Feeder Design and Testing Update

Following last week's difficulties with the capillary feeder, we returned to the plate feeder this week to try to improve its design.  After talking to fluid dynamics professors, Prof. Donaldson learned that both the geometry of the grooves on the plate and the hydrophilic properties of the plate itself affect how well sugar water moves into the grooves from the jar.  We first tried waxing the surface of the acrylic plate to encourage water to bead up inside the grooves rather than spread across the plate, but this didn't turn out as expected – as usual, water spilled all over the plate when the feeder was inverted (since the jar didn't have an airtight seal against the plate) and pooled on the surface of the plate rather than forming droplets on top of the grooves.  Since modifying the surface of the plate was unsuccessful, we tried changing the geometry of the grooves to get the water flow that we wanted.

In order to do this easily, we decided to design and 3D print a prototype plate.  This both gave us the results that we wanted and allowed us to fine-tune various aspects of the plate.  The grooves in the original acrylic plate, which did not work very well, were each 1 mm wide and 1 mm deep with vertical walls.  In order to encourage water to flow more readily through the grooves, I tested U-shaped, V-shaped, and beveled grooves on the prototype plate.  I also tested the effects of widening and deepening the grooves.  Each groove on the prototype plate had a different combination of shape, width, and depth, and after a few trials with various concentrations of sugar solution, I found that the liquid flowed less readily into wide grooves than into narrow grooves, and V-shaped and beveled grooves worked better than U-shaped grooves.  The V-shaped and beveled grooves that were 1 mm wide and 2 mm deep seemed to work best for both pure water and sugar water, so I made a plate with all beveled grooves of this size to try with the bees.  I also added a low wall around the jar, which greatly improved the seal between the jar and the plate, and an outer wall to hold the flowers in place.  Furthermore, after a couple of dimensional adjustments, the flower now interfaces perfectly with the photogate!

We brought the feeder to the farm yesterday morning to see how well the bees would be able to drink from the new plate.  When the feeder was tilted at an angle, sugar water pooled out onto the plate around the grooves, just as it had before.

However, after we moved the feeder to a more level surface and wiped up the pools of sugar water, we observed a more interesting phenomenon: not only were the grooves full of liquid, as we expected, but the small space between the inner and outer walls around the jar had turned into a canal, forming appealing droplets in the gaps in the outer wall that the bees were eager to drink from.  We observed bees drinking from both the gaps in the outer wall and from the top of the wall (as seen in the photo below), but they also drank from the grooves – which was a good sign.

Here's a bee drinking from our feeder!  The bee's abdomen expands as it fills with liquid, which is really cool to watch.

After returning from the farm, we immediately discussed our next improvements to the feeder.  The grooves work well now, and since the jar sits tightly against the inner wall of the plate, the feeder now leaks much less when flipped over than it used to.  Given the apparent ability of the canal to draw water out of the feeder, we decided to put the openings in the outer wall in front of the grooves to encourage water to flow into the grooves rather than onto other areas of the plate.  We also decided to mold the surface of the plate to the shape of the flower, so that the flower can fit more securely on the plate (in the current version, the flower sits loosely on the plate and can be easily knocked off).

I revised the plate and flower models yesterday, and this morning I'm supervising the 3D printers as they build the models.  This afternoon, I will check the fit between the new plate and the new flower and test the fluid flow properties of the new plate.  We have one week of work left this summer, so stay tuned for the final feeder design by the end of next week!