1

Design Report for the 15th International Submarine Races

Maryland Makos

May 17, 2019

2

SponsorsSponsorsSponsors

JFK Council 5482

3

Table of Contents

History ………………………………………………………….. 4

Executive Summary …………………………………………... 6

Design Philosophy & Goals ………………………………….. 7

Design and Fabrication

Hull …………………………………………………………. 8

Propulsion System …………………………………………12

Control Systems ……………………………………………18

Life Support System ………………………………………. 31

Safety Systems ……………………………………………. 32

Submarine Testing ……………………………………………… 35

Crew Training …………………………………………………… 35

Budget …………………………………………………………… 37

Appendix A: VO2 Max Testing ………………………………… 38

Appendix B: Sonar Ping Technical Specifications ………….. 41

4

History of KIDS

Kids Into Discovering Science (KIDS) is a 501(c)(3) non-profit organization which

exists to assist, energize and empower the youth in our community to learn about

science and technology. Our focus is on hands-on learning with the goal of

entering the International Submarine Races every second year.

The year 2012 marked the 100th anniversary of the sinking of the Titanic. SNAME, the Society of

Naval Architects and Marine Engineers, held a special conference in Oxon Hill, Maryland, to

commemorate the event. In the audience was a 14-year-old boy, a bicycling enthusiast, who was

working to earn his scuba certification for an upcoming Boy Scout adventure. The last presentation

of the SNAME conference was about the biennial International Submarine Races where many of the

human-powered submarines that compete are bicycle powered and all require scuba gear. The teen

was hooked.

This young man gained his parents' support and, after

consulting with other submarine teams, decided the

project was feasible. They were able to recruit three

other families and a Marine to work the project. After a

year of hard labor, their non-propeller entry, Il Calamaro,

crossed the finish line at a top speed of 1.32 knots. The

team also won the Spirit of the Races award.

The Il Calamaro crew learned so much through the submarine project--computer skills, basic physics

and calculus, hands-on fabrication skills, physical training, project management, and more-- that they

wanted to "share the wealth". The team brought

their sub to the local community center, and many,

many youth were eager to be part of the next

submarine team.

KIDS first official submarine entry as a nonprofit

was Nautilus. Once again, after a rigorous year of

learning and building, the team's entry crossed the

finish line! The top speed was 1.60 knots. And

once again, the team won the Spirit of the Races

award. Nautilus, our first official ISR entry as a non-profit . The team

almost tripled in size from the 2013 crew.

5

Next came Rubber Duckie, a single-person propeller

submarine whose hull was made from carbon fiber using a

mold. Despite steering issues, the team successfully

crossed the finish line.

Our fourth International Submarine Race entry, Maryland Mako, is a

single-person, propeller submarine which includes electronically

assisted steering and pilot alter system.

In addition to the submarine team, KIDS has sponsored the following

activities, as membership and interest suggests: Sea Perch, Junior

First Lego League and VEX IQ teams, Arduino programming classes,

Sea Glide construction opportunities, and a Team America Rocketry

Challenge (TARC) team.

6

Executive Summary

We are a team of families from Charles and Prince George’s Counties,

Maryland, inspiring our students to love science, who designed and built a

one person, human-powered, propeller submarine to compete in the 15th

International Submarine Races at the David Taylor Model Basin in

Bethesda, Maryland in June 2019. The carbon fiber, wet hull is driven by a

gearbox with a 2.5 gear ratio inline with a 3 knots propeller. This report

outlines our efforts, focusing on the scientific methods we employed and

nuggets of knowledge acquired on our journey.

Our focus during this season was to create a social environment where

young minds can prove that they can create an innovative technological

feat, while creating new friendships with their peers. Our goals were push

past our personal limits, expanding our boundaries and stepping outside of

our comfort zones by learning additional tasks, like leadership skills, time

management, and ingenuity, to enhance our skillset.

Team Members

Mini Makos: Elliot Kirkrup (1st grade), Nicolas, Gavin Vincent, Pragya Sitoula, Sanya Sitoula

(4th grade), Crisson Kirkup (5th grade)

Adult Mentor: Paola Addamiano-Carts

Propulsion Team: Griffin Hayes (7th grade), Sophia Gerstman (8th grade), Kritika Oli (9th

grade), Gavin Hayes (10th grade), Robert Polk (10th grade)

Adult Mentor: John Hayes

Steering Team: Scarlett Vincent (7th grade), Liam Vincent (10th grade), Ella Gerstman (10th

grade), Abby Gerstman (college freshman)

Adult Mentors: Martin Carts, Jennifer Gerstman

7

Design Philosophy & Goals

Our focus this season continued to be on hands-on, age-

appropriate stretch learning. This meant that younger

team members, lovingly referred to as “Mini Makos”

focused on activities and experiments targeted on the

basic science principles. Older team members, those in

middle and high school, spent time learning how to code

Ardunio microprocessors, intermediate electric

engineering, and increased their knowledge of 3D

computer aided drafting.

August September October November December January

High Schoolers:

Ardunio Basics Virtual

Class

Mini Makos: Barbie in

a bottle experiment.

Entire group: Sea Glide project

Entire group: Transition to ISR

Fall Activity Timeline

We chose to leverage the Sea Glide program to learn concepts related to underwater electronics

and coding. We intended for the Mini Makos to use the Sea Glides to prototype the efficiency of

various dive plane geometries, however we struggled for three months trying to get the code to work

and ran out of time to use them for prototyping. The kids did learn a valuable lesson—don’t over

tighten screws. The team was so concerned about leaking, they overtightened the screws

surrounding the servo motor which inhibited the functioning of the buoyancy engine.

