After completing my engineering prerequisites, the first class I took was ENGS 21: Dartmouth's introductory project engineering class. Working with three others, we interviewed local bee farmers to address difficulties of hive monitoring during the harsh New England winters. We developed a food distribution detection system using load cell configurations, which was eventually implemented on a live farm!
2024
Hardware Lead
Market Research
Product Design
Embedded Systems
The class was very open-ended: recognizing any problem, identify different approaches to a solution and build a functional prototype with good design principles. While other groups quickly launched into an idea, our group was stumped for a bit, disagreeing with each other over our own personal desires. We scheduled a meeting with our professor to talk it out, where we found good advice: focus on consumer research. For what good is a stellar product if no one wants to use it?
We reached out locally and learned from the Dartmouth Organic Farm that honey producers and beekeepers were struggling to keep colonies alive amid increasingly erratic winters. We contacted a regional beekeeper who manages thousands of hives across New England to explain the challenge: hive health is often judged by the distribution of food between stacked boxes. Bees eat from the bottom up, and an irregular pattern (such as upper boxes emptied before lower ones) signals a colony in distress. Monitoring this by manually weighing boxes is time-consuming and risks exposing bees to cold air. He expressed a desire for a beehive monitoring system (especially for hobbyists) for checking winter conditions.
Based on that market research, we set out to design a beehive weight-distribution system capable of tracking each box individually.
We began theorizing approaches and building mock-ups to fit an embedded system into the existing beehive setup. We decided on the approach seen on right, where each stack between boxes would have a separate layer of load cells, with each layer using an amplifier that connected to a microcontroller on the base layer. From this, we needed to machine the layers to the exact dimensions of each bee box to fit on top of each other. We decided to use wood based materials to fit the aesthetic and to attract bees to insulate between the wood layers.
As the person in charge of hardware, I chose to use a Wheatstone bridge configuration of load cells on each layer to measure weight distribution both between boxes and within a single box, since bees often die prematurely when they cannot find food along the perimeter. Using strain gauge load cells also required very little power input as signal amplifiers were used. Brackets for the load cells on the frame were 3D printed after being modeled in SolidWorks as seen below. An Arduino Nano was used for a compact form factor, so each load cell layer could be soldered to a data pin. A LED display was also integrated to display the weight distribution of each box independently.
I wanted the system to work entirely on solar power (lithium ion battery pack with solar cell), so that the device could be implemented into existing setups in the remote field. To maintain net charge, I decided to include a button that the use would press for the system to run for 10 seconds then shut off. The multiple layers of independent Wheatstone load cells and the "sleep" button delay function took a ton of coding (thanks Camry!) and calibration.
After installation, our major concern was insulation. We added siding to the frame and tested the heat loss versus a control in cold rooms. We needed the sidings to fit perfectly around the box frame, while still being able to be compatible with the insulation wraps beekeepers use. Load cell accuracy was measured to be within 0.5% of a precision scale. A 6 Watt solar panel was estimated to be by far enough for continuous deployment over the year.
Our final prototype (with some side panels removed to see internals) can be seen below! You can see I had fun with the laser engraver to make out logo design onto the panels! We were able to implement and test the system on a live farm. Our team managed to complete a first version of the prototype in 10 weeks, with the primary function being weight distribution. With the modular design in place, we also outlined ideas for future additions, such as a varroa mite control system using ultrasound or audio-based detection of bee activity.
This project gave me great insight on how to design a product with real market potential. I know many colleges do these project-based classes for intro engineering students, but I think our group got something special out of this experience. We develop a product that could be sold for others to use (focusing on economics and market research), rather than just something that we wanted to used. Throughout the process, we learned how to navigate team dynamics and set aside personal preferences to pursue what was most feasible and beneficial for the end user. It was amazing to work live with tens of thousands of bees and to learn more about their behaviors! It gave me greater appreciation for the people who work with bees to produce honey (which I love!) 🐝
After completing my engineering prerequisites, the first class I took was ENGS 21: Dartmouth's introductory project engineering class. Working with three others, we interviewed local bee farmers to address difficulties of hive monitoring during the harsh New England winters. We developed a food distribution detection system using load cell configurations, which was eventually implemented on a live farm!
