Portioning Seeds

Unsuccessful prototype produced valuable data.

Year: 2022

Role: Project Lead

Collaborators: (1) Junior Mechanical Engineer

Problem

While scaling our cold-plasma seed treatment process to higher throughput, our process was bottlenecked by the cycle time needed to portion, load and unload each reactor cell. Increasing the size of each cell would necessitate weeks or months of experimental work to validate the change, so I looked for ways of decreasing the time between treatment cycles. 

The most promising (least optimized) element was the weighing and distribution of seed, But measuring weight quickly was challenging, and parallelizing the hardware to do this automatically was beyond our scope and resources.

Strategy

I suggested to our head scientist that the measurement could be done by volume; we ran a design of experiment (DOE) around whether volume dosing was within our mass tolerance - it was. This change was successful and saved significant operator effort- it was eventually applied to our other projects. 

Lesson: it’s easier to measure with a teaspoon than a gram scale, provided you can validate your process tolerance.

The next hurdle was seed distribution - how can a n by n matrix of cells be quickly loaded with doses of seed, and how could this be parallelized, preventing cycle time from scaling linearly with n? After a survey of manufacturing techniques, I decided to emulate pharmaceutical tablet manufacture, where rapid, precise, parallel dosing is done by volume. Seeds would be brushed across a flat surface with holes of a precise volume, then dropped into each cell.

Prototyping

To validate the concept, we made a prototype from 8020 extrusion, on-hand fittings and CNC machined HDPE plates to simulate how seed could be dispensed in this way, and measure, among other metrics: cycle time, dose precision and seed health.

Results

Findings: as long as a minimum amount of seed was maintained in the machine, the dosing was within process tolerance, and could measure and dispense in seconds. Major issues were encountered: the sliding components tended to crush or abrade a portion of the seed, as well as progressive seed damage as seeds were moved back and forth before being dispensed. This decrease in seed health was unacceptable, and the design was ruled out.

Learnings: Seeds are fragile!