I don't know if anyone is still reading this one, probably not; however, I just wanted to say that the solution above works great.
I had a similar problem, when under no load I was able to check the 0-10v Analogue voltage out and it mapped perfectly as expected. When the spindle (in my case a servo) was connected and turned on it would only match for the first few hundred RPM (10v = 2500 rpm) and then would essentially pause at 5-600 rpm with lots of spindle speed fluctuation. I was ripping my hair out trying to solve it and found this post.
I installed a ferrite tube on + and - wires approx 1/3 along the wire length from the pokeys board; the tube was 8mm ID, approx 25mm long and I did 14 turns through the tube (as many as I could fit in, I probably could have pushed for another two but didn't want to break anything). I additionally installed a 22nF capacitor (0.022uF) across the wires at the board.
Now it works like a treat, rock steady RPM (the servo drive provides a real time RPM output from the encoder) with no noticeable fluctuations. Accuracy to target RPM within 2 RPM, I might be able to calibrate that out, but I think it is fine....
Thanks for posting your solution up here and if anyone has similar symptoms to either of us two, this method works! (total cost about $8AUD for parts, and I have 4 ferrite tubes left over).
Cheers