Bed Hoist for a 3D Printer - Final Version

A system for Z-axis movement with hoist action. For instance for the D-Bot, Core_XY etc.

This is a replacement of Version 3 of my "Bed_Hoist_for_a_3DPrinter (D-Bot, Core_XY)", https://www.thingiverse.com/thing:2846764. Texts etc. are renewed.
Version 2 on https://www.thingiverse.com/thing:3427719.

In this version the wave from Version 2 has gone. I call it now "final" in stead of "version 3" because it is essentially the same, but without that new wave, which was caused by using the wrong hardware, mostly the bolts I used to function as axes. In the "Notes" I describe the several phases in between.
"Final" does not mean that there won't be any changes, preferably improvements, just that I now think that these changes would not be essentially different from this. But you never know ;-)

I define this "wave" as a recurring irregularity upwards in the print with a periodicity of for instance 2 mm, depending on what worm gear one uses.
Though it seems that by using a very costly worm gear one can prevent the wave from happening at all, see the comments in version 2 by "the_digital_dentist", I still can't find that worm gear. And even then I would hesitate to risk such amounts of money with an uncertain outcome.

So, when I can't prevent the wave in version 2 from happening, I thought to shorten its length, for instance to 0,1mm by using a "differential pulley", here in the form of a pulley with 40 teeth and one with 36 teeth.

On Wikipedia you'll find lots of formulas for such a thing, but here it results in a 20 times reduction of the displacement of the bed for each rotation of the wormwheel. The 2mm wave that comes with a M0.5 worm gear with 40 teeth becomes a 0,1mm wave with this differential pulley.

Other consequences:

• half the force on the parts of the 6mm belt which hold the bed and 1/20 of the force with which to turn the the worm gear;
• 20 times the length of the belt needed, so for a displacement of the bed of 400mm one needs a 8m belt, which would be a nuisance, so I use a closed loop belt, which again has its consequences, like having to deconstruct parts of the printer to get it in place, which is a real bother in this thing.
• also, the belt gets into a twirl doing that, but that is not a problem with the low forces there.
• the stepper has to do 20 times as much steps per mm as before, in my case 32000 steps/mm, using a worm gear of M0.5, 40 teeth. I don't know the maximum speed of those steppers, but I had to lower the MAX_FEEDRATE_Z to 0.8 to get any sense out of it.
• with a HOMING_FEEDRATE_Z of 0.7 one has to be a bit patient when homing.

Though there is some to and fro with the belt, the system as a whole is a really simple one: one stepper, a worm gear, 2 pulleys above and a couple of pulleys down to hold the bed up. Some idlers to direct the belt and something to weigh down the belt at the back. That's it, in essence. No expensive stuff either.

Are ny prints perfect now? No, far from it, but I don't see that regular occurrence any more. The second picture above shows prints of a test object which I use because it shows off nicely the bad stuff (see test_tower.stl). Here you see a comparison of results with the bed hoist and with a lead screw instead of the bed hoist in the same place. Both not nice at all, but the bed hoist certainly is not the bad boy.
There is more work to do, but I feel that I now have a good case for this thing instead of one or more lead screws.

The photographs show the situation as of now. The belt is 2270mm, which is way too long. Around 1500mm would do fine, but I couldn't find one. This one needs some room behind the table but it's good for the whole height.

And again, as before:
I made this thing so I can use it with my D-Bot_Linear_Z-guide, https://www.thingiverse.com/thing:2779260, but it is independent of it, though when you want to replace the lead screw(s) you need good Z-guides.
No good using it in a system with wheels for the Z-axis, because without the support of the lead screws the bed angles so much downwards that this cannot be compensated for with the usual bed supports. Just too much play in those wheel systems, whatever I tried with 3 of 4 wheels. So be warned. ;-)
No such problem with the linear Z-guides. By the way: these guides work so much better than the original wheels that I only would need 1 lead screw, if I would want to go with that. Only, as I found out again when comparing, the bed goes down without power, something the worm gear is good at preventing, among other things.

What remains too in this version is that this thing is still much simpler and quieter than the lead screw configurations I tried. As accurate also, I think.
Edit (March 28, 2020): At the moment this is not true any more, since I found a motor/leadscrew combination of Robotdigg.com, part no. 17HS3001-400N. They claimed that the quality of the lead screw was better than the usual, so, despite the price I tried it and I like the results. I don't have any commercial interests with them, so please inform us if you know of lead screws of good quality and hopefully better prices, preferably without motors, so one can experiment more freely.
I only use 1 lead screw at the back, which is more than sufficient, even with the Z-Axis_Brake on it.

Compared to the loads they are made for, I feel my ca. 330x330mm bed assy of around 4kg won't stretch the belt so as to give wrong printing heights. Somewhere (can't find it any more) I saw a pre-load advised by the manufacturer of some 2kg for a 6mm belt. That equals the 2 parts of the belt over which the load is spread now. And all that with very low accelerations.
Though, of course, it doesn't harm to check when height is of the essence with one's print. But that would be true for the other belts too.

For much more images, materials, assembling specifics etc., see the new "Notes_on_a_Bed_Hoist.pdf" in the "Thing Files", which are also completely overhauled to fit the new situation.
Edit (17 aug 19) Fortified the Motor..._Mount (-> V2a) and changed other parts (->V2); changed the Notes (-> V2) accordingly and actualized the pictures
Edit (20 aug 19) Very small change: see second picture above and Notes -> V2a
Edit (March 28, 2020): I went back to one lead screw, see above.