In order to ensure enough attention on these new skills, the team decided to reuse the carbon fiber

hull from ISR 14. The hull details included in this report describe the activities to create the original

hull. The team intends to reconstruct the front third on the hull to incorporate a new PETG nose

cone.

We divided the team into three groups:

Mini Makos, seven elementary students, responsible for dive planes, emergency systems

Propulsion, two middle and three high school students, responsible for the drive train, propeller

and nose cone

Steering, one middle, two high school, and one college student, responsible for the electronically

assisted steering and pilot alert systems.

8

Design and Fabrication

Deign and Fabrication: The Hull

Hull shape:

Our previous two submarine had been manufactured using one-off construction, fiberglassing over a

foam plug that was circular at every cross-section. The pilots from these subs had commented that

there was extra space on the sides, but it was tight top to bottom. (We had to cut holes in one of

them to make room for the pilot’s foot motion.) Our idea for Rubber Duckie was to use a tuna-like

cross-section, one that was taller than it was wide. Tunas are good swimmers, so it seemed like a

good hydrodynamic choice, too.

We began by measuring pilot dimensions and tracing pilots on large sheets of paper. Next we taped

markers to the pilot’s foot while they pedaled a stationary bicycle, capturing their foot circle. We

attempted to map the critical dimensions out on a piece of

paper, but there was much uncertainty about what was

happening at the hip and knee locations.

We decided to build a “pilot sizer”. Using all “bound for

the trash” components, we generated a system that

mimicked the actual pilot motion. Thus we were able to

take measurements in x, y and z coordinates through the

pedaling cycle. We used these dimensions to generate a

prototype submarine cross-section at critical points and

cut these sections out of plywood. Then we aligned the

sections around a test pilot operating the pilot sizer. We

determined that we could refine the cross-section, making it narrower and less tall.

Using SolidWorks, we input the new sections at critical locations (shoulder, hip, knee and foot) and

generated a lofted design that connected these sections.

Fabrication:

The overarching concept was to make a plug of half the

hull, then make a mold from the plug. The mold would then

be used to make two hull halves which would be assembled

to make the hull. This fabrication technique, if properly

executed, would generate a beautifully smooth hull. It was

decided that all work could be done on a single table, so

our first order of business was to design and build a sturdy

10 foot long table.

Figure 1 . Pilot Sizer

Figure 2 . 3D wood plug frame

9

The Plug

The manufacturing of the plug began by cutting out a

wooden frame on a ShopBot CNC router at TechShop.

The frame pieces, essentially a 3D puzzle, were created

by using the STL file in SolidWorks and importing it to

AutoDesk 123D Make.

The wooden frame was assembled on the work table. As

the frame, made from thin Luan, was a bit wobbly, it was

necessary to scribe a centerline on the work table and

secure the frame to the table, using the line as a guide.

We also glued and stapled the frame pieces to each other.

Quarter circles of polystyrene foam (housing insulation

board) were cut and inserted into the frame. Although the

frame spacing was designed to hold exactly two pieces of

foam, we had to sand down some pieces so that they

would fit in the frame. All pieces were glued in place.

Next we had to fair the foam. We tried sanding and using

a rudimentary hot knife, but the rough fairing was most

effectively done with a grinder. Once the shape was rough

“hewn”, we moved to hand sanding in x patterns to follow the long curves.

After fairing the form, we applied a veiling layer of 2 oz fiberglass using

epoxy resin. Next we painted on a layer of Duratec Gray Surface Primer.

This was hand sanded for over 14 man-hours using finer and finer grit

sand paper.

To prepare the plug for the mold layup, we applied 5 coats of SC Johnson

paste wax, followed by 3 layers of PVA. NB: Due to subsequent

problems we had releasing the mold, it is likely we needed to pay

more attention to applying both wax and PVA.

The Mold

A team of four used paint rollers and brushes to apply orange tooling

gelcoat over to plug. This was followed by four layers of fiberglass

mat with polyester resin. Having your fabric pieces cut ahead of time

makes everything go easier. So does having one person with dry

hands.

Figure 3 . Smoothed plug

Figure 4 . Creating the mold

Figure 5 . Releasing the mold

Figure 6 . Slightly gouged mold

10

Indexing marks were built into the mold. These would help align the hull halves made from the

mold.

When it came time to release the mold, it stuck. We discovered with a little more in depth research

that a number of things might have caused this problem. We could have applied our wax and PVA

improperly. We could have damaged the PVA layer when we painted on the gelcoat; spraying

gelcoat is much preferred. We ended up destroying our plug. There were several patches of

Duratec primer that adhered to the mold. At the advice of our Fiberglast technical sales

representative, we sanded these off. The scratches left on the mold meant we had to do more

sanding. We consulted with a local body shop, a boat builder, two composites companies, several

engineers and the University of Maryland Composites Research (CORE) Lab to make sure we were

repairing our mold correctly. In the end, we borrowed equipment from UMD’s CORE Lab and

sanded our mold to a Class A finish.

The Hull Halves

The hull is constructed from 2x2 3k carbon fiber twill fabric donated by Argosy International. We

decided to use a ¼” Divinycell H80 core, adding both needed buoyancy and strength. The lay up is

as shown. We used epoxy resin. A layer of Fiberglast 1099 epoxy surface coat was carefully

brushed on to the prepared mold before laying down the carbon fiber cloth so that when the hull

halves were released from the mold they came out with a smooth, durable white surface.

Lay up ------------------------------------------------------------------ 2 layers carbon fiber (2x2 3K twill)

-------------------------------------------------------------------

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Divinycell H80 core, ¼” thick

------------------------------------------------------------------- carbon fiber (2x2 3K twill)

The Hatch

The hatch was cut from the hull on a beveled angle from the hull, so that the hatch fits flush to the

hull without dropping into the hull. The exposed hull core material (Divinycell) was coated with

resin after the cut was made. The Divinycell core makes the hatch positively buoyant, facilitating

opening of the hatch.

