After a couple of days of bad luck, Day 9 was the answer to all my prayers. Garry, technician for the Tech department, was back in school and I was able to discuss my progress with him. With only a couple of working days left I needed a fast solution to be able to cut the concave faces quickly. After showing which ideas I had already tried we decided that the way forward was to make a custom tool to insert into the CNC machine and then find a way to reconfigure the machine to make use of it. This was always the hypothetical ideal solution, as it would allow the machine to cut the faces extremely accurately However, without access to the right materials and without the knowledge of how the machine can be configured for custom tools, I was un-able to attempt it during the week.
While talking over the requirements of the tool, Garry pulled out a box of high speed steel square rods. It was as if they were made for the job. The small dimensions of the rods meant that it would be easy to get the required clearance that I needed to avoid surface collisions.
If we had had the time, we would have milled a slot out of a steel rod, finished the end and braised the HSS in accurately. But time wasn't on our side, so Garry used shortcuts and made the tool within 15 minutes. The slot was simply hack sawed out, the HSS hammered in, all braised up, and then the entire thing ground down. Then finally, the HSS rod had the end chamfered and ground into a cutting edge. It doesn't look pretty, but it's solid and does exactly what is required of it.
In the image above you can see the huge difference between the tools. The steep angle of the HSS rod means there is a huge amount of corner clearance. We don't have to deal with a fat bulky carbide tip. The new tool could cut concave faces with a tighter radii than I would ever require. The only slight downside which could be changed with a bit of re-grinding, is that the top face of the cutting tool is angled the wrong way. Although it is a boring tool, It can't cut a horizontal step. Therefore, this tool is ONLY suitable for concave face cutting as it is at the moment.
Cut down, the part is small and heavy for its size, and is like a present to me. The next step was to mount it in the machine and calibrate it. Firstly I loosely secured it into the tool turret and jogged it across to the chuck. I turned a point onto a rod of stainless and secured it in the chuck. Then, using 10% feed jog I slowly nudged the tool up to the point, and rotated and moved the cutting tip about so that the points met up. This should ensure the machine cuts accurately To make sure it was as accurate as possibly however I needed to calibrate the machine. I found old documentation under the machine about how to set tool offsets. Using this I made a new tool profile based off the original boring tool, and made test cuts to calibrate it as best as I could.
Finally, with the new tool ready to go, I took the code that had failed a couple of days previously and ran it. Before even seeing the result, one thing was very clear. The new tool cut the acrylic like butter!!! It came off so smoothly, the machine sounded a lot happier.
When the final part came out it was beautiful! The machine had sounded much happier and the surface quality proved that. It was so even and smooth all over, with no remelted swarf or surface damage, it was just the condition I needed to be ready for polishing.
The shot below also shows that the tool was cutting nice and accurately as well. It's difficult to meassure a the diameter of a concave face because there isn't anything to push the calipers against, however lining the edges up by eye told me that it was close enough. In reality, the important variable is the radii, which was totally defined by the CNC cutting path, so regardless of the resulting diameter, the surface would work anyway. The depth is irrelevant seeing as these elements all work on the thin lens principle, where the lens thickness is within a certain tolerance meaning it is so insignificant to the performance of the element that it can be ignored.
Hey who notices a difference? Look, a new flat screen! Why? Because the old CRT screen tried to blow up! Did I honestly expect a day to go perfectly with no set backs? 20 minutes out to get a new screen and install it wasn't too bad, but the fact the screen had decided to pack in mid-run on the machine meant I had to restart the cutting process, woo......
The next steps were to carry on cutting the concave faces! I started running some of the code files I'd written but stumbled into a minor problem. The billet length was sometimes to much for the machine to handle along with the working distances I had left myself in the code. So, when I parked the tool, it was too far away for the machine to work, as it ran out of room on the axes. During the code runs, rather than tell me in a simulation that it would be a problem, it would stop dead in the middle the a run and show me the error shown below. Why on earth the machine doesn't have limit switches on the axes is beyond me. The error requires the program to be terminated, the machine to be reset, it to be dumped into manual mode, and for the servos to be jogged continuously until the axes were freed up again. Annoying, but easily fixable by reducing the working distances in the code i'd written.
Before the end of the day I took the time to quickly buff up the concave surface I cut earlier. Lightly polished, the surface looked amazing. The higher quality cut had helped make the polishing process easier resulting in a clearer image through it. With only a couple of days left to finish all the elements and get the lenses assembled, I had everything I needed to press on and sprint for the finish line.
