Welcome to the SeeMeCNC Community

I'm seeking recommendations for a 0.9 degree stepper motor(need part number) that would be a good replacement/upgrade for the stock 1.8 degrees Rostock Max 2 steppers.

I've come up with a different method to attach the ball bearing to a screw for the Magnetic arms. I grind a small cup into a phillips head screw, and then silver solder the two together. I do sand a spot in the ball bearing to help flux, and solder adhere. I then have to buf the ball a little because it oxidizes a little from the heat.

why not just use traxxas ball ends?

I bought these before I read about he traxxis ball ends.

Anyone have a 0.9 stepper recommendation?

What problem are you trying to solve with the 0.9 steppers?

I'm just looking to see if I can get smoother curves. I've printed a few parts that should have been round, and were pretty round. But needed a little cleanup, and I'm looking at trying to reduce that cleanup.

Max wrote:I'm just looking to see if I can get smoother curves. I've printed a few parts that should have been round, and were pretty round. But needed a little cleanup, and I'm looking at trying to reduce that cleanup.

Did you create the STL files? Do you know how many segments the circles were made up of?

If you're not aware, depending on the CAD software that was used every circle will be a polygon made up of straight lines and angles. The difference between a 20 point and 50 point circle might be the difference you're looking for.

I have a beast of a system (16 cores/32 threads), so I have no problem with 1000 segments in a circle. Non-multi threaded slicers are not in my vocabulary.

But you're right a lot of the canned STL files are horrible. Sometimes they make me wonder if they even printed out the item to begin with.


The feeling I'm getting is that there is no significant difference going with a smaller step.

As I haven't actually installed my 0.9 steppers yet, I can't make this an official recommendation, but here is what I was able to figure out so far.

The stock Rostock Max steppers are 1.8 degree NEMA 17 wantai 42byghw811p2.
I can't find any info on that exact model, but there is information available on the 42byghw811
http://www.phidgets.com/documentation/P ... asheet.pdf

These steppers are rated at 47 n-cm of holding torque.

Sparkfun stocks the 0.9 degree NEMA 17 wantai 42byghm809
https://www.sparkfun.com/products/10846

These 0.9 steppers are rated at 64 oz-in of holding torque, or 45.19 N-cm, so they're almost exactly the same torque.

From perusing the forum, 0.9 steppers are smoother and make less noise.

Also, iirc the RAMBO has a hard time keeping up on some types of print jobs, because there are twice as many position updates per unit time.

Those 0.9 steppers from Sparkfun should be fine, the thing you need to be careful with is the he voltage (should be in the 3V range).

You should be aware that some slicers remove segments from the original model (both kiss and slicer do this I believe) if they are too dense, and if the Tessellation is too high on the sliced model you will run into issues where the controller simply can't keep up, and it will repeatedly pause during the perimeter.

Semi OT... When will Repetier firmware and slicing programs start doing interpolation and calculating curves like modern CNC machines? Is this waiting on closed loop motion control or just need to be coded still?

You'd have to be specific about what you mean by interpolation. Clearly all the axis move together when a G1 is sent, so I assume you mean something else by interpolation?

They do run motion planners, and the firmware supports G2 and G3, albeit by tessellating internally, using G2 is a good way to see how close to the edge the arduinos are running, even in a Cartesian printer you can't run circles at 1000mm/min without stuttering, and the USB link is completely out of the equation at that point.

Closed loop control only gives you position feedback, it doesn't really allow for anything new from a feature standpoint.

I can see were a finer grain stepper could cause miss steps, but only if the head speed exceeds the processing threshold. That will make the over all print time longer.

When the controller has limited processing capabilities, it makes more sense to put interpolation/curve smoothing in the slicing code. It will make the gcode much larger.

Is sending data over the USB a huge issue if you're not making positioning over long distances? I could also see if you have a lot of very short hops it might slow down because you can't feed the data fast enough.


3d printers for me are experimental, so I want to be experi-mental.


Here's a little more info on the stock steppers. http://www.wantmotor.com/ProductsView.asp?id=155&pid=80 The torque curve is actually pretty flat, which is nice.

Here's the info on that sparkfun 0.9 stepper, it's from the same company. But no torque curve info. http://www.wantmotor.com/ProductsView.asp?id=157&pid=80


Thanks for the info everyone.

Basically there are 2 issues that result in jerky movement and pauses, one is the USB serial connection, the buffers are tine, USB is a clock synchronous protocol and the firmware expects a response for every line sent. The work around is to use the SD Card to print from.

The second is purely a CPU performance issue, every linear segment has to be processed by the firmware by splitting it into smaller linear segments used to move the carriages, so the effector motion appears linear, this is done somewhat asynchronously using "spare" CPU cycles, IME this fails in two cases. Very long linear moves, and lots of very short segments, when it fails motion becomes jittery and often the print head pauses, this destroys print quality, because changes in extrusion pressure are not instantaneous, and the firmware tries to accommodate this by looking ahead.

Going to 0.9 steppers with the existing controller will increase the load on the controller, and at LEAST reduce the maximum speed you can attain.

The Sparkfun stepper came in today, they don't have connectors on the end, so now I have to go and get those... unless I have some extra in the rambo box..

Someone asked why I didn't just get the traxis balls. Well I've been tooling away in the shop, and figured I had most of the parts already.

I know most people don't have lathes at their disposal. But I happen to. I made a small jig for the lathe to screw in a small phillips head screw, turned it down a little, and put a cup in the top. That's mainly so the ball will sit level when the silver solder melts. The picture is the second jig used to silver solder it all together.
Yes buying the traxis balls would be faster, probably cheaper, and certainly a lot less time.

You know what a 3D printer is also good for. Making a funnel to pour new gear box oil into the horizontal fill plug in my lathe. Freaking awesome, wish I had a 3d printer 20 years ago...

There should be extra connectors in the box that the RAMBo was shipped in.

g.

geneb wrote:There should be extra connectors in the box that the RAMBo was shipped in.

g.

And if not, Ultimachine sells them on their website.

So does Hansen Hobbies.

g.

I think the reason that interpolated curves aren't supported yet is because it's hard to program. STL files are vertices; if you want to check to see if a geometric feature is circular, you need to run rediculous amounts of comparison logic statements. This is the geometry equivalent of facial recognition software.

If you have a file that has curves defined (AMF, for example) then you can easily get the curve data. The problem is that STLs are still standard and toolpath software is programmed for the most common system.

The STL circle recognition problem can be solved by checking the output gcode rather than the input STL file. If your points are changing planar angle at a constant rate, you have a constant radius, which means a circle. Hopefully the slicing engines will integrate this logic soon, since concentric parts are a weakness of all types of 3d printing right now. We want parts less than 0.001 out of round!


More on topic, if you get 0.9 steppers make sure you don't have any backlash or slop whatsoever. Magnetic arms may be the way to go.

Generic Default wrote:If your points are changing planar angle at a constant rate, you have a constant radius, which means a circle.

...or one wants to print a hex nut

Yup. If there are more than around 12 symmetrical sides to an object, it's usually a circle. But you've seen how low-poly some of the STLs on thingiverse are! It really wouldn't be that hard to add in basic feature recognition for slicers.