I have been testing the anit-backlash nut which is on ebay.
This was something I always wanted to design but never got around to... well someone has finally done it and exactly as I would have.
I tried to measure the back lash of the STD nut on the thread. I don't have a dial indicator just digital verniers but it looked like 0.2 to 0.3mm back lash... certainly feels like a lot. With this setup there is well none.
Here it is installed on a Wilson II I build for a friend.
What I did was cut 2 coils off the spring as it was way to long and the preload was a bit much. You could go a little more but I found this did work.
They fit together better one way, so once you workout the way mark them.
I found the threads a little tight. I got a battery drill, put some tape around the end of the leadscrew to protect it and put it in the chuck. On slow, I ran the nuts with spring assembled with some light machine oil up and down quite a few times till it felt nice and smooth.
After assembly I ran it up and down in Z and found it still a little tight. I stopped and checked my alignment and made sure all looked good. I bumped the Vref voltage to about 0.9v (I run DRV8825 drivers) I added more oil and run it up and down quite a few time... like 20 times.
Once happy I started a few random prints and it seemed good, so I stopped it.
I ran the Cube Torcher Test but in S3D I set Retraction Verticle Lift to 1mm so every retact it lifted 1mm and down 1mm and then scaled the test to 150mm.
It ran well I kept an eye on the motors and drivers but they never got warm despite the amount of Z lift every retract.
I would call that a win.