No striations, no shifting, no deviation. Everything turned out perfect! At last! View the whole set right here.
Archive for the ‘CNC’ Category
CNC finally working? Also: robot parts
Friday, April 8th, 2011Calculating your milling speed, feed rate, and more.
Tuesday, April 5th, 2011Custompart.net has a great widget to help you with all the math. Here’s an embeddable version!
Milling Speed and Feed Calculator
CNC calibration tests continue…
Tuesday, April 5th, 2011Well… it’s getting better. There was a lot of play in the machine. That means it’s loose and pieces move when they shouldn’t. For example, push on a part of a CNC machine and it shouldn’t budge. I tightened everything I could and it seems to be better. Now I suspect the problem is that the stock material is moving when the bit touches it. I need to put more thread inserts into the table and secure my stock better.
Leaning to mill on a CNC, one mistake at a time.
Monday, March 21st, 2011An accident is a series of mistakes. they key is to know when you are about to make one too many.
Last night I broke the smallest bit in my set of milling tools, and I should have seen it coming.
I have to go through a lot of steps to cut something on the CNC machine: First, a design has to be created in software like Alibre. Once all the interconnected parts are shown to be Ok, I take each piece in turn and export it to DXF, and then import those DXF into Cambam. In Cambam I set up all the machining operations – cut here first, then cut this bit but only so much, then profile this part, etc. Then the Gcode is walked down to the CNC. Once I fire up EMC2 in Ubuntu I power up the machine and I “home” the machine. Homing is how I tell the machine where 0,0,0 is located. That way it knows “I go X from here, drill down Z, move sideways Y, etc” and I get the shape I want in the spot I want. Generally z=0 is the table top.
At first I cut an involute gear, and everything was great. I was so pleased with myself. One step closer to profitability! There were a few “huh, that’s wierd,” followed by an equal number of “oh well, I’ll figure it out later. This software is pretty buggy.”
Drunk on accidental success, I tried to jump ahead. Jumping ahead is something I’ve been prone to all my life – I always wanted to go faster than my limitations allow. My next “test cut” was audacious – a 3D profile of the Lululemon logo. The logo was supposed to be 1/4″ deep in 1/2″ MDF. The machine did exactly as ordered …except that it cut the shape out of the bottom half of the MDF. Pleased that the profile had worked at all, I ignored the curious mistake and ploughed right ahead.
My last cut for the night was a piece for a robot I’ve been building. Again, I used 1/2″ MDF. The image on the computer didn’t look quite right but Hey! I’d learned not to trust it by now. I carefully homed the device, then hit the start button.
As best I can figure, because of a series of mistakes, the machine believed the design to be 1/4″ lower than it actually was. The bit, turning too slowly for the rapid movement, slammed through three inches of MDF before snapping in two, leaving the bottom half lodged in the MDF like the sword in the stone.
To prevent this from happening again, I’m going to be using CutViewer from now on to preview my work and check for mistakes. I’ve also learned how to set up the stock material (aka MDF) in Cambam and get the cuts to look right from start to finish. With luck I can replace my broken bit for $10 and never go through this again.
What kind of expensive learning experiences have you had lately? Share with the class!