SO3 style Z axis

With the reinforced X and Y axis it is time to work on the Z axis. As Wider Tyre tests have shown that the load bearing capbilities of the Makerslide wheels can be improved why not make a SO3 style Z axis.

The big question is what is better?

Solution A with a belt drive. Backlash tolerances defined by the belt only
Solution B with a belt driven ACME rod and nut, Backlash tolerances defined by the combination of belt, rod and nut.

How would you test the differences?

1 Like

That is option C.

Someting like this?

1 Like

My vote is for a direct drive ACME Screw as well.
I am running a direct drive ACME screw on my Z axis using a c-beam setup. (it has a anti-backlash nut on the ACME screw)

I am assuming the belt drive is added on the stock XC setup to add a bit of a gearing ratio to help the Z axis handled the weight of the spindle motor. This was probably needed for the nema 17’s. But my stock nema 23’s have handled it fine.

If you use the same diameter input shaft (pulley with teeth) and same diameter output shaft (pulley with teeth) where do you get this gearing ratio? I teach these things in class and I would like to know. links are relevant.

What I have been taught is same as this

Edit: same input diameter and teeth as output then there is no gear ratio (1X1) you would need to have one gear larger or smaller for a ratio.

Also I am not trying to be “that guy” just want to know if this is something I have wrong.

I could not remember if the 2 pulleys were the same diameter (as I replaced mine with a direct drive a while ago) as was too lazy to go digging through the setup instructions to verify. :slight_smile:
I just looked it up and they are both 20 tooth gears - so yes I was wrong, there is no gearing ratio on the stock design.

(Though I have seen other designs and mods that used pulleys with different teeth numbers to add a bit of gear ratio so you could if you wanted to.)

Thanks for clarifying that. had me worried.

A bit off topic, but here is a good article on micro stepping which I found via openbuilds 3D printer article on the problems of using 2 (or more) motors to move the Z axis.
It helped me wrap my head around the whole micro stepping process. As well as illustrates why you might want to use a mechanical gearing ratio to increase resolution rather than relying on micro stepping for an increase in quality.

For the Z-axis feed rate is not too important so gearing that drive down to increase torque is a good trade off. I think 4:1 pulleys are easy to find for GT2/GT3 belts IIRC.

The torque delivered is a function of mirostepping
Full step 100%
Half step 70% (used on the SO2 Z axis)
1/4 step 40%
1/8 step 19% (used on the SO2 X, Y and SO3 X, Y, Z)

The motor choice should be optional, this depends on the total weight to be moved and required acceleration, less on feedrate.

The openbuilds anti backlash nut is 13mm higher than the Inventables nut, has identical mounting holes and can be used.
This design makes use of the original X gantry plates of the existing machine and low Z profiles, this reduces the overhang to 21mm. The Openbuilds unit, well engineered with the double bearings requires 41mm.

Too bad nobody voted for the belt driven version, really easy to construct but with the reduction in torque due to microstepping not likely to work with a NEMA17

And with a belt driven ACME you can choose a ration but most of you voted for the direct drive.

What is the maximum speed and acceleration for the ACME nut? I make Model Airplane parts in 2.5D and the multiple bridges per part take ages at a conservative Z settings.

I am not sure on the part specifications but on multiple builds I have seen 500mm/sec used for plunge speed max. Though this may be a factor of the plunge cutting operation and not a limitation of the screw for rapid travel (not cutting) moves.

Looking into it a bit more, the R7 CNC design uses ACME screws on all axis.
They recommend “Max travel speed 6000mm/m (suggested 1500mm/m)” so it looks like the max travel speed of a ACME scree is comparable to that of a belt drive.

With the Openbuilds ACME nut I went to 1000mm/min and 100 mm/sec^2 acceleration, positioning is excellent with $1=255. (moved up and down the Z axis 10 times 20mm and came back spot on a 0.01mm gauge. At lower $1 settings it drifts away.

The ShapeOko Wiki gives a plunge rate guideline of less than 1000 for most of the materials.

This seems to come from the max rpm of the stepper motor (3000) and the 2mm pitch of the ACME rod. In the mean time I stumbled upon a second hand 8mm ballscrew, only a short stroke but perfectly for cutting balsawood at high z speeds. It is capable of handling 140 kg load, this should allow me to increase the Z speed considerably.

The design is ready:

Four WiderTyre Wheels and bearing blocks are ready.

rest of the parts ordered

1 Like

Slowly getting there, to make the project more complex I will ad an RC motor style Spindle to it. Reading the pwm output and sending a servo signal to the motor speed controller should allow me to switch the motor on and off with the software.

Design completed, as I do mostly 2.5D (max 12mm) I will raise the Spoilboard and connect both Y axis together with some 40x20 Misumi profiles. RC Spindle will be very compact. Now it is time to order enough aluminum to finish the project.

The SO3 style Z axis with ballscrew is completed. I had to abandon the RC type Spindle for the moment, the motor is not running well. So for the moment the Dremel needs to hang on.

Rigidity is slightly better than the improved stock Z axis, I think I reached the wheel limits here. Stepcraft publishes 0.07 to 0.12mm flex at 2 kg, my machine equals the 0.07mm with the spindle in low position.
The 8x2mm pitch ballscrew sounds great, at Z speed 1000mm/min and X-Y 2000 mm/min 30mm up, a square of 300mm in X-Y and back down 30mm gave no difference on a 0.01mm gauge, very happy with that.

The prototype on the table is the new fuselage nose section, I need to improve on my design skills but the machine is capable of making the parts now.

For machine control I moved to Estlcam, first impression after this prototype is that it is a lot less complicated to run than my previous toolchain.