Cutting and fraying the impeller works, but the NW specific impellers are the best way to go really.
You can also minimize the orifice for the CO2 tubing coming into the NW pump also, this will produce smaller bubbles. This works at lower rates of CO2, but at higher rates, the hole breaks and the bubbles stay larger.
The other thing you can do is add say 2-4ft worth of tubing between the NW and the reactor chamber.
I use a separate NW pump and do not use the Return pump as the NW. I also just recirculate the water in the sump from the NW near the Return pump so I end up with about 4-8 ft worth of travel. I also use a large UV in some cases as a chamber. Some folks use an empty canister filter.
The smaller clean water filter housing work very well, no modification is really needed other than a PCV pipe inside to pull water from the very bottom if you use the NW approach. Want to count the bubble rate? Turn the NW off for a second.
The issue with crippling your return pump is this: when you modify those impellers, sometimes they get unbalanced and lock/seize up.
What is ur preffered needle wheel?