Yes, @Griffin.Labs, that is accurate, and the hydrogen bonding crystallization of CBD and THCa is not missing anything, @8livedevil8.
Hydrogen bonding crystallization in some cannabinoids does not contradict Δ6a,10a-THC (the formal number name for Δ10a-THC) or Δ10-THC’s ability to crystallize, @MagisterChemist. I think the Δ10s just happen to crystallize by a different mechanism. I’m sure you have seen the very fine particulate aggregates that characterize Δ10 or 10a crystallization, as in @Kingofthekush420’s post below yours (post#11). Also, as @8livedevil8 posits, CBN can and does crystallize.
I have thought about this for a while now, and I think the extended conjugation of the Δ10 and the Δ10a double bonds with the substituted benzene ring in the THC molecule may hold the key for them, while the extension of conjugation through both benzene (phenyl) moieties in CBN allow it to crystallize the same way biphenyl does!
I believe that Δ10-THC & Δ6a,10a-THC and CBN might actually crystallize by similar mechanisms. All of them have either a double bond or a benzene ring conjugated with the central benzene ring, allowing for delocalization of the pi orbitals to spread across the edge of the oxane (formerly: tetrahydropyrole) ring moiety, and into that Δ10 or 10a double bond or the benzene ring.
This gives CBN a larger field of flat aromatic magnetism across the majority of the molecule. Basically, like sticks to like by van der Waals forces (remember my post on those?), as long as the repulsive forces do not overwhelm the attractive forces.
The conjugation likewise gives the Δ10 type rings an auspicious chiral bend to the Δ ring, allowing the other Δ10 or 10a molecules to be shaped in a way that allows them to stack by intermolecular attractive forces without various parts hindering or pushing others too far away to allow the large similarly electromagnetic (dipole) fields to rest closely together in a nested fashion.
To understand this, one must know that these are very bumpy, jagged molecules for the most part, when observed 3-dimensionally. Those OH groups’ oxygen atoms might line up with electropositive regions around the very electronegative regions created by the conjugated pi orbitals, exposing the electropositive areas on themselves to further molecules’ electronegative conjugated regions… but only if they can do so with out strongly repulsive parts getting too close together.
They might even nucleate as a hydrogen bonded dimer to start, then lattice out from that point, since OH dimerism happens pretty easily!
Like so: Ar-OH•[Ar-OH•OH-Ar]•OH-Ar
…where the part in brackets is the dimer nucleate.
In summary, organic molecules without 2 OH groups can and do crystallize all the time, and it is usually dictated by the most common conformational isomer being the most prevalent, and just right for intermolecular attractive forces to dominate.
Look at menthol on Wikipedia for a simple example with lots of isomers, but only 1 that is natural and common. It crystallizes like crazy! There are more obvious instances of this, too, like how coconut oil forms crystal looking structures when crystallized (frozen at room temperature) after melting. Slight polarity and induced dipoles go a long way to crystallization when the organic molecules are the right shape or floppy enough to be able to get out of their own way. Polymers and fats can crystallize, too, usually in laminar or spherulite aggregations/agglomerations. I mean, look at that CBG from @future4200! Looks like butter to me! 
I hope this helps explain the mystery a bit!