.
Here is another AFM (not my image) which shows the center ring to be about 20nm in diameter, and each of the arms of this multimer (three measured here) measure at about 40nm.  The main peak lateral (circumferential) to the Nterminus peak in dodecamers which lies in the collagen domain, is about 11nm in width, (as seen on many previous posts for dodecamers, on this blog about SP-D) and is also measures about that in this image… in terms of distance to the peak of the Nterminus.  I take this to mean that the center of the multimer is NOT at the center of the whole molecule… because it is just NOT a peak, but a valley, but instead at the center of each of the bright dots (in this case 7 bright spots in a ring of  this SP-D multimer of 14 arms).