I have just begun printing the parts, and I must say this is a thing of beauty. I appreciate the attention to detail in making the design printable. The top interior of the 5/8 holes was imperfect, but firm hand pressure using the screw as “broach” and it cleaned right up.
[attachment file=“mpcnc_roller.png”]
The rollers ride on the rough-looking conduit much better than I would have expected. I suppose the ugliness does not reflect its bumpiness, but in any case I am pleased.
I am curious why the instructions call for 2x each of roller and mirrored roller. I can see the parts are not technically symmetrical but it appears functionally they should be equivalent. Is there something I am missing?
After a mild printer SNAFU, I’ve finally got a complete set of the printed parts. I’m thinking 24" by 48" table top dimension, so conduit lengths slightly less than that.
I’m not sure yet what I’ll be doing with this machine, and right now I’m more fascinated by the MPCNC machine than I am interested in what I’ll be making with it.
One thing I’m noticing in the parts list is a large number of #6-32 machine screws and nylock nuts. (Thank you, by the way, for making them all the same size and length!) One thing I’ve been using in all my own designs lately are 1/2 inch #6 sheet metal screws (pan-head), which I can get locally in boxes of 100 for cheap. (To my untrained eye the threads look the same as wood screws, except wood screws come with predominantly flat heads, which tend to split the plastic.) To fasten two parts I make one of the holes 3.6mm, not intended for engaging threads, and the other hole 2.9mm. It holds very well, in that the joint is usually stronger than the part, and it’s super convenient. It is the only fastener I’ve needed.
I hope this doesn’t sound like a critique, or a request to change – I’ll be the first to acknowledge that I don’t know how demanding these joints are. But I would suggest to anyone who uses captive nuts, try sheet metal screws and you may be pleasantly surprised.
I think that is a common theme around here, and I love that!
I used to do that a lot. I think that is a great way to fasten things. For this machine though, all the repeated motion and the variance of everyone’s printers it just never worked very well. For personal projects though, I even use machine screws fastened right into the plastic.
Something I learned at my previous job there is actually a screw for this exact purpose (well injection molded plastic at least), McMaster-Carr
The are slightly triangular instead of round, they work so good. It is worth getting a box for sure.
Now that I’m looking at a detailed picture, I am pretty sure I’ve seen these before when disassembling plastic things. They look a little ‘funny’ compared to regular symmetrical screws, but I never realized it was its own thing. Very cool!
I’m sure nothing folks haven’t seen a hundred times before, but here’s my build so far.
[attachment file=89433]
Because I don’t know what I’m doing with it yet, I’m going for a relatively tall build. I might want to try an extruder, or I may want to try some liquid coolant/lubricant in which case I will want to put a tray for catching the splatter underneath.
My thought is that if I want higher stiffness and reduced Z range, I can add stiffeners to the legs and insert a raised work surface, or I could put holes in the table below the feet and lower the legs through the table without having to disassemble much.
I think I went too cheap on the 608 bearings and they do not run smooth. In my first test with a single linear rail, I used some bearings I already had, and apparently they were nicer than the bulk pack I got from Amazon. I am most likely going to tear everything down and replace the bearings but there are other things I can move forward with in the mean time.
The deck is basically a table with no legs. It is simply particle board supported by some 1x4 boards on edge. I’ve used this design (with legs) for work benches in the past and it’s surprisingly stiff for very low cost.
If I understand correctly, theory predicts that the stiffness is inversely proportional to the cube of the the span length. I thought it might be interesting to try to measure this, so I arranged a scale and a dial indicator, and took measurements shown below.
The two dimensions measured from the interior of the plastic are 19 and 31.375.
[attachment file=89584]
[attachment file=89585]
The stiffness I measured at 0.685 thou per pound for the short span, and 1.76 thou per pound for the long span
[attachment file=89586]
[attachment file=89587]
The shorter span is stiffer, no surprise, but the quantity does not appear to obey the relationship. I believe this is because of my setup, or it could also be because the corners are providing some stiffness and their contribution does not follow the same proportionality. I suppose I could loosen the corners and then the relative stiffness might be closer to the ratio predicted by the span lengths.
These are relatively easy measurements, but more interesting would be the lateral stiffness of the legs and the torsional stiffness of the center assembly, which could perhaps predict an estimated lateral stiffness as a function of Z height. But that’s later. I’m starting with the easy measurements.
My setup has numerous imperfections that I know about, and probably more that I don’t know about. A metrologist would be horrified. But this seems reasonable for a first order approximation.
[attachment file=89588]
Having trouble embedding or linking to youtube videos… they are “RVDDH4oV4b0” and “b_2RJtPqFOk”. (Or maybe the post is held in moderation?)
I will have to try the flat surface to see if it improves. I think I will also put a hole in the deck so I can pull straight down. Also it occurred to me that the gantry, depending on its position, could actually help support the side rails when loading it ‘artificially’ like this, and could explain the stiffness deviating from the prediction based on span length. I tried loosening the corners and the stiffness did not change that much. I will try later with the gantry all the way to the side and away from the rail being loaded, and see if it is closer to proportional to the cube of the span.
The high loads are mostly to determine the stiffness of the rail itself. In general the displacement vs. load could behave ‘funny’ (meaning nonlinear) near zero – probably not for Z loading of the side rails but for other deflections maybe. Even if there is no play or backlash in the normal sense, it could potentially be measurably softer near zero, which would be interesting if I get to that point.
Yup, that funny elastic zone is where we work. I had thoughts of pre-loading the rails a bit but I don’t think I could make it a linear relation so I will just leave it for now.
Tests from the rail with a load applied at the tip of the endmill will tell you a lot more than individual rail loading. Individual rails should act very close to what the tables and calcs spit out, but the system as a whole is more important. As in how much does one rail move when 4lbs of force is applied to the endmill, short rail vs long would be cool to see.
Need to tidy up the cables and find a decent home for the power supply and RAMPS board, and we’re in business! Oh there are so many things I want to try…
