ProDose

Design lab

UMass MIE capstone · pill dispensing module

Six ways to dispense one pill

ProDose has to dispense exactly one pill at a time from any bottle. Our first design, a sorting wheel, had known weaknesses, so we simulated it alongside five other approaches: a vacuum arm, a sliding pocket, a gravity lane, sensor-gated belts and a vacuum disc. Each one runs in a rigid-body simulation with real masses, friction, motor limits and suction force. We test them against the questions in the project brief: how often is it exactly one pill, how do we know, and does the pill survive?

The six concepts

Each is a working machine in the simulator with adjustable parameters, live sensor traces and the same test protocols. Colour identifies the concept on every chart on this site.

What the project brief asked, and where it is answered

Engineering areaWhat we didWhere
1 · Bottle / container interfaceMonte-Carlo of an adjustable iris with a compliant silicone seal against neck tolerance and out-of-round, versus a rigid clamp and a fixed sleeve; seal, hold force and neck crush limits.Labs → Bottle interface
2 · Single-pill meteringAdjustable pocket (length, depth, width) with wiper and compliant lips; five other singulation principles for comparison; sweeps of every adjustable setting; realistic size range from 25 pill sizes.Bench §1–4, Simulator
3 · Dispensing confirmationSix methods (edge count, pulse width, light curtain, dual beam, load cell, fused) against overlapping pills; confusion matrices and a live drop test.Sensing lab
4 · Transport and collectionFree-fall damage margin, ramp angle and lining, tray wall height and bounce-out with 30 pills per configuration.Labs → Chute & collection
5 · Cleaning and residueFiltered low-pressure air purge modelled from wall shear against particle adhesion; carry-over into the next product.Labs → Cleaning
6 · MaterialsStainless, polymers and elastomers ranked against requirements; a part-by-part recommendation.Labs → Materials
7 · Testing metricsSuccess, misses, multiples, jams, repeatability, damage, compatibility range, adjustment effectiveness, interface and confirmation reliability, each with an acceptance bar.Design → Test procedures
DeliverablesDesign alternatives and rationale, failure-mode table, decision matrix with rank stability, recommendations and next steps.Design

Why trust the simulation?

Real units

Millimetres, milliseconds and grams, so every force comes out in newtons. Pill masses come from size and density, crush strengths from size, drop ratings from real pill classes.

Validated engine

Free fall, projectile, incline sliding, restitution, momentum, actuator force limits and suction force are checked against closed-form answers, live in your browser on the physics check page. One known limitation is stated, not hidden.

Jams are real

Motors have force limits. A pill caught in a mechanism stalls the motor at its rated force; that stall is the jam signal and the squeeze the pill sees. Suction is a force limited by pressure × area, so a snagged pill drops.

Repeatable

Every result comes from seeded runs and the raw trial records ship with the site. Re-run any protocol in the browser on the bench page, or regenerate everything with node bench/run.js.