2024
Hardware Lead
Market Research
Product Design
Embedded Systems
The class was very open-ended: recognizing any problem, identify different approaches to a solution and build a functional prototype with good design principles. While other groups quickly launched into an idea, our group was stumped for a bit, disagreeing with each other over our own personal desires. We scheduled a meeting with our professor to talk it out, where we found good advice: focus on consumer research. For what good is a stellar product if no one wants to use it?
We reached out locally and learned from the Dartmouth Organic Farm that honey producers and beekeepers were struggling to keep colonies alive amid increasingly erratic winters. We contacted a regional beekeeper who manages thousands of hives across New England to explain the challenge: hive health is often judged by the distribution of food between stacked boxes. Bees eat from the bottom up, and an irregular pattern (such as upper boxes emptied before lower ones) signals a colony in distress. Monitoring this by manually weighing boxes is time-consuming and risks exposing bees to cold air. He expressed a desire for a beehive monitoring system (especially for hobbyists) for checking winter conditions.
Based on that market research, we set out to design a beehive weight-distribution system capable of tracking each box individually.
We began theorizing approaches and building mock-ups to fit an embedded system into the existing beehive setup. We decided on the approach seen on right, where each stack between boxes would have a separate layer of load cells, with each layer using an amplifier that connected to a microcontroller on the base layer. From this, we needed to machine the layers to the exact dimensions of each bee box to fit on top of each other. We decided to use wood based materials to fit the aesthetic and to attract bees to insulate between the wood layers.
As the person in charge of hardware, I chose to use a Wheatstone bridge configuration of load cells on each layer to measure weight distribution both between boxes and within a single box, since bees often die prematurely when they cannot find food along the perimeter. Using strain gauge load cells also required very little power input as signal amplifiers were used. Brackets for the load cells on the frame were 3D printed after being modeled in SolidWorks as seen below. An Arduino Nano was used for a compact form factor, so each load cell layer could be soldered to a data pin. A LED display was also integrated to display the weight distribution of each box independently.
I wanted the system to work entirely on solar power (lithium ion battery pack with solar cell), so that the device could be implemented into existing setups in the remote field. To maintain net charge, I decided to include a button that the use would press for the system to run for 10 seconds then shut off. The multiple layers of independent Wheatstone load cells and the "sleep" button delay function took a ton of coding (thanks Camry!) and calibration.
After installation, our major concern was insulation. We added siding to the frame and tested the heat loss versus a control in cold rooms. We needed the sidings to fit perfectly around the box frame, while still being able to be compatible with the insulation wraps beekeepers use. Load cell accuracy was measured to be within 0.5% of a precision scale. A 6 Watt solar panel was estimated to be by far enough for continuous deployment over the year.
Our final prototype (with some side panels removed to see internals) can be seen below! You can see I had fun with the laser engraver to make out logo design onto the panels! We were able to implement and test the system on a live farm. Our team managed to complete a first version of the prototype in 10 weeks, with the primary function being weight distribution. With the modular design in place, we also outlined ideas for future additions, such as a varroa mite control system using ultrasound or audio-based detection of bee activity.
This project gave me great insight on how to design a product with real market potential. I know many colleges do these project-based classes for intro engineering students, but I think our group got something special out of this experience. We develop a product that could be sold for others to use (focusing on economics and market research), rather than just something that we wanted to used. Throughout the process, we learned how to navigate team dynamics and set aside personal preferences to pursue what was most feasible and beneficial for the end user. It was amazing to work live with tens of thousands of bees and to learn more about their behaviors! It gave me greater appreciation for the people who work with bees to produce honey (which I love!) 🐝
After completing my engineering prerequisites, the first class I took was ENGS 21: Dartmouth's introductory project engineering class. Working with three others, we interviewed local bee farmers to address difficulties of hive monitoring during the harsh New England winters. We developed a food distribution detection system using load cell configurations, which was eventually implemented on a live farm!
