I'm not sure why you're seeing such poor performance from the single point position measurement (and I don't have the equipment to test it onhand), but you might consider using a buffered position measurement to get faster throughput. You can use the "Measure Buffered Position (NI-TIO).vi" as an example. Then used a fixed GATE, to make sure the sample rate is constant. You can use the paired counter (in your example you're using ctr4, so you'd want to use ctr5) to generate a continuous pulse train (Generate Pulse Train (NI-TIO).vi) at your desired sample rate. Make sure to use "other counter output" as the GATE of your buffered position measurement counter.
Hope this helps.
gus....