Hardware

Feb - Mar 2025 : Project Kickoff

Once the wet lab team finalized their project direction, the hardware team began brainstorming and ideating ways to build a complementary hardware system. We defined our main goal as designing a controlled set up that could test and validate the wet lab’s work under realistic conditions. Early ideas included developing a drone to demonstrate a potential to disperse the protein among agricultural fields and that would be able to detect the efficacy of protein dispersal. After weighing scope, cost, and feasibility, we determined that constructing a vegetation furnace would have the greatest impact within our timeline and resources.

We took inspiration from a similar large-scale coil wire furnace from a Northeastern research lab. Though this lab was interested in using the smoke produced due to sample combustion, the material and setup of their apparatus provided a good starting point to direct our project. Compared to their furnace, we envisioned ours would be scaled down in size, with an adjusted temperature range matching those occurring in forest fires. Additionally, instead of orienting the furnace horizontally, it would be oriented vertically to allow for better ventilation and the ability to capture better footage of the inside of the furnace.

Ultimately, this ideation and planning stage allowed us to clearly define the goals and scope of the project, setting the foundation for the design work that began in April.

Apr 2025 : Design Phase

Throughout April, the hardware team transitioned into the design phase, focusing on turning initial ideas into a workable plan. We sourced critical components from a range of vendors, including McMaster-Carr, Metals Depot (for the aluminum tube), and Amazon (for the PID controller, K type thermocouple, solid state relay, heat sink, furnace cement, and fire bricks).

At the same time, we created sketches to visualize the layout, define dimensions, and compare design options. This process involved anticipating potential failure modes (such as allowing proper ventilation or structural weaknesses) and proactively planning design strategies to mitigate these risks. By the end of the month, we had established both the base material set and the conceptual framework that would guide the furnace’s construction.

May - Aug 2025 : Build Phase

Furnace Shell:

To begin construction of the furnace, we started with the outer shell. This consisted of an aluminum cylinder as the main structure. Aluminum was chosen for its strength and availability, making it a cost-effective choice for housing the inner components. A cylinder shape was also chosen to make the design compatible with wire coiling, which would serve as the primary heating element. To line the inside, we used a ceramic fiber blanket, which is both flexible and highly heat-resistant. This material provided two key benefits: a secondary layer of thermal insulation to maintain stable furnace temperatures, and padding to ensure that the fire bricks installed in the next step would fit securely without cracking under stress. This layered design created a well-instulated, durable foundation for building the inner combustion chamber while keeping safety and performance in mind. This set the stage for the next phase of assembly.

Fire Brick:

To construct the furnace's inner chamber, we used four fire bricks: Two forming the inner radius of the cylinder and two each on the outside that were shaped to fill the outer gaps. Each brick was sawed and filed down to create a cylindrical profile, with the inner chamber refined to a 3-inch radius.

Next, a spiral heating coil pattern was mapped onto the inner cylinder using an X-Acto knife, keeping an estimated 0.5 inch spacing between outlines. Grooves were drilled along these traced paths to create consistent channels for the wire coil. These four fire brick pieces were then joined together with furnace cement, which also filled small gaps and smoothed imperfections on the outer surface into a uniform cylinder.

To finalize the structure, the assembled brick body was cured over an open flame. It was then inserted snugly into the aluminum tube with the ceramic fiber lining. This step established the insulated combustion core of the furnace, ready for wire coiling and further assembly.

Heating Element:

The next stage involved installation of the heating element by coiling Kanthal wire. We stretched approximately 15 feet of Kanthal into a uniform coil and carefully laid into the pre-drilled grooves of the fire brick cylinder, ensuring consistent spacing for even heating. Each end of the wire was twisted onto itself for stability and secured to ceramic terminal blocks, establishing the positive and negative ends needed to complete the circuit. To finalize the design, we fabricated a fire brick lid that enclosed the top of the furnace. This lid included two key features: a ventilation hole to allow controlled airflow and an opening for the thermocouple to be suspended inside the chamber, enabling accurate temperature monitoring. Together, the coiled Kanthal wire and fitted lid transformed the insulated brick body into a functional electric furnace core.

Temperature Control:

Using online resources designed for “DIY” furnaces and kilns, our team identified the Inkbird ITC-106VH PID temperature control system as the best fit for our project. This all-in-one package included the PID controller, solid state relay (SSR), and thermocouple. We chose this system because it could reach up to 1000 °C (1832 °F) while still being compatible with a standard 120 V outlet.

At the heart of the system is the PID controller, the “brain” of the operation. It continuously reads temperature data from the thermocouple (our furnace’s thermometer), compares it to the setpoint, and then issues commands to the SSR. The SSR, which acts as the power regulator, takes the low-power control signal from the PID and toggles the high-power current on and off, ensuring the heating coil maintains the desired temperature.

To make the system safe and fully controllable, we introduced an additional manual switch between the PID and the SSR. This way, even if the PID is powered, the furnace will only heat when we deliberately flip the switch. Once engaged, the SSR allows current to flow to the Kanthal coil, which generates the required heat for the furnace chamber. Additionally, the entire circuit was placed in an electrical circuit box to allow for easy transportation, improved wire organization, and ensure safety of the user.