2024
Hardware Lead
Market Research
Product Design
Embedded Systems
The class was very open-ended: recognizing any problem, identify different approaches to a solution and build a functional prototype with good design principles. While other groups quickly launched into an idea, our group was stumped for a bit, disagreeing with each other over our own personal desires. We scheduled a meeting with our professor to talk it out, where we found good advice: focus on consumer research. For what good is a stellar product if no one wants to use it?
We reached out locally and learned from the Dartmouth Organic Farm that honey producers and beekeepers were struggling to keep colonies alive amid increasingly erratic winters. We contacted a regional beekeeper who manages thousands of hives across New England to explain the challenge: hive health is often judged by the distribution of food between stacked boxes. Bees eat from the bottom up, and an irregular pattern (such as upper boxes emptied before lower ones) signals a colony in distress. Monitoring this by manually weighing boxes is time-consuming and risks exposing bees to cold air. He expressed a desire for a beehive monitoring system (especially for hobbyists) for checking winter conditions.
Based on that market research, we set out to design a beehive weight-distribution system capable of tracking each box individually.
We began theorizing approaches and building mock-ups to fit an embedded system into the existing beehive setup. We decided on the approach seen on right, where each stack between boxes would have a separate layer of load cells, with each layer using an amplifier that connected to a microcontroller on the base layer. From this, we needed to machine the layers to the exact dimensions of each bee box to fit on top of each other. We decided to use wood based materials to fit the aesthetic and to attract bees to insulate between the wood layers.
As the person in charge of hardware, I chose to use a Wheatstone bridge configuration of load cells on each layer to measure weight distribution both between boxes and within a single box, since bees often die prematurely when they cannot find food along the perimeter. Using strain gauge load cells also required very little power input as signal amplifiers were used. Brackets for the load cells on the frame were 3D printed after being modeled in SolidWorks as seen below. An Arduino Nano was used for a compact form factor, so each load cell layer could be soldered to a data pin. A LED display was also integrated to display the weight distribution of each box independently.
I wanted the system to work entirely on solar power (lithium ion battery pack with solar cell), so that the device could be implemented into existing setups in the remote field. To maintain net charge, I decided to include a button that the use would press for the system to run for 10 seconds then shut off. The multiple layers of independent Wheatstone load cells and the "sleep" button delay function took a ton of coding (thanks Camry!) and calibration.
After installation, our major concern was insulation. We added siding to the frame and tested the heat loss versus a control in cold rooms. We needed the sidings to fit perfectly around the box frame, while still being able to be compatible with the insulation wraps beekeepers use. Load cell accuracy was measured to be within 0.5% of a precision scale. A 6 Watt solar panel was estimated to be by far enough for continuous deployment over the year.
Our final prototype (with some side panels removed to see internals) can be seen below! You can see I had fun with the laser engraver to make out logo design onto the panels! We were able to implement and test the system on a live farm. Our team managed to complete a first version of the prototype in 10 weeks, with the primary function being weight distribution. With the modular design in place, we also outlined ideas for future additions, such as a varroa mite control system using ultrasound or audio-based detection of bee activity.
This project gave me great insight on how to design a product with real market potential. I know many colleges do these project-based classes for intro engineering students, but I think our group got something special out of this experience. We develop a product that could be sold for others to use (focusing on economics and market research), rather than just something that we wanted to used. Throughout the process, we learned how to navigate team dynamics and set aside personal preferences to pursue what was most feasible and beneficial for the end user. It was amazing to work live with tens of thousands of bees and to learn more about their behaviors! It gave me greater appreciation for the people who work with bees to produce honey (which I love!) 🐝