Hardware Design for Firefighting Drone 

Last Updated: April 28, 2020 
Authors: Jake Berkowitz, Will Durbin, Robert Lane 
 
 
About This Doc 1 

Project Background 2 

Progress 5 
Need Finding (Fire Department Call) 5 
Design Criteria 5 
Design Iterations (reasoning for certain decisions) 6 
Alpha Prototype Drawings 10 
Alpha Prototype Bill of Materials 11 
Market Research 12 

Future Work 13 

What We Are Changing and Why 13 
Beta Prototype Drawings 15 
Beta Prototype Bill of Materials 16 
Scaling to Manufacturable Product 16 
Plastic Motor Mounts and Connectors 17 
Wing, Tail Wing, and Electronics Platform 17 
Carbon Fiber Tubes 17 
Motors 18 
Electronics 18 
Propellers 18 

Appendix 19 

Resources 20 

About This Doc 

Executive Summary: 
The purpose of this document is to identify a problem, propose a solution, and showcase 
the efforts completed by the team. In the United States alone, wildfires typically account for about 
3000 deaths and $20 billion in economic losses [4]. The ability to gather real time information and 
allocate a finite number of resources are two of the biggest challenges facing firefighters when 

1 
combating wildfires. Given the necessity of a rapidly deployable visualization medium, many fire 
departments are looking into using advanced unmanned aerial vehicle technology in firefighting. 
Fire departments around the world are starting to use consumer drones such as the DJI Matrice 
and DJI Inspire, however being consumer drones these solutions have limitations in the harsh 
environment of forest fires. In large scale fires, fire departments will employ the use of very costly 
visualization methods such as planes with specialty observation systems or military MQ1 
Predator drones. The goal of this project is to make a cheaper more efficient drone specifically 
tailored toward performing in the harsh conditions of a forest fire. Through developing this 
specifically tailored drone, fire departments would not have to rely on consumer drones that have 
inadequate flight performance, or need to use costly methods such as military surveillance. 
This report will detail the efforts of the group throughout the semester in doing 
background research on the environment drones must fly in, as well as defining design criteria for 
development. The group was able to do so through conducting online research, as well as speaking 
with a California fire department chief. After defining these needs and design goals, the report 
then details the initial solution (alpha prototype) built by the group, detailing mechanical and 
electrical design. A preliminary theoretical analysis of this design is included as well. Given the 
limited time in the semester, the group was not able to complete a second design iteration but was 
able to plan out the next design. These future design plans (beta prototype) are elaborated on as 
well. Limitations of these designs, as well as ideas on how to scale up this design to a 
mass-produced product after further development is also discussed. 
 
Important terminology: 
Alpha Prototype: The current build of the drone. No fixed-wing capabilities, essentially a 
large quadcopter. 
Beta Prototype: The planned build of the drone by the end of the 2020 Spring semester. A 
fixed wing aircraft with VTOL and hovering capability. 
Minimum Viable Product (MVP): The theoretical version of the Beta Prototype that would 
be marketable. A fully functioning quad-plane made from custom parts with material 
properties able to withstand use in a firefighting capacity. A product that a fire department 
could actually use. 

Project Background 
The Problem: 
One of the most difficult aspects of fighting wildfires is gathering accurate real time information. 
Where is the fire burning, where is it spreading, what direction is it heading, how intense is it 
burning, are there multiple fires burning, are there hotspots, and is the fire controlled are all 
questions firefighters face every day. Quickly changing weather conditions and enormous areas of 
land to cover make answering these questions very problematic. The proposed solution is an 
unmanned fixed-wing aircraft with Vertical Take Off and Landing (VTOL) that is capable of being 
piloted by firefighters or flown autonomously.  
  

2 
2019 marked a year when fires dominated the news headlines for the entire world. Ranging from 
California, Brazil, and Australia, wildfires devastated enormous amounts of land, animals, and 
property. They also destroyed lives and cost governments millions to fight and to recover from the 
damages. With the climate changing and severe weather patterns becoming more normal studies 
show that wildfires will be an even bigger problem in the future. 
  
