*3.2. Zund*

Aluminum flexure (2024 0.51 mm THK) was machined on the Zund G-3 L-2500. Machining was carried out using a 0.76 mm diameter end mill with amorphous diamond coating (Harvey Tool 72030-C4). The end mill was driven by a 50 krpm spindle. A faced aluminum sheet was used to provide a level and rigid support to the stock. PSA tapes burnished with a small stainless-steel rod were applied to both surfaces of the support and the underside of the stock. The tapes were then bonded with CA glue.

The cutting time was 2.5 minutes that is comparable to the waterjet on the aluminum sheet 1.6 times thicker. Figure 5 shows the micrographs of the flexure (a), superposition of tool path flexure element (b), and the corresponding zoomed sections (c and d). Figure 5b shows that the overall geometry of the flexure matches well with the tool path. Magnified views of two segments of the mid-span and left end of the fourth and fifth loops from the top are shown in Figure 5c,d (flipped horizontally). They reveal that there are minor mismatches between the flexure element and the tool path. The maximum mismatch is about 0.1 mm. Figure 5d shows that the mismatches were also attributed to rotational distortion.

**Figure 5.** Aluminum flexure—Zund (CBA).

By comparing the magnified micrographs of flexures machined with the two waterjets and Zund, the match between the tool path and the flexure element appeared to be slightly better for the one cut with the MicroMAX than the one cut with the Zund whereas the match between the tool path and one cut with the OMAX 5555 ranked third. For the flexure machined on the MicroMAX, the mismatch was mostly attributed to undercutting at the end loops. All the horizontal segments of the flexure element show no observable rotational distortion about the axis perpendicular to the X-Y plane. For the flexure machined on the Zund, the mismatches were attributed to the rotational distortion of the several segments of the element about the axis perpendicular to the X-Y plane. Note that there were differences in the fixturing setup. During waterjetting, tabs were used to strengthen the flexure element. They were then cut away after machining of the flexure element were completed. On the other hand, the Zund machined the flexure without tabs. The support was provided by the PSA tapes bonded together with the CA glue.

A 0.5 scale flexure (2024 aluminum 0.51 mm THK) was machined on the Zund by using a 0.38 mm diameter end mill with amorphous diamond coating. It took about 8 minutes to cut that part. Figure 6a shows the half-size flexure. There was only a negligible distortion of the flexure element. One of the most important attribution to the success in cutting the delicate flexure was the use of the diamond coated end mill. According to Toress et al. [9], the fine grain nanocrystalline diamond (NCD) coating, with average grain size 30–300 nm, reduced the thrust and main cutting forces to less than 50%, compared with uncoated tools, when machining 6061 T6 aluminum. One of the failure modes was delamination of the coating. After delamination, the tool looks similar to an uncoated tool with worn and/or broken tool corners and adhered workpiece material. End mills coated with the NCD experienced delamination about one half of those coated with the fine-gained diamond (FGD) counterpart. At present, the materials most suitable for the Zund are limited to relatively soft materials (At present, aluminum is the only metal recommended by Zund (http://fab.cba.mit.edu/content/tools/zund/manual.pdf).

0.5 scale aluminum flexures were also machined on the MicroMAX. A beta 5/10 nozzle with IDs of a diamond orifice and a mixing tube, 0.13 and 0.25 mm, was used to machine the flexures. Several cutting parameters were set to pump pressure, 345 MPa, garnet 320 mesh at a flow rate 45 mg/min. For the powdery 320 mesh garnet, however, it did not flow well under gravity feed. A proprietary process and a novel feeding apparatus were developed to enhance flowability of the fine powdery garnet under gravity feed (patented). Figure 6b shows the waterjet-cut half-size flexure on 6061 T6 aluminum. Its element shows no observable distortion.

(**a**) Zund (**b**) MicroMAX

**Figure 6.** 0.5 scale aluminum flexures.

Subsequently, an experimental 4/8 nozzle with a φ0.1 mm diamond orifice and a φ0.20 mm mixing tube was used to machine 0.4 scale flexures. Most of the cutting parameters were kept the same except that the mass flow rate of the 320 garnet was changed to 34 mg/min. 0.58 mm thick 304 stainless steel was used as the material. Figure 7 illustrates the micrographs of 0.5 and 0.4 scale flexures machined with the 5/10 and 4/8 nozzles, respectively. Both showed little visually observable distortion.

**Figure 7.** μAWJ-cut 0.5 and 0.4 scales stainless steel flexures (0.58 mm thick).

Attempts to use the 4/8 nozzle to machine a 0.33 scale failed. At that scale, the gap between the flexure element was 0.25 mm while the minimum achievable kerf width cut with the 4/8 nozzle is 0.23 mm. The cross-section of flexure element was 0.17 mm (width) × 0.58 mm (thickness). In principle, the 4/8 nozzle should be able to machine the 0.33 scale flexure. It turned out that the large-aspect-ratio flexure element was insufficiently stiff to maintain its shape.
