I don't like backlash at all. Sometimes it can be compensated easily with a careful design and construction. Some other times backlash compensation could lead to something either not useful, or even something that may generate more troubles than a good operation.
Don't really spent too much in arguing about Wanhao Duplicator 7 leading screw and nut, but having a mechanism that provides some hundreds microns play when you look for tens of microns layer height, only this alone cannot fit my engineer mind :)
So, you can find attached both the stl file that I printed and works and the OpenSCAD file to generate yours.
The issue is that you cannot find (or at least I can't in a reasonable time) any specifications for the geometry of the thread. All you know is that is 8 mm outside diameter, 8 mm pitch trapezoidal thread with 4 starts. I'm not even sure that the designation used in the filenames is correct. Anyway, it works.
It is designed to reuse the backlash nut, just as a supplementary means of prolonging the life shortage by wearing.
It is longer (40 mm) than the original one since there is enough space under the build plate bracket and there is longer thread for better guidance :)
You may use any good material for this purpose, as igus, POM (Delrin), nylon.
I started with PLA, is true a pretty performant one from Sakata3D based on Ingeo 3D870 material.
Update March 16, 2018
Ok, after checking the first design and some debates on the Facebook groups, I decided to go 100% on the safe side and create an "inversed" version of the nut.
Read in the About backlash section why this could be beneficial.
Old version is still kept for... reference ;)
No longer work "in progress" since 2 of these nuts are working for weeks now with The Black Brothers.
0.1 mm (0.2 mm first layer)
Print slow (e.g. 30 mm/s to allow for cooling and avoid curling of threads margins.
Just clean any elephant foot from the first layer(s).
Play a few times with the screw to "refine" it.
The role of the "main" nut (e.g. this nut) is obviously to transform the rotation of the motor shaft and the leading screw attached to it into translation motion onto Z axis, which will move the build plate up and down.
Ideally, there should be no play between the threads of the screw and the threads of the nut, therefore no matter how small a troatation would be it will determine a proportionally small vertical movement. In practice this is not possible (actually such a screw and nut will not work at all) hence some dimensional characteristics of the threads allow for a gap between them.
Good news is that now the screw and nut will work. Bad news is that this gap will create a play, which means it is possible to have a small rotation which will not determine a correspondingly small rotation. This is called backlash. And this is not good.
There are many ways to compensate for backlash, but still allowing for some "elasticity" to make the things work. The objective is to let the mechanics do their job without any play under any reasonable load conditions (VERY important to keep in mind).
One of the most commonly used is the anti-backlash nut. This is a simpla not with a similar thread as the main nut, which is restricting from rotating around the screw axis by a cage that is part of the main nut body. A hexagonal cross-section is doing the job. Between the main nut and the anti-backlash nut a spring will push both nuts in oposite direction, but since both nuts are held by the same screw they will just keep a distance. The point here is that, if everything is designed correctly the aforementioned objective will be met.
But think about how forces are working in D7 in the original configuration:
- gravity is always downwards, and what is counts is the weight of the build plate and the entire sliding mechanism that is actually supported by the threads of the screw and the main nut.
- peeling forces also act downwards when lifting the build plate and part is still sticking on the FEP sheet
- however the elastic forces developed in the little spring between the nuts will just opose to this!!!
Will it hurt? In most circumstances perhaps not. But this would mean very well aligned Z axis (bars, linear rail, screw, motor). Any misalignment may induce supplementary forces that will opose the movement and sometimes "help" the spring to create some issues.
The worst result that could be is slightly different layer heights, which may alter the aspect of surfaces, or even create weak layers that may break when peeling.
This may probably not happen with a good profile, mainly with adequate speeds and exposure time, but may also be an overlooked cause of print failures.
The idea of the "inversed" nut now become obvious: the spring will push even more and not fight against all :)