I ended up the last update saying I had poured epoxy into the finished side of the chunk of aluminum I've been working with. I concluded with a picture of the epoxy-filled side after it had cured a while and said it would sit overnight.
Based on the experiences with the first side, I was going to do four passes: two rough and two fine, but very different from what I had done on the first side. Essentially, I broke the one rough pass from the previous side into two. There are two similar algorithms that my CAM program use to create tool paths, called waterline and contour. The major difference is that waterline takes a series of layered cuts down the vertical extent of the part while contour makes one single path around the part. For waterline, we (the user) defines how thick those layers are and how close they're allowed to get to exact size of the part. For contour, we have some control over how deep that one path around the part is - and we also set how close the cutting tool gets to the exact size of the part.
I chose to do a contour cut around the part and then do a waterline just deep
enough to ensure the top of the connecting rod was the proper distance (1/16"
- .062') below the raised, circular bosses.
Cutting a single slot, the depth of the part can be a problem because it can be hard for cooling fluids to get into the slot. In the case of aluminum (which this part is) that can cause the stock to melt onto the cutter. In addition, on one side of the slot, the action of the cutter is to pull itself into the stock while on the other side, it can push itself out of the stock. This can set up a vibration in the cutter and the mill, depending on how rigid the machine is.
As usual, there's a possible trick to get around this: cut several thin layers and not one pass to full depth. It turns out it's simple to change the single pass file using a plain old ASCII text editor. Simply change the depth of the pass from full depth to what you want (in this case it was change a single 0.220" deep cut to seven consecutive .031" passes) using block copy and paste in the tool path file, then editing every block's line where the cutter drops vertically to the right new depth.
It turned out even doing those thinner cuts I had a vibration and resonance
issue that caused my machine to vibrate so badly that a motor fuse shook loose
and shut the spindle motor off. It didn't mess up the position badly,
though, and using cutting oil instead of water mist helped.
Both rough passes: contour on the left and the waterline for three layers on the right. For the fine paths, I went back to one contour pass at full depth, just trimming away the margin around the part. That made the cut in the stock a little wider and had no vibration. For the waterline file, I just trimmed the margin around the part.
This went smoothly and by Tuesday the second side was done. Time to soak the epoxy in my shop toaster oven. It broke apart easily.
Those yellow chunks in the front left are the epoxy, and the leftover, machined metal is in the back. The first side machined is facing you. The dimensions I've checked are all within a reasonable tolerance of the intended size (.005" or less), and while a long way from done, if not for the machine screw-up on the first pass, it would be usable. The machine screw up is visible in this photo, at the lower right end.
Another view from the second (back) side. It's about .075 extra cut away along that one (top) edge.
I've done only one other piston connecting rod; the one on my Webster engine. I did that one a bit differently - more conventionally. I held the blank rod onto a tooling plate with two #10-32 screws and ran contour cuts around the rod. Then, to cut the equivalent to the circular bosses on the 1 by 1's conn rod, I cut circles around the ends with a few lines of G-code I wrote. For grins, here are the two rods, side by side, Webster in white.
Looking at the whole process of cutting this practice connecting rod, I'm not sure that doing the epoxy trick really bought me anything. I'm thinking of going back to the approach I used on the Webster's rod. The advantage to doing it with the epoxy method is I've already debugged the tool paths. That should speed things up considerably. No matter what else, there are a couple of fiddly little operations that need to be done on the rod after all this so those will be done the first time on the real one.