I've run into difficulties which I can't remedy nor explain, and I've very much like someone to help elucidate what is going on.

Background:
I had posted a couple recrystallizations from the same sample material on the "just some xtals" thread. Originally a re-x was performed on a combination of leftovers from old reactions (M. hostilis). The first re-x produced two different types of crystals differing in both appearance and color - a white set, and an amber set. I separated the white crystals from the amber ones, and recrystallized both separately using the same solvent I used previously, heptane.
The glass containing the amber crystals produced a fine end result, but the glass with the white crystals had a troublesome re-x. I changed nothing in the procedure, but upon transferring the solution of dissolved white crystals into the shotglass for re-x there was immediate precipitation, but of a different sort than I have encountered before. There were many particulates, all of which formed a nebulous dense network in a state of constant suspension. After 72 hours not a single crystal had formed anywhere, and eventually as all the solvent which had formed a top layer evaporated off the remaining suspension of particulates just slowly contracted more and more until in the end they a very thin layer of translucent white was left behind. The sparse crystals that did form manifested as small rectangular prisms, representing a new crystal form in this series of experiments.

So the problem here was no crystal formation in the way I expected.

Current Problems (Old Problems Persist):
Since this is batch I reserved for R&D purposes, I decided to recrystallize everything again, separately, hoping to purify further and get better crystals. So I start with the colored crystals and the [formerly] white crystals, like before.

Colored crystals:
There were immediate differences in getting all the crystals into solution. I used hot heptane for the third time, but here I had to continuously add more and more just to try and dissolve more crystals. After adding what should have been way more than sufficient solvent, I decided to just separate the solution from the insoluble material and reduce the volume until it was super-saturated. Yet, precipitate had already begun to develop, even at 4x the heptane I had used previously. Not wanting to over-saturate, I transferred this into a shot glass which just barely ended up containing everything.
But the same thing happened --while side crystals were simultaneously developing, here again was this suspension of particles which had amassed and which sunk to the bottom.
I tried seeding with a decent size crystal (which just pushed up a corresponding volume of suspended material, as shown in the picture below) but this just accomplished causing some minor crashing from the solution above the suspension. I even tried putting this on a Fischer vibration machine, sort of sonicating it. This made the cloud a little less opaque, but certainly didn't resolve anything.

White crystals: Same problem of suspension of particles which don't crystallize, and poor final product.

I might see how the more recrystallizations happen, the more each component compound gets separated out, changing the solubilities.. maybe the cloud was another compound than N,N-DMT? Although to me it seems unlikely, as heptane is pretty selective for that. But I still can't understand why the larger volume of solvent was needed to dissolve the crystals (insufficiently too), and why what I thought should have been a nice end product turned out to be underwhelming, especially since I didn't change anything up.

Any advice?
A thank you in advance.

Picture descriptions:
[1] Typical environment for re-x
[2] The separated white crystals from the original recrystallization
[3] First re-x of separated white crystals [right shotglass], showing the suspension of particles, and compared to first re-x of the amber crystals [left shotglass] (freshly set, normal super-saturation behavior)
[4] End product of the amber re-x
[5] The poor end product of the white re-x
[6] Re(-re)-x on the product shown in image 4, with the colored crystals, showing this cloud-mass.
[7] Where [6] is at currently, showing really not the greatest results

I have a partial answer:

The successive recrystallizations gradually purified by separating out similar freebase components. What looked like white N,N-DMT crystals turned out to be NMT I believe. Pretty insoluble in heptane, ethanol, and water, which is how I separated it this time from the N,N-DMT. Plus, it's melting point is high so it stayed fairly crystalline while everything else had well been put into solution. Hence also the erroneous thinking that more solvent would solve the problem, as well as thinking that the solution was super-saturated when in reality it was just the insoluble NMT (as I currently see it).
Attached is what was finally isolated and which was the source of my problems. The current re-x is looking normal again.

The only thing is, not understanding how so much NMT ended up in the crystals.
Reckoning the possible ways it was introduced, I wonder if there's a heat-dependent subreaction which took place from the heavy crashing when the crystals were first produced. Is there evidence of increased NMT conversion during heat-driven transition stages in the crystallization process?