SP-D dodecamer with what looks like side to side N termini binding between two of the four hexamers.
Image from Arroyo et al. supplement Figure 4b. (I placed the arrows at the juncture of two sets of N termini from hexamers).  The possible side to side N termini binding is more often seen in multimers but it looks pretty clear in this dodecamer.
Measuring the width of the much narrower N termini junction peak here also shows the LUT plots which identify a tiny peak before the largest peak (outside the N and C termini of the trimer). The tiny peak on either side of the side to side N termini association may be unique and it will be necessary to see whether the shoulder bump is actually part the LUT plots as seen in the most typical images, or if the N termini are side by side if actually in some cases a separate peak (as one can at least conjecture about here) is present on either side (black arrows).  The width of the N termini peak is only 9 or 10 nm where when the N termini are joined side o side as seen here but much wider (like twice) when the N termini are joined end to end (typical) and that number of nm is posted a couple different places on this blog.

No complaining here but sometimes it is hard to see progress in laborious tasks, and maintain forward momentum.
Case in point, trying to find out the number of “height” “brightness” “LUT” “grayscale” peaks in surfactant protein D.
The publication i have quoted the most is by Arroyo et al, and again it is here along with an assessment of four of their images
for peaks along the collagen like domain of the trimers. All the techniques previously employed to create easy to measure, to align, to adjust to a given-consensus-diameter, have been explored before in this blog, but this looks at other parameters. Arroyo et al subjected isolated hSP-D to different pH conditions and as I looked over the images i realized that pH apparently does not affect the presence of LUT peaks along the collagen like arm, thus i isolated, rotated, cut into 1mm slices and centered (horizontally) these images, and then subjected them to FIJI upversion of old ImageJ.
One of the good things about dicing and rotating and using two arms of the dodecamer is that one quickly realizes that there is relatively small effect of bounce, chatter, skipping, vibration in the plots of grayscale. In fact, in all these plots images taken at different angles to the trimeric arms, different magnifications, different resolutions, different HSL, different brightness, and contrast, and at least three different methods of sample preparation, and almost a half dozen different species/genotypes of SP-D really dont change the pattern of the LUT plots that much. That is consoling.
Image below has four images by Arroyo et al, and in each i have plotted a hexamer, and in each i have shown what processes i have used and given the visual peak number and shown the actual peaks created. In the 8 trimeric arms there is a mean of 3 peaks between the N termini and the CRD peaks. I have shown the nm measures for the N and CRD in each of the four samples.