The simple threaded mandrel that I made to hold the heads for their turning operations worked OK on the lathe since the cutters created only tightening forces. The heads do end up pretty snug on the mandrel, though, and have to be carefully 'wrenched' off. The milling operations, though, create both tightening and loosening forces; and so I made a second, threaded and expandable, mandrel for use on the mill.
I created a few single and double groove test disks in order to try my hand at generating code to eventually undercut and fully radius the fins between the valve towers. I was pleasantly surprised to find that my CAM performed well on these new (to me) operations. After coming up with the machining parameters for my fin cutting tools, it was time to generate the tool paths to continue the machining on my two test heads.
I then started a week long trip to Hell. It all began during my CAM simulations when I began getting tool collision warnings created by my fin radiusing tool. These warnings surprised me because in my CAM these particular operations work only with curves and aren't supposed to consider the part or workpiece at all. But, they were legitimate, and I eventually realized they were happening because I hadn't fully taken my tooling into account in my head design. The troublesome areas were where the profiled fins in the tower valley and intake/exhaust flange areas blended into the existing fins in the lathe-turned blank.
About this time I suddenly came down with some really bad flu-like symptoms. Instead of resting, I decided to make a new cutter to see if I could avoid redesigning the head, So, I started with a carbide reamer, a diamond hone, and a handful of Harbor Freight diamond Dremel points. The carbide decision was irrational and totally flu-related. Anyway, two full days of work later I had a working tool with perfect contours and proper cutting edge clearances, but it still wouldn't pass my simulations. I had also made a .031" corner rounding cutter out of carbide reamer.
At this point, and in no physical or mental shape to make such a decision, I suddenly didn't like my head design anymore. So, instead of fixing the tool interference problems with the current design, I started over - again. This time, knowing my CAM was capable of what I needed, I became obsessed with the minutia of perfectly blending the fillets, radii, and undercuts together at all the complicated intersections I was creating in the new design. Getting the necessary tooling clearances simultaneously with the esthetics I was looking for turned out to be much more complicated that I had ever imagined. I spent dozens and dozens of hours on such incredibly small details that probably only I would notice.
All the CAD file saves I was doing caused me to bump into a SolidWorks bug. My file sizes had begun growing exponentially even though the many design changes I was continually making to the design were minor. This turned out to be known bug with the old (2007) version I'm using. My one meg design file had blown up to some 200 meg and 199 meg of it was SolidWorks bloat that slowed my computer to a crawl. Just when I was almost finished with the design, I had no choice but to delete all my cumulative design work and start over. I had never run into this problem before, but then I had never made so many changes to the same file either. Several days later, I had a design ready to submit to my CAM. The CAM work went smoothly, but after my flu symptoms began subsiding, the jpgs of my original design started looking better to me than my latest design. Unfortunately while in a mental stupor earlier, I lost all my early design versions in a major delete accident.
After getting a successful simulation, I set one of my two blanks up on the mill, started the program, and held my breath. To my amazement, the part came out exactly as I had visualized it. I found three minor tool gouges at the rear of the part that my CAM, as expected, had failed to flag. I was really lucky I didn't break either of my one-of-a-kind cutters. I tweaked the design slightly to improve the tool clearances before regenerating the tool paths to start the second blank. The second part ran cleanly and looked as good as I had hoped. After seeing the two parts in actual metal, I'm very happy with the result.
There are four more operations needed to complete the head, but these are simple compared with what I've already done. Figuring out the work holding for them will be the most difficult part as all four operations are at unique angles to the axis of the head and new coordinate systems and fixtures need to be created for each one of them.
I also plan to re-work the rocker arm support design as what I've shown earlier has only been a placeholder. That should be an opportunity for me to jump down another rabbit hole. -Terry
