Week 6: Computer Controlled Machining

Design, mill, and fabricate something BIG using computer-controlled machining.

The group assignment this week entailed completing EECS's milling safety training and testing the runout, alignment, spindle speed, feed rates, material properties, and toolpaths of the lab’s CNC mill. Our LA, Anthony, provided the safety training while also demonstrating several of the mill characteristics for the group assignment. Anthony had recently assembled the mill and had already honed feed rate, spindle speed, and cutting depth for the material being used this week. Experimenting with these variables demonstrated that larger diameter end mills can be run at higher feed rates and cutting depths without sacrificing cutting quality. Smaller diameter end mills need to be run at lower feed rates and shallower cutting depths to limit flexure of the cutting tool, which effects both cutting quality and tool longevity. Machine runout is a measurement of cutting variation along a tool’s cutting surface due to misalignment. Prefect alignment of any rotating machine cannot be achieved, so all mills will have some amount of runout, but it can be minimized by ensuring the cutting tool is concentric with the milling head and aligned in the vertical axis.

While testing the mill it was apparent that OSB is a less than ideal material for milling due to its inhomogeneous structure, which caused entire wood chips along the cutting edge to rip out rather than cut cleanly. This was further aggravated by the use of up-cutting end mills, which clear chips by pushing them up to the material's top surface. This method of chip clearing had a tendency of detaching entire wood chips adjacent to the tool at the top face of the material; resulting in a poor finish appearance along the top face. However, it was found that if the OSB's weather resistant coating was faced upwards (green coating show in the image below), it increased cut quality along the material's top face by acting as an additional binding layer for the wood chips.

We also experimented with different methods for securing OSB to the machine's bed. The use of composite nails which can be milled through without harming the end mill were both effective and quick to install. They proved effective for securing large components, but were difficult to adequately place for smaller components. Another method to secure components is by adding tabs to the G-code (see image to the left). These worked great, but did require additional hand tooling and sanding post milling to achieve a smooth surface finish. Similar to tabbing, leaving a thin layer of material by not cutting completely through the stock material is referred to as onion skinning and can also be used to keep components in place while milling. I will show this method in section 6.4 of this webpage.

Due to my schedule and machine availability, I ended up milling this week's assignment on the C. R. Onsrud 3D milling machine in the architecture wood working shop. This is an incredible machine that is fitted with a 4 ft by 8 ft vacuum bed and automatic tool change. The machine is shown in the below images.

The above images show milling in progress and the end product. A few of the smaller components detached from the vacuum bed during the fourth tool path, so the fifth tool path to remachine joints was ended early. This meant final joint preparations would have to be done by hand.

As with all building supplies, my plywood was not perfectly flat nor did not have uniform thickness. The thickness tolerance was +/-100 mils with the lower limit around 0.6 in. While developing the G-code in MasterCAM a stock thickness of 0.6 in was used knowing a thin onion skin layer may remain in some areas. Using a utility knife and sand paper the onion skin was easily removed.

Having not completed the final tool path to remachine the dog bone joints, I had to use a small file to remove excess material in the joint corners. In this image an unfinished dog bone joint is shown to the right and a finished joint after hand tooling is shown to the left.

After removing the onion skin, finishing the joints by hand, and some light sanding I had a finished final product; a frame to support the induction machine and pipe section for my final project.

As part of my ongoing research in the RLE, I am setting up a system to analysis refrigerant compressor loading. Instead of purchasing the suction and discharge ports for the compressor, I decided to machine them with scrape aluminum from the RLE machine shop. The shop is outfitted with a two axis CNC Avid mill which is shown below. I created my model in Fusion360 (CAD file is hyperlinked in the below rendering), but the G-code for machining was developed in Rhino.