First 'Real' CNC task : Machining the Fins

5 Bears Home Homebrew CNC bench mill

The toolpath was created, and the G-code generated and imported to Flashcut on my shop PC. Because the mill still doesn't have proper swarf guarding, I taped a few paper towels so as to surround the work and keep the chips out of the linear motion hardware, and mounted the KaVo block into the little 3" KURT vise.

If you have read of the projects on this web site, you'd find that just about every "first", such as a first engine run, first test, etc, has been comical, disastrous, or both. I was determined this time to break the jinx chain with this, my first CNC cut of any consequence. I don't count the simple engravings that I have done so far, they are just too benign.

To prepare, I chucked a 0.187 carbide end mill, and tightened everything down. With the toolpath entered and displayed into the Flashcut interface, I set a temprary Z=0 at about the position you see in the first photo, and started the CNC motion to check the path of the cutter at a level well above the actual block.

Flashcut allows you to override the G-code feedrates, so by selecting 10X speed, I verified that the toolpath was, in fact, what was expected. Everything looked good! I reset the spindle to the correct XYZ 0,0,0, and let rip.

The first cut made by the controller was a skim around the perimeter. The spindle then repositioned, and began to cut the first fin. Here is where the odd, angled tool entry became clear to me. The spindle assumed a position clear of the work, and about 1" to the right of the path shown cut here. It then diagonally plunged into the aluminum, and simultaneously curved around the left edge of this fin. The cut channel here is ramped. Very nice! The tool continued it's cut at 0.050" depth around the first fin, ultimately removing the ramp, and then retracted.

Repositioned to the second fin, it repeated the cut, until all six fins were milled on the first Z level.

The block after 0.050" removed. I now know I'll need some form of flood coolant to clear the chips if I want higher speed, otherwise, the cutter will load and break. I was forced to use a toothbrush and continually clear chips for the entire process, which became pretty tedious.

So far so good!

After four successive passes, the fins were close to their final form. When I created the roughing toolpath, I specified perhaps 0.004" all around for cleanup.
The Z level of -0.196 indicates the spindle is 0.196" below the original surface of the block, and the spindle is on its last cutting pass in the roughing phase.

It would have been smart to place a machine Pause command between roughing and finishing, so I could really clean the chips out, apply some lube, and prepare for the finish pass. I did the best I could, and during the transition from horizontal roughing to horizontal finishing, I boosted the RPM by about 50%. The machine doubled its feed (as I had programmed), immediately went to full depth (0.200") and finished off the fins in a couple minutes.

The mounting block, fresh off the mill, no work done except to clean off the chips. I tried to photograph the flanks of the fins, which were just about perfect. The finish was excellent, especially for a spindle that doesn't have the fastest RPM.

The toolmarks in the valleys between the fins are typical of aluminum... smooth to the touch, and in this case, attractive. A super-fine wheel brushing would work nicely, or some fine steel wool to break the edges on these fins, and the block would be ready to go. Still need to split it.

One more view! The fins dimensionally are close to perfect. I have a long way to go both with hardware, and my expertise with software, to refine the entire process.

Oh yes, the chain of buffoonish "firsts" has been broken! It went off without a hitch. Yay!

One last operation remained to make the KaVo clamp functional, and that was to slot the face. This was a no-brainer in my big mill with a 0.064" slotting saw.
Finally in place. I am pleased with the block that was just made. I'm not really thrilled with the "wedding cake" appearance of the entire Z-axis carriage, but function over form in this case.

The boring in the block turned out perfect. You know you've done a good job with a split clamp like this, when the absolute slightest torquing of any of the three screws instantly grabs the shaft. This spindle will go nowhere. The possibility exists to adjust Z by moving only the spindle, by releasing the clamp, rather than trying to jog the entire Z-axis into contact with a workpiece. Hmm that gives me an idea... how about a micrometer plunger device which will move the KaVo unit a precise amount? I'm still struggling with setting program Z = 0 before switching on the G-code.

At the top of the motor is the pneumatic input for the collet release. The air input for the bearing housing pressure is in the middle, next to the power input jack. It'll be interesting trying to plumb this spindle... I'll need an armored, reticulating sheathe to handle 9" of Z travel without fouling the cables and tubing.

If I can figure out a way to hold cutters perfectly oriented and rigid in a little rack of some sort, I could easily automate the tool change by positioning the spindle, releasing the current tool via a solenoid valve, reposition the spindle, and grab the next tool. That's what's fun about this type of project, there are so many ways to refine the process, both hardware and software.