Tool Selection
Hello,
Tool diameter is largely a preference, within the constraints of machine capabilities. For me, 1/2" tool diameter is preferred for wood products, 3/8" for plastic or aluminum. The 1/2" choice for wood products is because that size is a commodity size, it is only about 12% more money than 3/8", and allows for more chip load. The 1/2" tool is also much stronger, so it is far less likely to break than the 3/8", the finish is also better at the elevated feedrates I typically run at as well. For your product mix, I would use a 1/2" compression tool most of the time, unless product details demanded a different size, such as 1/8" max inside radius or similar specs on the customer prints.
The first CNC I put together used Porter Cable 3-1/2 hp routers for spindles, they had a both 1/4" and 1/2" collets, which still is pretty limiting, but workable. That type of spindle is going to be the limiting factor in tool selection, as a 1/2" compression tool with 1-1/4" cut length is quite capable of stalling the router long before you get to the designed tool limits. My current machine is not a DIY effort, so I have to dig through the cobwebs in my brain to get in your shoes.
In short, use the largest practical diameter tool, use the shortest cut length. The tool will have more flute than cut length, so for most applications, a 1" cut length tool (which will likely have 1.02 to 1.032" cutting edge) is valid for 1" material, provided you can hold Z reasonably well. If your machine deflects easily in the cut, a bigger diameter will impart more deflection, but be more resistant to breakage, you have to decide on the trade offs.
What feed rates is your machine capable of? I usually program feeds and speeds based on material and chipload. Different tools have different chiploads, as do different materials. Bigger tools can carry bigger chiploads as a rule. For a 1/4" compression, in 1/2" mdf, I used to run about 18K rpms at 16500mm/min or 650 ipm feedrate, which is about 0.018" chipload (feed per flute per revolution). Bear in mind that my spindles are modestly beefy router spindles with hsk63 tool holders. Back in the old days, the old Digital could barely make 400 ipm on a good day, and the spindle mounting was not all that rigid. When they upgraded to a Perske with variable frequency drive, that help alot, but still, the machine was not all that rigid or fast.
Eric Neumann
http://www.cncrouterworks.com