Junior Design · Omega Design & Norwalt
Automated Pill Bottle Capping Line
A compact, fully automated back station that takes filled pill bottles, orients and places caps, tightens them past the child-safety lock, and rejects anything capped incorrectly — built by an 11-person team for two automated-packaging sponsors. I led the part-rejection subsystem.
- My role
- Part-rejection lead
- Team
- 11 engineers
- Timeline
- Aug 2024 – May 2025
- Tools
- SolidWorks, Arduino Mega, 3D printing, laser cutting
The problem
Large pharmaceutical producers can afford big automated bottling lines. Small producers can't, so they fill and cap bottles by hand. Omega Design and Norwalt asked us to close that gap with a small, low-cost machine that caps bottles at least twice as fast as a person, runs without an operator, and fits inside a 26″ × 20″ × 40″ footprint on a $400 budget.
Our team built the back station: it receives bottles already filled by a partner team's front station, then has to (1) descramble loose caps into the right orientation, (2) place and tighten each cap until the safety lock engages, and (3) catch and remove every bottle that isn't capped properly.
How it works
-
Conveyance
A guide-railed conveyor lines bottles up single file and hands them to a laser-cut rotary table. Pockets sized to the bottle hold each one steady while the table indexes it from station to station.
-
Cap descrambling
A stepper-driven linkage oscillates a platform inside an acrylic hopper. The caps' off-center weight settles them the right way up, and a laser-cut chute drops them onto the bottles.
-
Capping
A lead screw lowers a rubber-lined 3D-printed "twister" onto the cap, then a second motor spins it until the child-safety lock engages. Side rails stop the bottle from turning.
-
Inspection & rejection My subsystem
Three whisker sensors check bottle presence, cap presence and cap height. A bad bottle gets swept off the table into a reject bin by a servo arm. Good bottles drop through a cutout into the collection bucket.
My subsystem: part rejection
Components
Pick a component here, or click it in the model, to isolate it. Drag to rotate, scroll to zoom.
I owned inspection and rejection end to end: the mechanism, the sensors, the mounts and the control code.
- Sensing. Three whisker sensors: one confirms a bottle is in the pocket, one that a cap is there, and one checks cap height. A loose or missing cap sits high and trips the height whisker.
- Actuation. A servo swings the 3D-printed rejector arm down from above and sweeps the bad bottle out of its pocket and into the reject bin.
- Control. I programmed the rejection logic on an Arduino Mega, syncing the three sensors with the rotary table's indexing.
- Result. It correctly sorted 11 of 12 bottles on average and passed every rejection benchmark in final testing.
How the design evolved
- Concept Break-beam sensors plus a solenoid that pushes bad bottles off a conveyor.
- Fall prototype The belt was wider than expected, so I swapped the solenoid for a servo with a long custom arm. It swings out 90°, waits for the bottle to reach it, then sweeps back 120° to push it into a bin.
- Final The conveyor was replaced by a pocketed rotary table. Whisker sensors replaced break beams, and the arm now reaches down from above. That meant fewer parts and more accurate detection.
Testing & results
We started by timing a person filling, capping and tightening 12 bottles by hand: 28 seconds per bottle on average. Then we ran 12 bottles through the machine and tested each subsystem against the sponsor's requirements.
| Requirement | Target | Measured | Result |
|---|---|---|---|
| Speed | At least 2× faster than by hand | 15 s per bottle vs. 28 s by hand | Pass |
| Part rejection | 11 / 12 bottles sorted correctly | 11 / 12 rejected · 11 / 12 accepted (average) | Pass |
| Bottle handoff | 12 / 12 bottles collected | 12 / 12 | Pass |
| Cap tightening | 12 / 12 caps tightened | 12 / 12 | Pass |
| Footprint | 26″ × 20″ × 40″ | 26″ × 20″ × 40″ | Pass |
| Cost | 20% under $400 | $341 | Pass |
| Pill spillage | ±1 pill | 0 pills | Pass |
| Cap descrambling | 12 / 12 caps oriented | 4 / 12 average | Fail |
The weakest link was getting caps from the descrambler onto the bottles. Once a cap was seated, the capper and rejector worked reliably. The report's path-forward section covers the descrambler and cap-placement redesign.
Final design review
The slides from our final design review for Omega Design and Norwalt cover benchmarking, concept selection for each station, the final design, the bill of materials, failure analysis (including a fatigue check on the descrambler drive shaft), and test validation.