Hi all,
Whenever I get in touch with a project like the MPCNC, I start thinking about modifications, variations, building on the given material. So, to get things verbally rolling, I first purchased a couple of stainless steel tubes, polished, from my favourite metal store.
I found that the tube diameter a) is not perfectly circular and b) varies between 25.00 and 24.92 mm, according to my calipers.
So, I am seeking an answer to the question how the mechanistry of the MPCNC interacts with such imperfections - do the roller bearings’ holders represent some sort of elastic mounts putting some pressure/tension on the bearings’ axes?
What are acceptable tolerances? I can imagine that other hardware is even less straight than the polished VA tubes (2mm wall thinckness, btw).
I also thought about replacing the set of rollers by an axial bearing, like the LM8UU which is heavily usd by reprappers. There are larger variants, even an LM25UU exists, and igus offers a plastic one designed for heavy duty loads. Single pieces of that sell for about 8 Euros.
The LM25UU puts pretty close demands on the diameter tolerance of the carrier tube. Smaller diameters will work, of course, but rattle more or less. How will the 608s behave? Are they mounted under enough mechanical tension to at least minimize lateral mobility?
Most of us are using galvanized steel conduit, it is much less circular than the stainless. Being .08mm off is very good your tubes which probably flex more than that over there span from there own weight. The bearing holders flex to compensate for unevenness.
If you want to replace the bearings with another type you will be making a whole new machine. You are now making modifications off your own assumptions of how you think my design will behave. I strongly urge you to try the machine before you redesign it. There are plenty of untested modifications already out there.
Tolerances are dependent on your design tolerances. Tons of people are cutting plywood I could only imagine that they would be very happy being within a 1/16" over a few feet on a 4’ x 4’ machine. If you are making PCB’s your machine will be much smaller and your tolerance will be expected to be much higher. User dependent and a question I can not answer.
You should print out a piece and see if it meets your expectations, sounds like you might be expecting a little too much from a $500 cnc.
Please let me know how you feel about it if you decide to try it.
Sorry if my questions sounded like criticism - nothing less than that is intended! A good design just triggers my mind to start thinking of extensions, alterations, modifications…
I actually started building your machine, as told somewhere else (@thingiverse). Just now I am printing the second part - I started with the cornerblock locks. The first one matches the tube in an almost gas-tight manner. Printing on my (also heavily modified) OrdBot. Slowly but precisely, .35mm nozzle, 0.2mm layers, E3DV6 - the lock alone needs almost 5 hours, but the result is fine. Very well designed for printability!
And nope, I don’t expect too much from your design, in contrary, I am positively impressed by the reduction to the essentials that make this possible. For the 0.01mm job, I have a small MF70 (Proxxon). The next machines are planned and being assembled already, slowly and stepwise. and this one fills a wide gap between the two ends of the spectrum.
I’ll surely report on my results!
Cheers,
Ulli
(wrote this the second time this morning, Firefox decided to stall immediately before I hit the send button…)
I didn’t take it as criticism, It’s just hard to hard numbers on this machine because of all the different things that people do with it and how they get built. I do get nervous when people want to change things without trying it first, but I do like parts that have a use and have been tested.
Just have to let you know that I am elated, enthusiastic, thrilled by the first parts that I printed. I started with the corner connectors, they match the 25mm polished stainless steel tubes almost gas-tight. Today, I put together the first two rollers, mounted the bearings, threaded the zwo rollers onto a tube, just as a test - and couldn’t believe how absolutely well they fit, how tight their position on the tube is held, yet how easily they move along. That is so much better than I expected - wow! The next roller is printing over night. Not too far from now, I’ll hopefully assemble the rest.
All ordered parts have arrived. Due to my good experience with linuxcnc, I’ll put together a simple controller consisting of one 8825 per stepper, some digital glue for the various endstop switches, and an old PC saved up for applications like this.
Thanks to Allted / Vicious1 for this design! Fun to follow your ideas. And yes, I’ll make my own aluminium motor mounts, maybe stepwise and iterated.