Virtual Impactor

I designed and machined the first prototype from 2 inch waterjet segments which allowed three critical geometries to be independently varied. Once the virtual impactor was machined it was placed in a closed loop setup with orifice flow meters and an aerodynamic particle analyzer connected through isokinetic samplers. Several parts of the system were custom cast using a vacuum chamber I built from a crock pot and venturi type vacuum generator.

The device was tested under open loop, and closed loop conditions. The setup eventually evolved to pull air rather than push air through the inlet. In the open loop setup, material was captured in vacuum cleaner bags and analyzed with a Malvern Mastersizer diffraction type particle size analyzer.

The results with this prototype were not promising and our problems evaded us until high speed video revealed flow reversal owing to unanticipated 3 dimensional flows. This prototype was scrapped and a new one was explored.

A second and third prototype was made. The first was a quick test of a concept and offered no adjustments. Given the positive results we continued to develop a third prototype which used inserts and shims to control the critical dimensions. We used high speed photography to follow the trajectories of larger particles (>15micron) which assisted us in design choices.

A fourth prototype was dreamed up by my supervisor. The idea for this one was for a curve to shoot larger particles along a boundary layer then to the reject chamber. This prototype didn't work. We were all sad so we cried and held hands.

Dispensing the powder required a dispenser capable of removing clumps and capable of aerosolizing the particles. To achieve this a three disk assembly with two half openings 180 degrees out of phase on outer portions rotate about a central portion where small slivers of powder collect then get passed to the exit. This system allowed us to control dispensing independent of the pressure drop between the bottom and top of the dispenser tube which in previous trials would cause powder to dump quickly and uncontrollably.

To liberate the dispensed powder and to mix it with air, I designed and machined a venturi with a flattened tube spanning the smallest diameter. This flattened tube was fed with compressed nitrogen which further dispersed the powder.

Labview software was used to automate entire setup and collect useful flow information from the orifice flow meters.

I built a control box to convert low voltage signals from the computer into higher voltage signals...Darlington transistor arrays are the best!