3 Axis CNC Mill
Shoo! I have not posted in a while. This is most likely do to the retarded number of units going into my term. But, I am having the time of my life, so it is most likely worth it.
These days I’ve been mostly working on my 2.72 project, which is the design and construction of a CNC desktop mill. Most students in the course design a high precision desktop lathe, but our group decided the lathe had been done too many times and had too few design variables left over. In light of this, we followed a senior student’s thesis for the professor, which was a small, easy to build CNC milling machine.
Boy, what a class. I’ve learned more from this course than I have from entire terms, and we’re only a month into it. In fact, our first design review is in several hours, and I’m somewhere between excited and completely boned, a pretty typical state for any MIT student at a checkpoint. That being said I’ve got an extremely strong team, and I believe we’ll have this one on lockdown my friends.
Boy, real mechanical design is crazy fun. First of all, I’ve spent the past two terms in electronics and materials land, so I’ve missed a lot of mechanical thinking. Revisiting it all is great fun though. An example is the error analysis of the machine, of which we need a nice list in about four more hours. Error is anything that puts our cutting tool in the wrong place, and while you can never model everything, you can come close enough to be an engineer. Our list so far includes the following:
Thermal growth
Vibrational modes
Bearing run out
Cutting tool collet run out
Stepper motor resolution
Lead screw play
Connection compliance
And needless to say the list goes on. What’s fascinating is the true intensity of estimating any one error source. Let’s just look at bearing run out on the cutting tool spindle. In order to ensure the bearings don’t flop around, a pre-load must be applied equal to the manufacturer’s specifications. This is usually 2-5% of the max load, so a pretty easy number to calculate and build in. With that number present, we design a nut, spring washer and shoulders smartly such that a spanner wrench can be used to implement the pre-load exactly. But the fun doesn’t end there! As the machine makes a cut, the friction created by the pre-load heats up the bearings. Further, operating speed alone creates bearing heat through friction, and taking a cut into material with the tool creates a lot of heat which carries through a variety of paths into the air and spindle. All of this means thermal growth of the spindle and it’s associated pieces. Thermal growth has to be allowed, as otherwise enormous pressures will be generated internal the shaft, utterly destroying it and the bearings. So the thermal growth is expanded into the bearing pre-load, releasing the original compression applied across the bearing. This means increased run out but decreased friction, rolling our heat generation off a bit but increasing vibration. Somewhere, deep inside the math, exists an equilibrium point that the machine would truly run at. But finding it is nigh near impossible. And as engineers, we must have an answer for that which is not knowable, and we trudge forth with the tools we have and the knowledge of their limitations to find: The Engineer’s Approximation.
So we FEA it, we run some empirical numbers on it, we examine a similar system, and we trust our gut. And then we defend it to one of the best engineers in the world. Just another day for an MIT student.
With all this in mind, I’ve mostly been doing modeling of the electronics system, because electronics and I get along real, real well. That and most MechE kids don’t know a MOSFET from an op-amp, a real shame if you ask me. I’ll rant at you about the horrors of the course 2 curriculum at MIT later, for now I’m going to do some saturation limits on a power transformer and bridge rectifier to supply our drive motor.
Expect some specs, diagrams, and how-to’s in the near future. For now I’ve got to tool hard and hearty.
There’s now a full page following this project! Check it out over at http://codydaniel.wordpress.com/cnc-mill/