11

The Windows

Our dream is to have a PETG nosecone. Thus far our manufacturing processes have not be

successful. Our backup plan is to have many small windows as used on SpaceShipOne.

Testing

Some of the testing we did along the way included:

lay up of preliminary carbon fiber squares to see how hard the material was to work with;

how compatible were various materials with the solvents and resins we were using;

lay up carbon fiber with various core materials to see which was the most robust;

was it possible to apply PVA using a foam brush? YES!

could we brush on the epoxy surface coat over PVA without damaging the PVA? YES!

could we drape form a PETG nosecone? (This test was aborted when we melted our heat

gun housing.)

Decisions

The most time-consuming decisions we were confronted with were:

Fiberglass or carbon fiber for the hull;

Divinycell or cork for the core;

When to stop sanding the plug;

How to proceed in repairing the mold once it was damaged.

12

Design and Fabrication: Propulsion

Overall Propulsion/Drive Train Concept

The overall Propulsion/Drivetrain concept for the 2019 ISR submarine is a one person, propeller

driven submarine with improvements from the 2017 drivetrain. The problems with the 2017

drivetrain include:

The gearbox was too big and heavy, and needed to be cut down in certain areas to fit

within the limited space in the rear section of the submarine.

The gearbox wasn’t mounted well in the sub and it wasn’t easy to adjust the location of

the gearbox/pedals to accommodate different pilot heights

The design included a “cassette” of different size sprockets to allow for the ability to

change gears while pedaling. This was a great idea, however, it made the gearbox too

heavy and way too big for the back of the sub

The weight of the gearbox and the size of the gearbox made the back of the sub weighted

down and therefore we had to use a lot of counterbalance weights in the front of the sub

The amount of sprockets in the gearbox made the output shaft off-center at the rear of the

gearbox, which meant that the gearbox was not centered in the submarine (see illustration

“A”). Since the gearbox was off center, the balance of the submarine was off and that

made the sub rotate in a clockwise direction when pedaling. This situation also made if

very difficult for the pilot to properly “steer” the submarine because the submarine was not

upright.

We had issues with the freewheel function of the bicycle hub because it was not meant for

underwater use.

Figure 7: ISR 14 Gearbox

13

New Gearbox Concept and Design

Our new gearbox was designed to be smaller

and lighter weight, but designed with the same

¾” aluminum material as before for durability.

Instead of incorporating multiple “sprockets”

for a gear changing capability, the team

decided to simplify the design and include one

carefully selected sprocket to reduce the

needed space within the gearbox. With the

reduced space in the gearbox, the goal was to

center the output shaft on the rear side of the

gearbox so that the gearbox itself could be

mounted in the center of the submarine. The

bicycle sprocket was used in combination with

readily available bevel gears for two reasons.

First, the bevel gears were needed to

transition from our chain driven sprocket to the

output shaft. Second, the bevel gears

provided part of the gear ratio needed to reach

our designed shaft/propeller RPM based on pilot input

RPM (described below). Our modified bicycle rear

wheel hub was no longer needed and replaced with improved bearings for more reliability

underwater. The illustrations to the right show our current gearbox concept.

The team decided it would be a good idea to build a prototype gearbox, outfitted with the bevel

gears we planned to use along with pedals, a chain and bicycle pedals with a front sprocket. Once

the prototype was completed, we could experiment with different sprockets to select the appropriate

sprocket that would give us the needed gear ratio (see picture of our prototype drivetrain below)

Sprocket Selection Process

The sprocket to use in our gearbox was carefully selected to be used in

combination with our bevel gears to provide the necessary gear ratio to

achieve our designed 200 RPM at the output shaft/propeller. After some

experimenting on stationary bicycles, the team determined that our pilots

would generally pedal between 70 and 80 RPM when estimating the

resistance of pedaling underwater. Based on this information, our

estimated overall gear ratio that was needed was calculated as follows:

a. 200 Output RPM/80 Input RPM = 2.5 Gear ratio

b. 200 Output RPM/70 Input RPM = 2.8 Gear ratio

Figure 8: ISR 15 Gearbox Sketch

Figure 9: ISR 15 Gearbox Prototype

14

With this information , our prototype was used to select our rear

sprocket size to provide the necessary gear ratio. The team

temporarily installed several different size sprockets, one at a

time, and observed our output shaft revolutions, based on one

pilot input revolution using the pedals. The sprocket that we

decided to use (24 teeth), provided 2.5 to 2.8 output shaft

revolutions based on one revolution of the pedals. Once the

team confirmed that the prototype would operate as designed

and provide the gear ration needed, the decision was made to

develop CAD drawings of our gearbox using Autodesk Inventor.

The youngest member of the propulsion team was currently

learning Autodesk Inventor in middle school (age 12), so he

developed the drawings for the team. With the drawings

completed, the team met with a local engineering company who

offered to manufacture our gearbox using the drawings we

provided. The team described what we were able to

accomplished on our drivetrain, how the drawings were

completed and also reviewed the drawings in detail with the

engineers. Our contact at the engineering company also took

our kids on a guided tour of their shop, equipment and facilities

which was very enlightening and entertaining. Figure 10: Gearbox Prototype

Figure 11: Gearbox CAD image

Figure 12: Output shaft

from gearbox

Figure 13: Top-down

view of gearbox showing

sprocket assembly

15

Propeller Design

The design for the 2019 propeller included two

separate propellers incorporating a 2 blade

system with an overall diameter of 1.81 to 3.15

feet overall. One propeller was a bit smaller than

the other. The smaller propeller was designed

for our more powerful pilots that could pedal at

higher RPMs, while the larger propeller was

designed for lower RPMs, and designed to

achieve roughly the same speed.

