At small axon diameters (<1 µm), why does myelination not increase neuronal conduction velocity?

Graphs like these are based on a complete cross-sectional area. Myelination adds thickness, so if you want to add myelin outside and keep the same diameter, you have to have a smaller lumen inside the axon.

As you can see in the "unmyelinated" plot versus diameter, smaller fibers have lower conduction velocity. Since the benefits of cross-sectional area are most important at small diameters (because the velocity changes with the squart root of diameter), for small diameter axons you end up losing more from losing internal area than you gain from adding external insulation.

If you keep the internal diameter the same and just add myelin, you will always increase the conduction velocity, but the total diameter including the myelin won't be the same anymore.

That's observed due to passive conduction. With more area of cross section resistance is decreased (ohm's law). Myelin has just function of insulation the graph showing conduction in neuron fibre varying with thickness of neuron (hereby also thickness of myelin sheath). Conduction is ruled by multiple factors not only just thickness of insulation. check this out https://www.ncbi.nlm.nih.gov/books/NBK10921/ When the area is so less than certain level the resistance is really high and insulation provided by myelin sheath works against conduction rather than supporting it. Conduction happens superficially instead in the bulk. Where as in non myelinated fibre the insulation of myelin is not barrier.

Straight line tells me that my reason is legible probably since current is proportional to (1/R and hence) A