The traditional methods of detecting and monitoring wildfires relies on manned helicopters and 
fixed wing planes. These types of aircraft are able to complete the mission but have major 
limitations. First, weather conditions and visibility are notoriously terrible in the airspace above 
wildfires which means manned aircraft are constantly grounded and not allowed to fly. Fighting 
wildfires is already a dangerous job and having aircraft fly in unsafe conditions is putting more 
people at risk. Second, planes and helicopters are extremely expensive to operate. According to 
the LA Times, most departments do not have their own fleet or aircraft and are therefore required 
to pay thousands of dollars per hour to use the helicopters or planes. Even if there are funds 
available, there is often a shortage of planes and helicopters and fire commanders aren’t able to 
scan all the land they want. Third, wildfires are often in remote locations hours away from the 
nearest airports which means more time is spent flying to the fire scene rather than at the scene. 
  
Other methods of detecting and monitoring wildfires include satellites and boots on the ground. 
Satellites are great at offering trends and patterns over long periods of time but are not being used 
for real-time monitoring. Boots on the ground is as real-time as it gets but is obviously limited in 
the amount of area that can be monitored at once. 
  
Consumer drone manufacturers like DJI have converted popular photography drones into 
firefighting drones by adding thermal imaging sensors. The birds-eye view and thermal imaging 
have helped firefighters manage urban and small-scale fire scenes greatly. They are relatively 
cheap and not difficult to fly. These drones, however, have not been as successful fighting wildfires. 
The main problem boils down to flight time. On the best day with no wind, these types of drones 
can fly for only about 30 minutes. Realistically, flight times are usually much less. This limits the 
amount of ground the drone can cover which is why they aren’t being implemented to fight 
wildfires. These drones are also not able to fly in the weather conditions that usually come along 
with wildfires. Winds typically reach 40+ mph above the fires but the traditional drones are limited 
to flying in winds under 20 mph. 
  
To better understand these problems, we talked to two Californian fire departments. The first 
department located in Silicon Valley is one of the most technologically advanced departments in 
the US. They are already utilizing drones and have found them to be useful but not perfect. They 
are mostly satisfied with their current drones for managing house fires and industrial fires. 
However, the department does not currently have anything for wildfires and sees this space as an 
area for new technology. Compared to their current drones, the department wants better flight 
times, more durable design, and something that does not require a pilot. The second department in 
Stinson Beach has no drones in operation and indicated that in addition to wildfires, they often are 
called upon for medical aids, trail rescues, cliff rescues and water rescues. The department 

3 
frequently uses rescue helicopters for these calls but noted that due to weather, it is commonly 
unsafe for the helicopters to fly. The department mentioned that in these situations, a drone 
would be very helpful for large area search and rescue missions. 
 
The Solution: 
In order to solve the problems discussed above, our idea is to create an unmanned aerial vehicle 
(UAV) that combines the functionality and capabilities of a manned helicopter/plane with the cost 
effectiveness and ease of use of a consumer photography drone. In order to solve the flight time 
and range problem facing the current drone solution, our design features a quad-plane flight 
configuration. This type of configuration includes vertical propulsion to provide vertical thrust, a 
fixed wing to provide lift during forward flight, and horizontal propulsion to enable forward flight 
speed. This setup allows for vertical takeoff and landing, stable hover, and forward fixed wing 
flight. Fixed wing flight is much more efficient at both covering ground and battery consumption. 
Using this technology our UAV will have a flight time measured in hours rather than minutes. This 
is comparable to manned helicopters and planes. The small form factor, ease of use, and cheaper 
cost, however, will make our UAV more like a consumer drone in terms of operating and 
implementing. 
 
Value: 
The value of our solution is giving firefighters and agencies a game changing tool to help manage 
and control wildfires and conduct search and rescue operations. Our device will be unmanned. 
This is important because even when helicopters and planes are available, they are often grounded 
due to unsafe weather conditions. When a wildfire is just starting, gathering information quickly 
can prevent it from turning into a much bigger disaster. If the weather is unsafe for manned 
vehicles to fly in these conditions, then there is not much that can be done to prevent the spread of 
the fire. An unmanned vehicle like our solution would be able to fly in more adverse weather 
conditions because there is no risk of losing life and the equipment is much cheaper. Deploying our 
solution at the beginning of a wildfire when the weather is bad could prevent the fire from 
spreading and becoming something worse. If a hiker goes missing and the weather is too unsafe to 
fly a manned aircraft, using an unmanned drone could literally be the difference between life and 
death. 
  