MATLAB-Based Image Process Script

Like most fire-retardants, FloraGuard is designed to be a liquid-based retardant that can be dispensed through pesticide drones for farms or planes for forests. With this in mind, our hardware team determined that it’d be best to have a device directly attached to the dispensing system. This device would capture images in real time, detect fluorescence, calculate the extent of coverage, and relay back to the user whether the amount dispensed is sufficient to protect against flames.

To achieve this, we developed a script for the device designed to analyze drone or plane-captured images and detect red fluorescence as a marker of cellulose-based flame retardant binding. Unlike manual inspection, which is time-consuming and subject to bias, the script provides an objective and reproducible way to process large image datasets collected during aerial application.

1. File Import
   • The script scans a specified directory for all .png files.
   • Each image is read into MATLAB sequentially for analysis.
2. Red Pixel Detection
   • A red mask is generated by applying thresholds to the red florescence
   • Post-processing includes:
     • imfill to fill holes inside red regions.
     • bwareaopen to remove small noise objects (smaller than 20 pixels).
3. Black Pixel Detection
   • The image is converted to grayscale.
   • Pixels below a low intensity threshold are classified as black.
4. Pixel Counting
   • The number of red and black pixels is counted.
   • Each pixel is assumed to represent an area of 1 m² (scaling factor can be adjusted if actual calibration is known).
5. Coverage Calculation
   • Red area (m²) is divided by black area (m²).
   • Coverage is expressed as a percentage:

$$\text{coverage \%}=\frac{\text{Red Area}}{\text{Black Area}+\text{Red Area}}\times 100$$

6. Classification
   • If coverage ≥ 50% → Good coverage
   • If coverage < 50% → Bad coverage
7. Output
   • For each image, the script prints:
     • Red area (cm³)
     • Black area (cm³)
     • Coverage percentage
     • Classification result

Theoretically, if the percent coverage falls below 50%, the amount dispensed may not provide adequate protection. However, this 50% threshold is currently theoretical. Further testing, including controlled experiments with this script and furnace trials, is required to determine the true minimum coverage necessary for flame protection.

Below is codeblock containing the MATLAB script described above:

%creates file directory to master file with all trials and replicates
HYPER_MASTERFILE = dir(fullfile('\Users\preet\OneDrive\Desktop\CODING ASSIGNMENTS\BMES\Lab_4\iGEM\', "*.avif")); 
for i = 1:length(HYPER_MASTERFILE)
    A = imread(strcat('\Users\preet\OneDrive\Desktop\CODING ASSIGNMENTS\BMES\Lab_4\iGEM\', HYPER_MASTERFILE(i).name));

% Red mask
Rthreshold = 80;
maskRed1 = (A(:,:,1) > Rthreshold) & ...
          (A(:,:,2) < Rthreshold) & ...
          (A(:,:,3) < Rthreshold);

% Black mask (dark regions)
Bthreshold = 80;
maskBlack1 = (A(:,:,1) < Bthreshold) & ...
            (A(:,:,2) < Bthreshold) & ...
            (A(:,:,3) < Bthreshold);
    
    % Cleaning masks
    maskRed2 =imfill(maskRed1,"holes"); %filling any holes in cells made by the threshold
    maskRed3 = bwareaopen(maskRed2,20); %deleting anything too small to be considered a cell
    
    % Determining size of photo
    [rows, columns] = size(maskRed3);
    
    % Counting Red pixels
    [row_Indices, col_Indices] = find(maskRed3); %Identifying the rows and columns of pixels
    R_count = length(row_Indices); %Counting the number of times a red pixel shows up
    
    % Counting black pixels
    [row_Indices1, col_Indices1] = find(maskBlack1); %Identifying the rows and columns of pixels
    B_count = length(row_Indices1); %Counting the number of times a black pixel shows up
    
    % Area of Red
    pixelArea = 0.35 * 0.35; %(mm^2)
    Red = pixelArea * R_count; %(mm^2)
    
    % Area of Black
    Black = pixelArea * B_count; %(mm^2)
    
    % Calculating coverage
    percentCoverage = (Red/Black) * 100;
    
    % Results
    fprintf('Image %d: Red Area = %d cm^2, Black Area = %d cm^2, Coverage = %.2f%%\n', ...
        i, Red, Black, percentCoverage);
    if percentCoverage >= 50
        disp('Good coverage');
    else
        disp('Bad coverage');
    end
end
              

Figure 12. Fluorescence Microscopy of bacteria expressing RFP-CBD

Figure 13. Output from MATLAB-Based Image Process Script

Petri Dish Mold

A side project that the Hardware Team engaged with at the start of May was printing a mold for the Wet Lab to place in the petri dish. A solution of media (LB agar and kanamycin, an antibiotic) and IPTG was poured into the negative space of the mold and the whole plate was then streaked with BL21DE3 bacteria, a specific strain of E.coli. When the bacteria interacts with the IPTG, it fluoresces and will fluoresce in the shape of the Northeastern iGEM logo, providing both a satisfying visual effect and a tool for the Wet Lab Team to use during their experiments. This logo was created on Onshape and FDM printed, allowing it to be cleaned and sterilized before use in a wet lab environment.

Figures 14-16. 3D-printing process of the mold.