We used software called OpenProp to design our

3D propellers. The software allowed us to input

different parameters and experiment different

propellers. The team established the appropriate

OpenProp inputs (number of propeller blades,

shaft RPM, thrust, designed speed in knots, hub

diameter and water density – see illustration

below for some of these calculations.

Figure 14: Propeller diameter calculations

Figure 15: OpenProp calculation parameters

16

Figure 16: Calculation and measurements of the prop

Figure 17: Propeller sketch

Figure 18: 3D

representation of the

propeller

Figure 19: 3D representation of the full

propeller and hub assembly

17

Propeller Hub Design

After designing the propeller in OpenProp, we imported the three dimensional propeller blade ge-

ometry from OpenProp into Autodesk Inventor where the propeller blades were formed into a solid

body to prepare for 3D printing. We also used Autodesk to add an interface section at the base of

the propeller blade to connect to our hub.

Figure 20: Prop hub sketch

Figure 21: 3D

representation of the

prop hub

18

Deign and Fabrication: Control Systems

Introduction:

A primary goal of our entry in ISR15 was that the steering will be electrically controlled. The ISR15

rules caution to have a mechanical back-up system is noted but the major differences in an electrical

and a mechanical system make the implementation of both beyond our abilities. Taking all things

into consideration we decided to approach our first attempt in automated steering by breaking out

control systems into three components: electronically-supported steering, a sensor-based pilot alert,

and control surfaces.

Electronically-Supported Steering:

The electronically-supported steering system functionality

includes pilot input (joystick) and four control surfaces. In

addition it includes propeller counter-torque roll

compensation, control surface position indication and

data logging.

Physically, the Controller four control surfaces & their

servos, and an absorbent glass mat battery comprise the

six discrete Components* of the steering system,

connected by a wiring harness.

Technologically, the system is built on a single power

source and digital instead of analog

processing. Computing is distributed at each

Component. Arduino-type microcomputers are used

because of their low cost and provision for very high flexibility in behavior. Data communication

between Components is by I2C protocol, which is a two-wire bussed digital protocol.

Microcontrollers

Arduinos at each Component allow development of each Component to be significantly advanced

before system functionality is tested. Teensy brand arduinos by

PJRC were selected as their form factor is as small as possible

without shrinking into the ATTINY DIP sizes, which have

substantially reduced capabilities. Teensy-LC is a slightly less

capable, less expensive, version which is capable of the simple task

of completing a servo. The Controller is built on a Teensy 3.2, a

more capable arduino.

Figure 22 shows the Components* and inter

communication of the steering system.

Figure 23 The brain of the steering

system—a Teensy 3.2..

19

Communications Protocol

Intercommunication between Components is relatively low bandwidth, primarily involving trading

position commands and position responses between the Controller and each of the four Servos. At

update rates of 10 per second to each Servo and perhaps 32 bits per command and per response,

total communication traffic might be 32 bit * 10 /sec * 2 transactions * 4 Servos < 3 kbps.

Multiple protocol possibilities for this small area network exist, including SPI, RS-232, CAN and

I2C. I2C or IIC or Inter-Integrated Circuit, also known as TWI or Two-Wire-Interface, was selected

because it is a bus which allows addressing without additional data lines.

I2C network distance is limited not by actual length but by total data line capacitance which slows

signal rise time. 200 pF is the nominal limit, which includes device input capacitances and

transmission line capacitance. Coaxial cable has typical values of ~25 pF/ft, which would limit total

length to 8'. Twisted shielded pair has capacitance of 50 to 100 pF/ft. Unshielding the twisted pair

reduces that to about 22 pF/ft and untwisting the wires improves that to as low as 8 pF/ft. However,

the permittivity of water outside the conductor insulators significantly increases capacitance and thus

decreases the bandwidth of the system. In the design phase it is recognized that additional

accommodations will be required. Building the wiring harness out of discrete wires only loosely

gathered increases the overall conductor separation, decreasing capacitance. In addition the I2C

implementation within arduino libraries has the occasionally implemented capability of running the

interface at reduced speeds. Since the nominal 100K or 400 K bps bandwidth is not needed, there

is the possibility of running at least 10x slower, resulting in a proportional 10x improvement in

capacitance limit.

Controller

The Controller assembly consists of the main arduino microcomputer, pilot input in the form of a

custom-adapted joystick and momentary switches, tilt sensing, displays, and a data logger. The

Controller is located underneath the pilot interface with the joystick and buttons.

Construction

The Controller assembly is constructed on plug-in breadboards both for ease of re-construction

during the design phase but also because a tiling of multiple breadboards can conform to the hull of

the submarine better than a printed circuit board. Wiring within the assembly is cut-to-length and

formed flat to the breadboard for robustness to handling.

Joystick:

The joystick includes a small, commercial commodity joystick used in the

PlayStation 2. This joystick has a very fine touch and a limited range of motion to

be used by an exercising underwater athlete with thick gloves on. So a custom Figure 24 Joystick

20

over layer provides a larger handle and motion limiters in order to not break/dislodge the joystick. It

is expected that the force feedback may be too small to effectively inform the pilot of neutral

positions so two accommodations are made: A display indicates both the commanded as well as

the actual position of the control surfaces, and the overlay has provision for additional springs to

increase the return-to-neutral force.

Pushbuttons:

Momentary push-activated Switches as inputs to the arduino will allow the pilot to interact with the

system, performing functions such as light intensity control, emergency beacon control, and

others. Uncommitted lighted buttons are included in the design for evolution of pilot control

functions.