Furthermore, most departments and agencies in the US and especially around other parts in the 
world do not have the funding or resources to keep and maintain a fleet of helicopters or planes 
that will cost millions. The alternative is to lease or hire aircraft that will typically cost thousands 
per hour. Neither of these options are cost effective which means many departments and agencies 
across the world have little or no aerial technology to combat wildfires and help with search and 
rescue missions. Our UAV will provide nearly the same capabilities at a fraction of the cost. This 
will not only help the taxpayers save money in countries like the US but will also provide aerial 
operations for a whole new segment of departments and agencies across the world that never had 
these types of capabilities. 

4 
Progress 
This section contains the current progress of the project. Last updated: April 25, 2020. 

Need Finding (Fire Department Call) 

To aid in deciding on the appropriate goals for the drone performance, the group had a call with 
Jon Johnston, the Fire Marshal/Division Chief of the Menlo Park CA Fire Department to get 
visibility on important features of the drone. Jon was an invaluable resource because his fire 
division was one of the first groups to integrate drones with fire fighting. The division has been 
using drones to fight fires for 3 years, and has spent a long time refining the use cases of their 
drones. The drones are used for search and rescue, insurance claims, history keeping, and (most 
relevant) wildfire visualization and studying of burn patterns. The station hopes to deploy drones 
on a daily basis for inspections and fire responses, however the biggest problem the group has run 
into is getting trained pilots and trained observers for these drone flights, given a designated pilot 
is needed to fly even autonomous drones. While the group cannot fix this problem given FAA 
regulations, they can provide support making the most effective drones possible tailored to 
firefighting. An important thing the group learned is that fire departments mainly use the DJI 
Mavic and Matrice 210, which are then fitted with thermal cameras. Given these drones are both 
consumer drones that are not primarily meant to fight forest fires, the group is presented with an 
interesting opportunity to make a drone specifically tailored to fight fires. 
 
Jon elaborated that the most important factors for a specifically fire-fighting drone, explaining 
that durability was most important--the drone would need to be able to handle diverse weather 
conditions, such as 60 mph winds, extreme heat, and flying debris such as ash and fire embers. 
Next important, Jon explained that battery longevity is fundamental, optimally having a flight time 
of 2 hours. It would also be important that the drone could transmit live video data from a distance 
of 5+ miles, as well as be easy to use (given firefighters do not specialize in drone operations). 
When asked about what sensors would be important to use in the drone, Jon explained that the 
more the merrier, though it would be especially important to include an HD camera and thermal 
camera.  

Design Criteria 
In light of the discussion with Jon, as well as after doing background research into other 
firefighting drones and having several discussions, the group started to develop design criteria to 
provide guiding performance metrics for future development. Table 1 contains a list of optimal 
performance characteristics of this drone: 
 
 
 

5 
 Design Metric  Optimal Value 

Water/Dust/Debris Resistance  IP54 

Max. Heat Resistance  1000 °F 

Flight Time   2 hrs 

Video Transmission Distance  5 mi 

Max. Gust Resistance  65 mph 

Takeoff Method  VTOL 

Assembly Time  5 min 

Table 1: Design criteria from phone call with Menlo Park Fire Department. 
 
The group decided to move forward with these specific design criteria to holistically provide a 
robust and effective drones that could fly in forests and handle the conditions of forest fires. Jon 
recommended the drone be able to handle 1000 °F heat and 65 mph winds, hence the group 
decided these would be reasonable environmental conditions to aim for. Jon also mentioned that 
having the ability to fly the drone 5 miles away and fly for 2 hours would also be optimal, hence 
that design criteria was chosen. A VTOL drone would be optimal because this capability would 
enable the drone to take off in a confined space with little lateral space needed (especially in dense 
forests), as well as hover in place while visualizing part of a fire. The group also wanted to ensure 
that the drone could be quickly and easily assembled in a manner similar to competitors’ drones. 
Most competitor drones can be assembled for flight in several minutes, hence the group decided 
that a 5 minute assembly time would be reasonable. An IP (Ingress Protection) rating of 54 was 
chosen to ensure reasonable environmental protection of the drone. The 5 in this rating is for 
intrusion protection. The group assumed that it would be impossible to have complete protection 
of the drone’s electronics from dust given the open stator and rotor configuration of the brushless 
motors, hence a partial protection rating was chosen. With respect to the second digit of the 
rating, the group wanted the drone to be able to handle rain--a rating of 4 in this category reflects 
protection from water splashes in all directions, hence it was chosen. 

