● ● ● fatima_ramirez / portfolio.2026 v.01
Project 02 · Trihton

EDAPI
bee venom collector

A controlled-stimulation system that gets bees to sting glass — safely.

01 / Hero
EDAPI COLLECTOR IN THE FIELD /assets/images/edapi-hero.jpg
year 2019 – 2020
role Lead electronics & firmware · 2-person design team
team 2 designers (electronics/firmware + mechanical), in collaboration
stack ESP32 · H-bridge driver · custom PCB (EAGLE) · C/C++
status Deployed · 2 systems in active use · repeat order 2024
01 · Context

Why this matters

A beekeeper learned from his daughters — biology students — that bee venom (apitoxin) is being studied as a treatment for rheumatoid arthritis. He wanted to expand his apiary business into apitoxin harvesting, but couldn't find affordable collection equipment. Commercial alternatives were scarce, expensive, and not designed for field conditions.

He approached us through a personal connection with one of Trihton's co-founders. The challenge was three-fold:

02 · Approach

How I solved it

The first iteration tried self-contained collectors — each with its own signal generator. It worked but was expensive and complex. We pivoted to a centralized architecture: one "brain" unit generates the stimulation signals and manages power, while up to 20 slave collectors connect to it in parallel via RCA cables. The slaves carry no electronics — they're rugged, replaceable, and low-cost.

Multi-domain ownership

Electronics Custom PCB integrating H-bridge signal generation, power management, ESP32 controller, and an output selector for DC vs. square-wave bipolar output.
Firmware C/C++ on ESP32 — signal pattern generation with configurable frequency, amplitude, and duty cycle; activation cycles tuned to research-backed safety limits.
Field-ready power Dual power input (internal 12V lead-acid battery + external automotive battery via alligator clips); ~7 h autonomy at full load (20 collectors); rugged RCA + BNC connectors.
Signal research Reviewed existing devices (some at up to 33V); ran field tests to find the safe envelope. Settled on 9–12V output, ≤10 min activation cycles — minimizing bee mortality while maintaining stimulation.
BRAIN + SLAVE ARCHITECTURE
fig.01 — one brain, up to 20 parallel collectors elec · firmware

Key technical insight

Existing devices used DC signals. In our field tests, we discovered that square-wave bipolar signals (alternating polarity) provoked stronger, more localized stinging — bees seemed disoriented by the alternating field and stung more actively at the collection surface. This became the dominant operating mode of the final product.

Iteration discipline

Three iterations, each driven by direct field testing with the beekeeper:

Every iteration involved on-site testing alongside the client, whose decades of beekeeping experience interpreted the bees' behavior in ways no datasheet could.

03 · Result

What it achieved

Delivered & in use
2 complete systems
Each with 1 brain + 15–20 slave collectors. In active use since 2020.
Strongest validation
Repeat order in 2024
The client came back for another full system — the clearest signal the product works.

Final specs

FIELD · MP4

Video — EDAPI collecting venom in the field

What I took from it

EDAPI taught me that engineering decisions in field-deployed hardware are not just technical — they're collaborative. The single most valuable design choice (bipolar signaling) came from observing real bees in the field, not from a datasheet. Working directly with the end user changed how I think about specification gathering for every project since.

04 · Gallery

Process & details

Next project →
CAHD
Unbreakable IoT biometric attendance