Tilt Sensor:

An arduino-connected 3-axis accelerometer is carefully mounted to align to the submarine

axes. The side-to-side axis value reads zero normally when the hull is at rest and upright. An

unbalancing force like propeller torque will give a non-zero reading. The magnitude of this reading

will inform the arduino of the differential angle in the dive planes needed to counteract the

disturbance. Multiple algorithms are possible, from a fixed differential angle (simplest, easy to

implement and thus most likely to be of benefit) to a proportional differential angle (suggests best

control of the tilt) to an time-integrated correction (perhaps this will offer better control if water speed

information is not available).

Displays:

Smart LEDs (SLEDs), brand named "Neopixels" are implemented. These integrated modules are

approximately the size of dimes when mounted. These contain red, blue, green and white LEDs and

a controller which sets the intensity of each, from off to full brightness. Each SLED takes two power

inputs and a single data input line (and provides a single data output line). They are designed to be

daisy-chained, connected in series so that a controller (the arduino in this case) can send an

appropriate bit stream through the entire series to set each SLED. Arduino libraries

make this functionality very easy to implement. SLEDs are capable of an intense

brightness but will be operated near their minimum intensity to be of compatible

brightness with the murk of the race course.

A string of at least five SLEDs arranged in a vertical line indicate the dive planes

position. Likewise a string of horizontal SLEDs indicate the rudder

position. Depending on the joystick functionality these may be critical to pilot steering. If the joystick

operates essentially as an off-full on switch for each direction, the control surfaces will integrate the

time of switching and move across their entire range only within several seconds.

Figure 25 Neopixels

21

Battery

Power requirements were not well known at the outset of the Steering system design but some

parameters emerged affecting the selection of a battery: Weight is of minor-to-negligible

importance. Safety is a poorly understood parameter but it is known that some lithium battery

technologies have propensity for sudden discharge involving

intense heat. Lithium was excluded from consideration for this

reason. Sealed lead acid batteries, in the form of either the gell

cell or absorbent glass mat (AGM) variety, are both ubiquitous in

home and office and thus highly available, inexpensive and easy to

operate. For this reason the AGM battery type was chosen. A

very standard size used in small Uninterruptable Power Supplies

was entertained, pending power needs. These batteries are 12V

nominal.

Maximum power capacity needed was estimated so: Each servo might draw 0.25 Amp and have a

50% run time. All other electronics might consume 0.5A. Total current then is 4 x 0.25 Amp x 50%

+ 0.5 Amp = 1.0 Amp. Run life set to of 6 hours at full utility for no recharge needed during a race

day. 1 Amp * 6 hour = 6 Amp Hours. This power need computes to be compatible with the intended

battery was selected.

Servo

A selection of inexpensive waterproof servos in the torque range anticipated do not seem to exist

but it is known that inexpensive motors do. The decision to build servos out of motor and other

components for that reason, for educational purposes and for waterproofability was made.

A servo has three main parts to perform the function of positioning a shaft in accordance with a

position command. Obviously a motor to position a shaft is the core of a servo. A sensor of actual

shaft position is also required. A means of discerning the 'error', the difference between the

commanded and actual shaft positions is called an error amplifier.

Range: The maximum angular range of the Servo is estimated to be +/- 30 degrees.

Torque requirement:

Depending on the position of the shaft on a control surface the control surface may be

unconditionally stable or unstable. Torque testing has not been performed so potential instability

had to be avoided. A shaft centered on area of the control surface was selected for nominally

unconditionally stability.

Figure 25 Neopixels

Figure 26 12V Battery

22

Torque required varies with submarine speed and while it is hoped that control surface control will

not be the speed limitation a compromise with available motors and budget was necessary. A back-

of-the-envelope estimate follows: An 8.5" x 11" envelope has a surface area of ~2/3 sqft. Imagine

having that pulled that through water without providing the translational force but having control over

the rotation of the surface. The torque required to hold that surface at some angle is the torque

required of the Servo. If the axis of the envelope is at the center of the area that torque ought to be

due to things like surface lift and turbulence. Poling of Steering team members resulted in

consensus that the torque might be about the weight of a soda (1 - 1.3 lb) at a distance of a foot, or

192 oz-in ~= 14 Kg-cm. This value was used, with the recognition that if it is low, the maximum

speed of the submarine may be affected.

Motor:

High speed low torque DC motors permanently coupled to a high reduction

gear box both provide magnified torque as well as the benefit of shaft

locking, the reluctance of the shaft to turn when the motor is de-

energized. Such a motor has been identified. It is a 12V motor geared to

turn at a maximum of 10 RPM, thus providing a maximum time-to-steer of 1

second for full 60 degrees angle change. It has an advertised torque of 40

Kg-cm, approximately three times the guestimated torque required. Which is

good. They draw a maximum of 850 mA so they have the potential for

exceeding the battery current budget. These double-shafted motors retail for

~$30 and are vended by Uxcell.

Position Sensor:

A simple panel mount potentiometer, joined concentrically with the back shaft of

the motor, provides a resistance value that corresponds to the shaft

position. Note that these potentiometers do not swing a full 360 degrees so it is

critical to not allow the motor to turn without constraint. The ends of the

potentiometer are energized with DC voltage, and the wiper feeds an analog

input to the Servo arduino. Unless additional care were to be taken to over-

voltage the potentiometer ends so that within the possible wiper positions the

voltage spanned the entire arduino analog input, the resolution of the

reading will be reduced. In addition there is significant noise from inter-

bit uncertainty and digital noise within the system. Hence when the shaft

position (potentiometer resistance (voltage)) is read multiple readings

are taken and averaged.