Design Iterations (reasoning for certain decisions) 

Our team decided to use a rectangular airframe early in the design phase, as it combined 
simplicity, constructability and versatility, all while using the prefabricated carbon fiber tubes that 
we had access to. Because our team had more experience building and flying quadcopters than 
fixed wing aircraft, we decided to build a fully functioning quadcopter base, onto which we could 
attach pieces to allow for fixed wing flight. Accordingly, the main aesthetic of our drone has looked 
very similar throughout the beginning iterations. In the early stages, our team mostly iterated 
design of specific pieces, such as joints and mounts. Most of the larger-scale design iteration was 

6 
to come in the addition of fixed wing flight functionality. Indeed, just before spring break, our team 
cycled through different design ideas for attaching a horizontally oriented motor.  
 
3D Printed Pieces: 
The tee joints have three design iterations. Initially, we printed both a full joint, and one cut in half 
in the plane of the body tubes. Because of its print quality, ease of construction and 
deconstruction, and ability to be far more adaptable, the tee joints in two-half pieces proved to be 
a significantly better option. The final iteration of the tee joints involved simply changing print 
qualities to reduce weight by nearly 50%.  
 
The motor mounts also went through an iterative design process. Initially, given the team’s design 
iterations led to changes in drone weight (thus motors needed), as well as different motors being 
found, the motor mounts were designed to be compatible with different motors. The mounts had 
the same base component that would attach to the carbon fiber tubes; these base mounts would 
never be taken off. An adapter component (tailored to the specific hole pattern of the motor) 
would then be fastened on top. Two different adapters were made for current motors that were 
being considered. The adapter was designed to be attached to the base component through nut 
and bolt fastening--the holes for this fastening were designed to be kept out of the way of the 
mounting hole pattern. This primary iteration can be seen below: 

 
Figure 1: Primary Motor Mount Iteration 
 
After printing this preliminary design, the group integrated the motor mounts with the current 
design iteration of the drone frame (Alpha Prototype). While the motor mounts fit relatively well 
within the design, as well as having well-fitting motor adapters, it was hard to attach the motor 
adapters to the motor bases. It was difficult to attach the two printed components because the 
bolt and nut were difficult to attach--the nut was quite hard to turn while keeping the bolt in 

7 
pressure between the two parts. During this time, the group was also finalizing a motor choice. 
Because motor changes were less common, the group decided to re-design the motor mounts to 
hold only one motor type and have no adapter. Therefore, a new iteration was made that had just 
one specific motor mounting hole pattern, as can be seen below: 

 
Figure 2: Secondary Motor Mount Iteration 
 
After attaching these motor mounts, the group found another problem with the mount design. 
There were recesses in the design for an allen wrench to screw the bolts into the base of the 
motors, however there was not enough clearance for the allen wrench to screw in the bolts. The 
group had to effectively drill the recesses further down in order to fit the allen wrenches in. In light 
of this problem, the group made one final iteration of the design with more space for the allen 
wrench to fit in, as can be seen below: 

