Eureka Moment

When I worked at Boeing, I was a CNC Maintenance Tech. I came to work and listened to the brief safety meeting, was given my tasks to begin work. As I riding my work tricycle out to the gantry mill, I heard this loud noise coming from a different gantry machine nearby.

When I arrived, the noise was deafening so I put on ear plugs and ear muffs. As I walked up, I saw a very small part which I estimated to be about 45 feet long and 3 feet wide attached at the bottom much like you would hold a tuning fork from the bottom.

This part was vibrating so intensely, that it was shaking the fixture holding the part, the machine frame, the 8 feet wide 8 feet deep by about 75 feet long concrete foundation. When I stood about 10 feet away from the machine base, I could feel the vibrations shaking the soles of my feet though my heavy work boots and 2 sets of insoles. This was worst machining I ever saw in my life. I said to myself there has to be a better way.  I finished both of my shifts that weekend.

The following Monday morning (I usually worked second shift), I was reading through my mail as I was interested in machining knowledge & skills for my job.  I read the latest issue of “Tooling and Production” and was skimming through when I started reading an article about polymer enhanced concrete and its vibration damping. When I learned about the amazing vibration damping attributes, I started thinking about the worst machining I witnessed 2 days prior.

In that Eureka moment, I realized I had almost all the basic skill sets to design, engineer, fabricate, test and finish machine [grind smooth flat surfaces] a vacuum a fixture base plate to hold the wing part at work. My Dad was a carpenter so I had helped him pour concrete in a dozen different projects on the farm growing up. Having commissioned very large vacuum heat treat furnaces, I know how to design for vacuum. I designed in large diameter piping for highly efficient vacuum pumping.  I could fabricate [saw, assemble and align just as I had done with my Dad] the precise engineered wood forms to tight tolerances for casting on a simple but accurate table saw. I also designed the matching aluminum cooling plate with integrated vacuum ports to match the vacuum ports below and the rubber vacuum seal groove design for optimum effect. Also, I had experience precisely bending tubing from my days when I worked on construction. 

The only portion I was missing was an optimum concrete mix to greatly increase stiffness, 4.5X compressive strength, inherent natural porosity to essentially zero [leak tight] from 7-9% porosity and design the mix for a very high resistance to heat flow or essentially be almost an insulator.  Also I wanted the mix to almost match the TCE of steel so it expands almost the same amount.

I started researching standard admixtures, mix design, and reached out to a few concrete engineers asking how to design my mix. With the help of 3 different engineers, I was offered samples of admixtures by the companies they worked for.  I started worked in my double garage with a small concrete mixer, a standard carpenter’s table saw and all my many hand tools.

With a very systematic “design of experiments” approach that I had learned about from my research, I worked at creating my mix designs, optimizing the mixing sequence, weights not volumes and started creating 24” by 2” by 4” ingot samples to test.  I also incorporated post tensioning for extra added strength.  Within 6 months, I could pour any shape I desired with embedded steel threaded inserts so I could “bolt into” or “through” my castings. Initially, I started by creating castings to extend the “ways” [aka linear guides] of my Bridgeport knee mill and 2 standalone grinder bases. I could now machine and grind bigger, longer parts as the moveable axis would now travel twice as far.  

When that worked, I bought a much bigger scrap, cast iron, 16 feet long, 6000# casting for a song and started by adding roller ball linear guides along with a rack & pinion drive for motion.  The new X axis was a double offset, 2 layers of 1” thick steel plates which I machined by hand hanging off of a small steel gantry hoist on top of my newly enhanced Bridgeport knee mill in the garage. The plates are annealed, straight ground and bolted together so they can slide upon the dual linear guides bearings on the 16’ casting.  The 16’ casting with a horizontal axis casting was mounted on top of a custom poured, high strength, and post tensioned concrete casting 24” high by 48” wide and 20’ long which I poured myself.  I used an old, $100 fork lift nobody wanted and fixed it up so worked. I also bought a used 3 ton rated engine hoist which nobody wanted for $200 to lift the opposite end of this 6000# casting up on top of the 24 inch high concrete foundation with 2 helpers.

I hired a local Computer Aided Design (CAD) young man whose dad owned a machine shop not far away for about 3 months.  I sketched out portions of the rest of the machine design, did all the force and moment calculations on the fly.  I had him draft it up in CAD so we had very accurate dimensions on all parts.  I could then saw, cut, drill and bolt up the steel frame.  I also designed and fabricated two very complex 3 meter vertical columns which sat on two 8 foot long polymer beams. These held the moveable Y and Z axis perpendicular to the 16’ long horizontal axis which I mounted the new temperature controlled vacuum fixture on.

Once the milling machine was assembled, aligned and the controls moved the motors which moved the axis, it was time to test the machine. You start with tooling foam which is hard enough that it replicates the cut surfaces exactly.  You can measure the cut features after to see how far off the mill is.  As long as the mill is repeatable, then you can laser compensate for 99% inaccuracy in all axis.  Then re-cut foam again to verify the laser comp values are correct.  Then it’s time to machine a soft material like aluminum to see if it is rigid and repeats and then measure your “as milled” part and then you can comp the program with offsets to dial in the features for accuracy.  I achieved ± 0.0012” over 10 feet in X, same over 80” in Y and same over 36” in horizontal Z axis.

I initially used a scrap aluminum plate of 24” by 24” by 1″ thick.  We milled one large pocket 22 by 22 by .940 or just shy of 1 inch deep for a very thin layer left on the bottom which was straight after we measured flatness with a machine tool grade straight edge.  

We then went on to mill out a 48″ by 48″ by 1 inch thick scrape plate with a pocket of 44″ by 44″ by the same 0.940″ deep a few times.  It came off with a thin 0.060″ face left over and it was also very straight; not warped which is what I wanted to prove. This is extremely hard to do on any mill with dry machining as the heat builds up quickly. I wanted to prove to myself this was a repeatable process which it was.

Initial testing with dry machining process will work down to 0.060″ which is very susceptible to warping. The thinner the material is, the easier it is to warp, twist and bow.  I used the 0.060″ as a worst case scenario. This 0.060″ plate is 6.25 times thinner than 0.375″ (9.5mm) skin edge right at the wing tip. 

We ran soft tap water through the copper tubing in the fixture that was measured at 52°F to 54°F at 38 psi pressure through the cooling of the fixture so it definitely was cool to the touch.  We did this in the middle of summer with about a 75 to 80° F degree day.  I calculated the cooling water flow rate at 0.61 gallon/minute.