Figure 27: Motor

Figure 28: Potentiometer

Figure 29: Limit switches

23

Error Amplifier:

The Servo arduino takes commands from the Controller in the form of Go To Angle command. It

responds with its current angle and other information (limit switches states, etc.) and subtracts the

commanded angle from the potentiometer reading and if needed, drives the motor to reach the

commanded angle. The Servo will be commanded regularly, and will respond each time with

information but will not drive the motor if it is already as commanded.

The arduino does not have current drive capability (nor voltage range) to

drive the motor directly. Instead it uses low power digital signals to

command an H-bridge driver (L298-based) to drive the motor. The H-bridge

driver has the beneficial feature of having a 5V output. The Arduino,

although a 3.3V device, can be powered from 5V, so only 12V power is

needed to the Servo.

Assembly:

The components of the servos were assembled by constructing a housing from sheet metal. Figure

31 provides the details of the housing, while figure 32 shows the completed servo assembly.

Other

Waterproofing:

Few of the chosen electronic circuitry is waterproof or can be expected to operate reliable under

water. Examination of previous ISR attempts at electronics show a near-universal application of

hard waterproofing to components. Putting electronics into a hard waterproof container results in the

entirety of the increased pressure from water depth on every seal and the entirety of the

container. Depths of ~20' result in pressures of ~10 pounds on every square inch of the watertight

boundary. Home-making, improvising pressure vessels is not a trivial effort and the consequences

Figure 31 Servo housing schematic

Figure 32: Assembled servo

Figure 30: H-Bridge Driver

24

of one failure may be the destruction of important components.

Instead of fighting pressure, the approach of equalizing pressure across a watertight

boundary is being tried. Putting everything in bags, with enough air inside to allow for the bag to

compress to about half it's regular volume, should reduce the pressure differential to almost zero. (It

is not zero, because the total height of the airspace inside a bag results in a different pressure at the

top and bottom of the bag. The top of the bag is at a higher pressure than the water above it and

tries to push outwards. The bottom of the bag is at a lower pressure than the water below it and

tries to push inwards. However, this pressure differential from top to bottom is about 1/2 PSI per

foot of height. A bag that is 6" high has a maximum water pressure of 1/4 PSI.

FoodSaver brand food vacuum bagging is fairly robust and the associated tool repeatedly makes

high reliability seals quickly. Experimentation will be required to estimate the enclosed volume of

compressible air and the volume of air outside structural portions of the enclosed electronics; once

the bag has collapsed against internal objects crushing force and pressure differential start to

manifest.

Both electrical and mechanical (motor shaft) feedthroughs will be required. 0.1" spacing rectangular

headers have been shown to penetrate the bag material easily and seal robustly with hot glue. All

electrical feedthroughs are made in this fashion.

The selected motor has most of its shaft ground with a flat for a set screw. However there is enough

room for a 1/16" O-ring to fit against the body of the gear housing. The inside of the O-ring is

lubricated with silicone grease *after* the outside has been hot-glued to the gearbox. Then the shaft

is passed through a hole in the bag and the bag is hot-glued to the flat surface of the gearbox,

completing the waterproof seal for the shaft.

Wiring Harness:

Removable connections to each waterproof Component are required for development and

maintenance. Power is 12V and maximum current is expected to be less than 1A. Maximum

harness length is approximately 10', 20' with return. If 10% voltage loss is acceptable then 1.2V at

1A = 1.2 ohm over 20' = 0.06 ohm/ft is the maximum resistance for the gauge of wire used. Any

wire larger than 27 AWG will suffice.

Contacts can be expected to corrode if attention is not paid to the less-than-deionized water of the

test basin. Both headers and receptacles are plated with 1/4 micron gold. In the event of systematic

degradation of contacts, the crimped receptacle contacts can be soldered and ultimately the wires

can be soldered directly to the header pins.

Servicing:

25

Each Component is sealed before installation and before electrical and mechanical connections are

made to the system. Each arduino has the USB programming connector routed to the Component

connector; the arduinos may be re-programmed without reaching the waterproofing. If needed the

battery can be replaced with a fresh battery already sealed, or it may be charged through its

connector.

Glossary:

Arduino--Arduino is used in several senses here. Capitalized, Arduino brand of hardware not

included. Use of arduino is for the style of miniature self-contained microcomputer, for any instance

of this type of hardware, for the integrated development programming environment (IDE)

Teensy--Brand name for small arduinos

Component--When capitalized, refers to the six major portions of the Steering electronics: Battery,

main controller and four servos

Controller--When capitalized, refers to the main controller

Sensors--When capitalized, refers to another system operating in cooperation with but

independently from Steering.

Servo--When capitalized, refers to the assembly including motor, servo functionality and

communication

26

Pilot Alert System

Our team has had repeated problems with steering. Mostly, steering our submarine away from

crashing into the sides of the basin. So, this year we decided to venture into the world of sensors.

We found BlueRobotics, which has a wide variety of sensors. We developed a system that uses the

following:

The system is created to sense the distance from the wall, and then if the submarine is too close to

the wall of bottom, the corresponding LED indicator will light up, and alert the pilot.

If {Input (11-20 ft from wall)}

Then {Led(50%)}

Else if {Input (1-10 ft from wall)}

Then{Led(100%)}

Else{Led(0%)}

Figure 32 Pilot Alert System Schematic

Figure 33 Model Basin width considerations for risk of collision with the walls of the model basin.

27

All of the sensors and LEDs we are using are

not supported on 1 teensy board, which are the

boards we are using throughout the steering

system, but we need 2. The pings are the main

issue, as they take up the most room. So, we

have done research of I2C and board to board

communication, so that we can use 2 boards ef-

fectively.