8 
 Figure 3: Final Motor Mount Iteration 
 
 
Wing & Horizontal Motor Placement: 
Finally, just before spring break, our team had begun to experiment with wing fabrication and 
horizontal motor placement to begin transforming the quadcopter into a VTOL fixed wing aircraft. 
Our first wing was simply a loose airfoil shape created out of foamcore and taped to the body to 
get a general feel for flight characteristics. We quickly realized that this wing was far too heavy 
and created too much drag due to partially squared edges. We were trained to use the hot wire 
cutter in The Foundry, and we did research to determine the airfoil shape that would be most ideal 
for our usage. We wanted to maximize the flight time, or endurance, of the aircraft. To do this, we 
wanted to make the aspect ratio of the wing as large as possible; making the aspect ratio large 
means making the wing very long and slender. However, to balance usability, we couldn’t make our 
wing too long, so we compared the drag polars between several airfoils on airfoiltools.com. To get 
a starting point, we researched several reasonably slender wings (thickness<12%) at low-speed 
flight for different types of aircraft [1]. We knew our aircraft would be flying between 20-60 mph, 
and assuming a chord length of 10 inches, this provided a range of Reynold’s numbers 
≈150,000-500,000. The drag polars of three sailplane airfoils, a flying wing airfoil, and a free flight 
airfoil were then compared in the appropriate range of Reynold’s numbers [2]. Our team decided 
initially to use the S7012 airfoil, as it was very thin (8.75% thickness) and had a very high lift/drag 
ratio, a quantity that should be maximized to maximize endurance. However, we quickly realized 
that this wing was too thin; when we cut one from styrofoam, it was extremely floppy and 
structurally fragile. We then found the S7055 airfoil, which had even better flight characteristics, 
10.5% thickness, and a flat bottom that would make its attachment to the body much easier. This 
was the next airfoil shape that we were going to try. 
 
The other structural aspect of our drone that needed to be modified to allow for fixed wing flight 
was the addition of a fifth motor, placed horizontally in either a pusher or puller configuration. 
Initially, a motor mount was created and placed directly in the center of the shorter cross body 
tubes. However, we quickly realized that this placement was far from ideal, as the propeller would 
interfere with the VTOL propellers, as well as wash a significant amount of air over the body tubes 
before hitting the wing. Instead, a new mount was created to be placed in the same location, but 
hold a fifth carbon fiber tube that would project out in front of the two leading VTOL propellers 
and hold a puller motor at the very front of the aircraft (see Figure 5 for reference). 

9 
Alpha Prototype Drawings 

 
Figure 4: Alpha Prototype final drawings. 

10 
Alpha Prototype Bill of Materials 
Structural Bill of Materials 
Part  Quantity  Note/Material 

Body Tube ID Ø1” –– 32.5”  2  Prefab carbon fiber 

Body Tube ID Ø1” –– 17”  2  Prefab carbon fiber 

Motor Mount (Detail A)  4  3D Printed (PLA) 

Tee Joint (Detail B)  4  3D Printed (PLA) 

Electronics Board 17.6” x 12”  1  Foamcore 

Motor  4   

Propellor 16 x 5.4L  4  Carbon fiber 

Motor Screw  16   

Body Screw ¼” x 20 – 2.5”  12  Stainless Steel 

Body Nut ¼” x 20  12  Stainless Steel 

 
Electrical Bill of Materials 
Part  Quantity  Note/Material 

5200 mAh, 4S, 10C Battery  1   

Electronic Speed Controller  4   

GPS  1   

GPS Mount  1   

Pixhawk  1   

5V Converter?  1   

Transmitter?  1   

Buzzer  1   

Killswitch  1   

Connective Wiring    Various gauges and lengths 

Tables 2 & 3: Structural & electrical bill of materials for Alpha Prototype. 

11 
Market Research  
 
Consumer drones (quadcopters) first showed up in the early 2010s. They were initially used for 
recreation but as they became more advanced they began to be used for photography and 
videography. Once drones were established as camera carrying devices, they began to be used for 
industry. This mainly included helping with inspections for the oil and gas, agriculture, and 
construction industries. The first professional thermal drones were created around 2015. It wasn’t 
until around 2017 that they began to get popular and actually utilized. Today, DJI dominates the 
consumer and professional quadcopter market. They have a handful of thermal capable 
firefighting drones ranging in price from $3000 to $20,000. These drones are typically used for 
structural fire surveillance. Table 4 summarizes some of these drones. 
 
Consumer Drone Companies With Thermal Capabilities 

Company  Capabilities  Price Range  Country 

DJI  Excellent quadcopter systems with best thermal  $3000-$20,000  China 

imagery provided by FLIR.  

Yuneec  Excellent quadcopter systems with decent  $8000  China 

proprietary thermal imagery  

Parrot  Small and portable quadcopter. Not for professional  $2000  France 
use. Mediocre thermal imagery provided by FLIR 
Table 4: List of Consumer Drone Companies with Thermal Imaging.  
 
