Micro machining a gem

Anybody who knows me is very well aware I hardly pass on a challenge, so when I heard about the Bantam Tools Micro-Machining Challenge I had little choice.

Out of the three key aspects of the challenge, I decided to start focusing on surface finish: I needed to find out a model capable to put on display the reflections I could machine out of the tiny 20mm thick block of aluminium I had lying in my hands.

After a couple of days of thinking and sketching different shapes on my ideas block I decided a diamond could be both challenging and great to put on display on my work desk. There are many different shapes into which a diamond can be cut, but I believe the most effective is what I also consider the most classic one, the round shape.

Creating the 3D model wasn’t too difficult, my skills with Fusion360 modelling have become quite decent lately and the first rendering took me only a few hours.

I knew the milling process would be very challenging: I had little experience with aluminium, the shape required the flipping of the stock and I had no room for failure as I had material for one attempt only.

A start from the top

I decided to work on the top surface first, trying to remove no more than a 0.1 mm of material so to be able to use some double sided tape to keep the item in place when performing the last operation on the other side.

So the first setup in Fusion360 had the stock material defined as fixed size and the model placed at 0.1 mm from the top and in the lower left angle, aiming at easing the alignment when flipping over.

Obviously I started with a 3D adaptive clearing strategy with a 3.175 mm flat 2 flute endmill, for which I set a 1 mm roughing step down and a 0.2 mm fine step down with a stock to leave setting of 0.25 mm, both axial and radial: this should have left enough material for refinement but not too much to stress the thinner endmill I was planning to use.

After fiddling with many web sites and tools and simulations I decided to go with a 330 mm/min feed rate, 220 mm/min ramping speed and 16400 RPM.

Looking for the best surface finish I had to rely on Fusion360 simulations and my imagination: after uncountable changes, I decided to go for a radial pattern with a 0.5 deg step angle using a 2.5 mm ball 2 flute endmill. As the clearing strategy has only a tiny stock leftover I decided a faster speed would have left a shinier surface so I went for 16400 RPM at 350 mm/min with a plunge speed of 70 mm/min to avoid any brutal hit on the stock material.

I wanted a very shiny top surface, so I went for a radial pattern using a 3.175 mm flat 3 flute endmill which I pushed to 20000 RPM at 400 mm/min with a step over setting of 1 mm.

It was time to engrave the Bantam Tools logo, but I decided to add a little touch: the roster eye would be drilled with a 0.6 mm micro drill 1 mm into the diamond. At that depth, it shouldn’t have made any difference, but I went for a chip breaking strategy nonetheless.

The roster profile would have been made out of three separate engraving passes, each one 0.05 mm deeper into the material using a very delicate 0.1 mm 15 deg v-bit I usually use for PCB manufacturing. At a 12000 RPM and 300 mm/min feed I carefully set the plunge feed rate at 20 mm/min: boring to death, but safer for the tool.

The whole top side took 1 hour of machining time, but more than 3 days to elaborate and double check any operation.

The bottom, at last

It was at this point I had to go all-in: the stock material needed to be flipped along the Y-axis and any misalignment would badly reflect on the final result.

I put plenty of time in the effort to minimize any error: multiple measurements of the stock piece, accurate clean up of the machine fixture plane and fine adjustments on the Fusion360 setup.

To ensure the last cutout operation would not send my precious gem flying inside the machine I put a piece of double-sided tape on the top surface, now facing the machine plate, but I discovered I had a much greater space than I was expecting: instead of 0.05 mm gap I had a 0.15 mm, which required three pieces of double sided tape one of top of the other.

After triple checking I decided I had to give this a go, so I once again started with the 3D adaptive clearing strategy, with the very same work parameters I had used on the top: a 3.175 mm flat 2 flute endmill, set for a 1 mm roughing step down and a 0.2 mm fine step down with a stock to leave setting of 0.25 mm, both axial and radial. You can admire the result of this step down below.

Once this step was over I also tried to ensure I had enough leftover material to keep the piece in place for the next operation.

Next operation in the list was the finishing of the sides, so I installed the 2.5 mm ball 2 flute endmill and have it running on the 3D radial pattern for 1.5 hours: when it finally parked away I discovered with horror the surface was far from being shiny and smooth as the top surface. On the contrary, it was very rough and opaque, with bits of metal everywhere. It was only after I went back to Fusion360 that I discovered I had forgotten to set the spindle speed for this operation: instead of running at the intended 16400 RPM I had it running at only 12000 RPM: a huge difference which, probably combined with vibrations and resonance left me with a horrible result. Luckily for me the piece was still perfectly sitting in its position so I had the operation parameters set at16400 RPM and set the stock to leave parameter to a nice -0.05 mm(both axial and radial): enough to have some material to remove but potentially not noticeable on the final piece. I also decided to bump up the feed rate from the intended 350 mm/min to 600 mm/min to reduce friction and heat on the surface: this decision also reduced the execution time of the second attempt down to a mere 50 minutes

You can find pictures down below capturing both the ugly result of the first attempt and the outcome of the resulting fix.

Now that I was happy with the bottom part of the diamond I was ready for the last and riskiest operation: the final cutout.

Remember, I was relying on three pieces of double tape stuck on top of each other to keep the piece in place while the final 2D contour operation slowly bit (ramping at 2 deg angle at 300 mm/min) into that 0.5mm thin piece of leftover aluminium: if the tape didn’t have enough glueing action I would see my tiny gem thrown away by the 3.175 mm flat 3 flute endmill spinning at 20000 RPM, probably with big mark on the side. Luckily for me what happened is the stock material was actually thrown away: yes, I was so stupid to put the tape under the diamond but nothing under the actual aluminium block.

More than 3 hours after beginning the mill operations on the bottom side I was now left with what I considered a splendid gem made of aluminium.

Showoff time

So, after a grand total of more than 3 hours of machining time, 2 facepalm moments, 4 days of Fusion360 designing and planning, and 5 minutes of polishing (top face only), allow me to introduce the most precious treasure of my one-item collection: the eye of the bantam!

With a radius of 24.44 mm, a height of 15.8 mm, a 2 ct weight and it’s classic, round shape cut based on a hexadecagon, the eye of the bantam stands out from all the other aluminium gems on this planet and is considered a unique object by both the Tooth Fairy and the Queen of England!

2 thoughts on “Micro machining a gem

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s