The pings are able to send their distance to the

teensy, which then processes the information

and decides whether or no to turn the LED on or

not. This same series of steps occurs in the

Bar30 depth sensor. If the LED is turned on then

the pilot knows that they need to start moving in

the opposite direction to self correct.

We are hoping that this system is successful in

aiding the pilots, so that in a future race we

might develop a fully autonomous steering sys-

tem.

If {Input (10-22 ft from bottom)}

Then{Led(50%)}

Else if {Input (1-3 ft from wall)}

Then{Led(100%)}

Else{Led(0%)}

D1= to close to bottom

D2= too far from the bottom

1P= left ping

2P= right ping

Figure 34 Model Basin depth considerations

Figure 35 Pilot alert light configuration

Figure 36 Calculations to determine optimal sub depth. The sonar ping has a b. of 30% and the team was concerned that the ping would detect the floor of the basin, rather than the wall, at certain depths.

28

Control Surfaces

The Mini Makos were responsible for designing and constructing the control surfaces. These

include the dive planes, rudder and stabilizers. The location of each surface was chosen to mimic

the mako shark fin locations as closely as practical to integrate with other systems. Fin geometry

was also guided by typical mako shark fin geometry as defined in the detailed SharkFin Guide:

Identifying Sharks from their Fins published by the Food and Agriculture Organization of the United

Nations.

Figure 37 Detail from the SharkFin Guide showing the meticulous research on shark fin geometry that was done.

The Mini Makos identified a number of materials that the fins could be made from: metal, plastic,

clay, wood, cement, rubber and cardboard. Each member of the team was to research one material

and report back to the group on pros and cons of using this material for dive planes and with a

proposed manufacturing method. Based on the research the team members brought back, the

group decided to choose cardboard because it was cheap, easy to get hold of, and easy to

fabricate. Obviously, though, the material would need to be waterproofed, which led to another

experiment.

Prototypes of cardboard dive planes were made and a waterproofing test was conducted (see

photos). The outcome of the prototype suggested that a better dive plane could be made by having

29

the shape precision cut on a laser cutter. A SolidWorks model was drafted and the shapes cut out

at a local manufacturing facility. Also, the prototypes were rectangular in cross-section, rather than

the hydrodynamic airfoil shape one might expect. Several of the Mini Makos expressed concern

over this boxy cross-section. We decided the drag penalty would not prevent us from crossing the

finish line, but that additional testing would be required to see if we could sand cardboard so that

the dive planes could be made into a more hydrodynamic cross-section.

The waterproofing tests suggested that either epoxy resin (West System 105 epoxy resin + 205

hardener) or FlexSeal would make suitable waterproofing coatings. The importance of sealing all

surfaces, particularly the exposed corrugations, was underscored, as well.

Figure 38: Prototyping the first mako shark pectoral fin dive plane from

corrugated cardboard.

Figure 39: The cardboard sandwich. Note that the cardboard sheets

were cut so that the corrugations alternated for 0 & 90 degrees for

improved strength.

30

Waterproof Testing Materials

The candidates:

West System epoxy

Cabot Spar Varnish

FlexSeal

Plaid Royal Coat

Drylok

Rustoleum LeakSeal

Each coating (2 coats) was used to cover 3 cardboard test squares.

After the test squares were dried, any exposed cardboard was

sealed with silicone. The squares are being suspended in a water

tank and observed over a one week period. Results were not

available at the time of this writing, but FlexSeal and epoxy are

hypothesized to be good waterproofing materials.

Figure 40: Waterproofing candidates.

31

Design and Fabrication: Life Support Systems

Primary Air

[UMD VO2 max report included in appendix. Do we need to include calculations?]

Secondary Air

A pony bottle will be used as the secondary air supply. As Team Mary-

land Mako intends for pilots to be lead to the submarine by a support

diver (tender), buddy breathing off the tender’s alternate air, the pony

bottle is intended for the pilot to have an air source to reach the surface

in event of an emergency. The pony bottle selected for this application

is a Catalina Cylinder size 6 outfitted with a Sherwood Scuba 5000.

The regulator has been customized to include a button pressure valve

and a regulator. The equipment was serviced/inspected in April 2019.

The design team opted to use gravity (the pony bottle is negatively

buoyant) and magnets to hold the pony bottle in place on the bottom of

the submarine, immediately adjacent to the primary air tank. Since the

bottle is aluminum (hence, non-magnetic), a sleeve was sewn for the

bottle and a steel tang inserted into the sleeve. Magnets are fitted into

the hull to align with the steel tang in the pony bottle sleeve. The

strength of the magnets is sufficient to hold the air tank in place, but not

so robust as to make it difficult for the pilot to remove the pony bottle.

On-Board Pneumatics

Maryland Mako does not have any on-board pneumatics.

Figure 41: The prototype pony bottle sleeve. The strength of the magnets in the magnetic level stuck to the sleeve was strong enough that the pony bottle could be lifted up.

32

Deign and Fabrication: Safety

Submarine coloration

The Mini Makos conducted a disk using Secchi disks of various color to verify which color was most

visible from the surface. Results are shown in figure 45. Using this information, the team

recommended that the main color of the hull be painted white.

The Mini Makos thought to decorate control surfaces

with the brightly colored Maryland flag motif, thus

relieving the monotony of a white submarine and

highlighting the Maryland in Maryland Mako. Drawing

on surfboard decorating practices of embedding the

design under the finish layer, the Mini Makos conducted

a few simple tests finishing computer-printed designs

with epoxy resin and fiberglass.

All control surface and propeller tips will be spray

painted blaze orange to alert divers to potential hazards.