DJI, Yuneec, and Parrot all offer thermal capable quadcopters at a variety of price points and 
capabilities. These quadcopters are great except they all have flight times under 25 minutes. When 
targeting wildfires, the amount of land to cover is usually so vast that flight times in hours are a 
must. In order to be competitive, our solution specifically targets fighting wildfires and large scale 
search and rescue. With current battery technology, longer flight times are usually accomplished 
using a fixed wing design. Table 5 shows some smaller companies that are either targeting the 
firefighting/disaster relief drone market or are utilizing a fixed wing design. 
 
 
 
 
 
 
 
 

12 
 Specialized Companies Leveraging Thermal Imaging or Fixed Wing Design 

Company  Capabilities  Price  Country 

Range 

ElistAir  Tethered quadcopter providing unlimited flight time    France 

for aerial reconnaissance for stationary targets. 
Thermal imaging capabilities.   

Applied  Simple, affordable, fixed wing UAV with modular  $20,000  USA 
Aeronautics  payloads and capabilities. Offers thermal imaging by 
FLIR. Possible VTOL design. 

ALTI  VTOL Fixed wing design with up to 20 hours of  $50,000  South 
endurance. Many applications including search and  Africa 
rescue and thermal capabilities. 

ZeroTech  Hybrid powered VTOL fixed wing UAV. Used for many    China 
applications including forest fire monitoring. 
Table 5: Specialized Long Endurance UAV Companies. 

Future Work 
What We Are Changing and Why 
An important component to study more in depth is brushless motor performance. While brushless 
motors are excellent in their thrust to weight ratio and efficiency, they are not able to operate at 
high temperatures. This problem is due to properties of magnets. These motors are operated 
through changing the charge of an electromagnet that through changing charge moves magnets 
on the outside of the motor (the rotor), making the rotor rotate. Magnets in general start to lose 
their magnetism as their temperatures increase, and completely lose their magnetism at a point 
called the Curie temperature. Given that these motors are operated by magnetic attraction, the 
efficiency and operation of these motors could be compromised by temperatures as high as 
needed in the design criteria (1000 °F). Neodymium magnets (the most effective magnets for use 
in these motors for their high magnetism levels) start to lose their magnetism at around 176 °F, 
and the Curie temperature of the best temperature-performing neodymium is around 608 °F. 
Given the needed operational temperature of the drone, the group may need to explore other 
motor options that are not magnetic. However, the group could pursue finding coolant or 
protecting materials for these magnets to keep their magnetism, or could find the amount of 
convective heat transfer cooling these magnets with the rotational motor of the rotor. A 
preliminary convective heat transfer calculation was conducted to conceptualize this problem. 
Using properties of air at 1000 °F (and calculating the velocity of air over the magnets when the 
motor was at maximum speed) the Reynolds and Prandtl numbers were calculated, being 
1511.209 and 3.697, respectively. Because the Reynolds number was below 2300, the flow would 

13 
be assumed to be laminar. The following Nusselt number correlation was used to find the 
convective heat transfer coefficient, which is models heat transfer for forced convection over flat 
plates: 
N uL = 0.664(ReL )1/2 (P r)1/3   = 39.913 = hL
k  
f

Through completing these calculations, the convective heat transfer coefficient was 571.75 W
m2 K . 
This is a relatively high heat transfer coefficient, which means that there would be a relatively 
small temperature difference between the air and magnet despite the convection. In addition, 
according to 0-D Conduction, if the magnets are in this hot air for long enough the magnets will 
have the same temperature as them over time. The formula for this new temperature over time is 
as follows: 
­
T (t) T ambient
= e­ mcp  
hAt

