After precipitation & centrifugation use a pipet to remove the ethanol, then add a second short centrifugation step to concentrate the remaining ethanol from the vial walls at the bottom of the vial and use again a pipet but be careful not to transfer the pellet.
I don't know if you might be doing this already but after the initial removal of ethanol, it might help to run another centrifugation step to ensure anything further drops to the bottom of your tube. I find that is most often the culprit, a drop somewhere not near the pellet that I couldn't see. Depending on what use we are talking about, maybe it's also reasonable to allow the pellets to dry in a fume cupboard?
the above answers are good. Do not be tempted to heat the samples to evaporate off the last of the alcohol. If the dna dries too much it becomes very difficult to dissolve properly again
spin, then carefully remove the liquid by 10uL pipette tip, then allow the tube to dry in a heat block 37 degree. Monitor the pallet every few minutes, and add buffer to dissolve the DNA when the pallet trends to be transparent.