­
T i T ambient
As can be seen from the formula, the RHS would eventually go to 0 over time, hence the 
temperature of the magnet T(t) would become equal to ambient temperature over time. For this 
reason the group has significant work to conduct in order to find solutions to make these motors 
function in high temperatures. 
Another important factor to consider in the future would be the protection of the integral 
electronics components, such as the antenna, flight controller and battery. These components 
would not be able to handle dust or debris contact, or temperatures as high as 1000 °F. The group 
would need to find the necessary operational temperature of these electronics, then find a way to 
keep their temperature regulated. The temperature could be regulated through a form of coolant 
or air circulation above them, and could be enclosed by a metal enclosure that could potentially 
double as a heat sink to help protect the electronics. 
Along the same vein of temperature resistance, the foam core used to make the wing, tail 
wing, and platform holding the electronics would not be able to handle temperatures as high as 
specified in the design criteria. Polystyrene, which has a melting point of 464 °F, would definitely 
not be able to handle temperatures as high as 1000 °F. While this material is excellent in its 
cheapness, low density and adequate rigidity, the group would have to select a more durable 
material in the future for these components. If the group still wanted to use the same material as a 
base, the group could coat the material with a material that can handle higher temperatures such 
as fiberglass or carbon fiber. Both fiberglass and carbon fiber have very high melting points: 2075° 
F and 6686 °F, respectively. The primary deciding factor between these two materials would be 
the additional rigidity they provide--fiberglass is more flexible and carbon fiber is more rigid. The 
group would need to decide which characteristic they would value more in these components 
moving forward. 

14 
Beta Prototype Drawings 

 
Figure 5: Beta Prototype final drawings. 

15 
Beta Prototype Bill of Materials 
Only the additional materials needed to modify the alpha prototype to the beta prototype are 
listed in this Bill of Materials. Combine Tables 2, 3, 6 & 7 for full beta prototype BOM. 
Structural Bill of Materials 
Part  Quantity  Note/Material 

Body Tube ID Ø1” –– 8”  1  Prefab carbon fiber 

Body Tube ID Ø1” –– 10”  1  Prefab carbon fiber 

Elbow Wing Mount  4  3D Printed (PLA) 

Angled Tee Joint  4  3D Printed (PLA) 

Cross Tube Tee Joint  2  3D Printed (PLA) 

Wing (S7055 Airfoil Section)  1  Insulating Styrofoam 

Tailwing  1  Foamcore/Styrofoam 

Motor  1   

Propellor 16 x 5.4L  1  Carbon fiber 

Motor Screw  4   

Body Screw ¼” x 20 – 2.5”  10  Stainless Steel 

Body Nut ¼” x 20  10  Stainless Steel 

 
Electrical Bill of Materials 
Part  Quantity  Note/Material 

Electronic Speed Controller  1   

Connective Wiring    Various gauges and lengths 

Tables 6 & 7: Structural & electrical bill of materials for Beta Prototype. 
 

Scaling to Manufacturable Product 

While the group plans to make many more design iterations of this product prior to mass 
manufacturing, it is important to consider a plan for scaling the design manufacturing in order to 
understand how the product could be made accessible and cheap for firefighters around the world 
(especially given the up front cost of the sensors that would be placed on the drone). In order to 

16 
scale up production of this product, it is important to consider the cost and efficiency of making 
each component of this product. Below the most important components of the drone (in terms of 
scalability) are discussed. 

Plastic Motor Mounts and Connectors 

Given the low density and cost of using plastic, the pipe connectors and motor mounts 
would be made of plastic in high volume production. In order to minimize cost and maximize part 
creation, the plastic pipe connectors and motor mounts would be injection molded. Injection 
molding dramatically decreases assembly time, is more exact than 3D printing, and is most cost 
efficient if a high volume of parts are made. In injection molding, casts and machines are generally 
around $10,000-$12,000 up front, but would be approximately $0.03 per part made. As can be 
seen,the cost of this operation would be dramatically lower over time with a high volume of parts 
made. Following the design criteria of the project, a very high temperature-resistant plastic must 
be chosen (that can withstand temperatures as high as 1000 °F). PLA is currently being used in 
prototypes, primarily because it is the plastic that is readily available to the team. The 3D printers 
at Duke use this plastic--while it is relatively light, inexpensive, and holds its structural integrity 
during testing, it does not have the adequate temperature performance characteristics. The 
melting point of PLA is around 350 °F, which is far too low for the needed temperature 
characteristics. For that reason, while PLA will still be used for injection molding, it will not be used 
when production is scaled up. Most of the most temperature resistant plastics that can be 
injection molded are nylon-based. Among these most flame-resistant plastics is Vespel. This would 
likely be the plastic of choice because it can handle a temperature close to 1000 °F. The plastic is 
considered to maintain mechanical properties at 900 °F, which could likely be increased by adding 
heat stabilizer additives to the plastic. However, if during testing the team could not obtain 
adequate performance at 1000 °F even with these additives the team would have to switch the 
joint and motor base material to metal, likely to a particularly heat resistant metal that is not too 
heavy or expensive such as brass. In order to make these parts in high production, the parts would 
be die cast, a process which has a relatively high upfront cost but much lower cost in high volume 
production (like injection molding). 

Wing, Tail Wing, and Electronics Platform 

As discussed earlier, the material of these wing, tail wing and electronics platform is still 
being decided upon. If the group is to continue to use polystyrene (which would be covered in a 
temperature-resistant material such as carbon fiber or fiberglass), the components could be 
CNC’d or injection-molded. For the same reasons mentioned in the Plastic Motors Mounts and 
Connectors section above, the material could be CNC’d for lower volume manufacturing for a 
lower cost, while the polystyrene could be injection molded for a higher up front cost but much 
lower long-term cost. 

Carbon Fiber Tubes 

The group would likely continue to use carbon fiber tubing in mass production if the final 
design were to feature the carbon fiber tubes. These connecting tubes need to mitigate vibration 

17 
(which can be mitigated through their stiffness) and need to be strong enough to handle potential 
impact forces. The rods also need to be temperature resistant as well in order to follow the design 
criteria. Carbon fiber tubes are extremely strong, having a yield strength 10x that of steel (3.5 
GPa). Carbon fiber has a very high rigidity as well, having a modulus of elasticity approximately the 
same as steel (228 GPa). Most importantly, carbon fiber has an extremely high melting point (6686 
°F) which is significantly higher than the operational temperature needed. Carbon fiber is also 
significantly lighter than almost all metals. Though the carbon fiber tubes are likely more 
expensive than metal tubes, the group decided the heat, rigidity, strength, and density 
characteristics of the material deem it worthy to use in the drone. Also, while carbon fiber as a 
material is relatively expensive, tubes such as the ones used in this design are mass produced so 
they would be cheaper to buy. For that reason the group would still use these carbon fiber 
connecting tubes. 

Motors 
Though there is more research to complete in order to narrow down which motor type and 
model to select (as referenced earlier in the Future Work section), it is likely that regardless of this 
research the motor selected will be purchased from a manufacturer. Motors are a very specialized 
component, and it would be beyond the scope of this project to make custom motors. Motors 
would be purchased and then attached to the drone via screws in an assembly line process. 

Electronics 
While each electronic part is currently purchased from online vendors, the group would 
ideally integrate all of these different components onto a single PCB in order to minimize space 
and weight. These PCBs would be made in an assembly process with a PCB manufacturer. 

Propellers 
Given the propellers are made of carbon fiber, they would be able to handle the worst 
environmental conditions being considered for this project. The group would likely still purchase 
these propellers from manufacturers, and would attach them to the motors via assembly line 
operations. 
 
 
 
 
 
 
 
 

18 
Appendix 
 
The link for a youtube video of the drone’s powered flight test: https://youtu.be/sDmcpWNwq-Y  
 
 
 
 
 

 
Figure 6: The drone under construction with the first version wing placed on top. 

19 
  
Figure 7: The fully instrumented drone on the final day of testing, immediately before the test 
flight in the linked video. 
 
 

Resources 
[1] Summary of Low-Speed Airfoil Data Michael S. Selig, Christopher A. Lyon, Philippe  
Giguere, Cameron P. Ninham and James J. Guglielmo. Department of Aeronautical 
and Astronautical Engineering University of Illinois at Urbana-Champaign. 1996. 
https://m-selig.ae.illinois.edu/uiuc_lsat/Low-Speed-Airfoil-Data-V2.pdf 
[2] Airfoil comparison. Airfoil Tools. 2020. http://airfoiltools.com/airfoil/details?airfoil=s7012-il 
[3] Johnston, Jon. Menlo Park Fire Department. Phone call on 1/23/20. 
[4] FEMA 2018 Wildfire Data https://www.usfa.fema.gov/data/statistics/#tab-4 

20