Figure 46 Maryland flag motif printed on tissue paper and epoxied to a piece of cardboard. The right hand side has also been covered with a veiling layer of fiberglass with barely noticeable change in clarity of design.

Figure 42: A few of the colors tested.

Figure 43: One of the disks at the water's surface.

Figure 44: Mini Makos observing for visibility

Figure 45: Mini Makos visibility test results

33

Rescue egress

The hatch is located on the topside of the submarine. It is

hinged with a short length of webbing at the leading edge of

the hatch. The webbing is attached to both the hatch and the

hull by a rivet. All cut edges or perforations in the webbing

were heat-sealed so that the webbing would not fray.

The hatch was cut from the hull on a beveled angle so that

the hatch fits flush to the hull without dropping into the hull.

The exposed hull core material (Divinycell) was coated with

resin after the cut was made. The Divinycell core makes the

hatch positively buoyant, facilitating opening of the hatch.

The hatch is held tightly in place by a spring-loaded latch pin. A hole was drilled in the latch pin to

allow cables to run from the pin to two locations. The first cable is short and runs to the exterior of

the submarine so that rescue divers can pull the exposed grip and, in turn, the latch pin, so that the

hatch is released. The location of the exposed grip is marked by a 4” blaze orange square marked

“RESCUE”. The second cable runs through a series of guides until a grip reaches the pilot’s field of

vision. Should the pilot need to escape, he/she can pull the grip to release the hatch.

Strobe marking light

Maryland Mako is equipped with a SOLAS approved strobe light, the ACR

Firefly Pro. The light is held in place by a formed sheet metal holster. The

holster allows for the light to be easily removed so that the lithium batteries

can be changed or the light can be turned on/off, as needed.

In previous races, KIDS team members have noted that

SOLAS and USCG approved lights have failed during ISR.

Indeed, KIDS ISR13 entry Nautilus experienced a malfunction with its USCG

approved light. For this reason, Maryland Mako has a backup strobe light, an

inexpensive eGear strobe light rated for underwater use at 300 ft. Although it is

does not meet ISR contestant rules, similar lights have been observed to work

well for other ISR teams.

Figure 48: The ACR Firefly Pro light, about $60.

Figure 49: The eGear Guardian light, $15.

Figure 47: Spring-loaded latch pin

34

Emergency pop-up buoy

Team Maryland Mako’s youngest members, the Mini Makos, staged a revolution. They took great

exception to the term “deadman switch” and elected to use the term “unconscious man switch”,

hopefully drawing a more accurate picture of the emergency situation. After much brain storming,

the group opted to use a cable-operated bicycle brake that would prevent the reel containing the

buoy line from spinning. A conscious pilot would squeeze the lever holding the cable tight and, thus,

the buoy secure to the hull. When rendered unconscious, the pilot’s hand would relax and the spring

-loaded lever would allow the brake to release the reel and the buoy line to feed out freely. A simple

prototype was made and tested at the local swimming pool.

The final buoy will be made from incompressible foam in the shape of a mako shark dorsal fin.

Figure 50: Prototype buoy release system.

35

Submarine Testing

Each of the primary components (hull buoyancy, steering, propulsion, pilot alert, emergency sys-

tems) are tested separately upon completion. Integrated wet tests will take place in a swimming pool

during the month prior to ISR 15.

Crew Training

Pilot Training

Pilots were responsible for their individual training.

Pilots primarily trained on bikes, this prepared pilots for the actual propulsion system in the

submarine

Support Crew Training

Pilot Support Protocol

The support diver team typically consists of four divers. One diver functions as the pilot escort, this

diver ‘carries’ the diver to the bottom while the diver is breathing on their secondary octopus until

they transfer to the air supply in the submarine.

Three support divers are

assigned a spot on the

submarine. One diver is

positioned at the back of the

submarine, this diver supports

and holds the submarine up

while the pilot is being loaded

into the submarine and ensuring

that the rudders and propeller

are not scrapping against the

bottom of the test tank. Another diver is positioned at the hatch. This diver is responsible for

opening and closing the hatch, checking the octopus in the submarine is ready for the pilot the dead

man switch is armed and there are no pieces of equipment in the submarine that might impede the

pilots movement. This diver is also responsible for helping the pilot in to the pilot alter system. The

final support diver is positioned at the front of the submarine, supporting the nose cone and holding

the submarine in place while the pilot is being loaded in. This diver is also responsible for

communicating with the pilot once the hatch is closed and walking the pilot through the final check

before the submarine is brought to the starting line. After the final check, the escort diver will either

signal to a surface swimmer, or ascend themselves, that the pilot is ready to race.

Figure 51: Support crew positions

36

Submarine Recovery

Submarine recovery requires two surface crew members. One crew member drags the front of the

submarine with a hook-device, while the second crew member is positioned behind them with a

second stick to keep the submarine from banging against the side of the tank. If there is only one

surface crew member available, the pilot will stay in the water if possible and swim along with the

submarine.

Surface crew members responsible for submarine recovery position extra life vests in the

submarine to make is positively buoyant, and provide one to the pilot if the pilot remains in the

water. The surface crew members are also responsible for recovering the hatch and ensuring that

the pilot is okay to swim back if they choose to do so.

37

Item Amount Category First Race Fee Payment $ 300 Race Fee

Second Race Fee Payment $ 950 Race Fee

SeaGlide Kits (2) $ 750 Prototyping

Rolling Crates (4) $ 80 Tech Kit

Soldering Stations (3) $ 120 Tech Kit

Laptops $ 1,300 Tech Kit

Microsoft Office Professional (4) $ 100 Tech Kit

SolidWorks Licenses $ - Tech Kit

Epoxy Resin $ 750 Hull

PETG $ 300 Hull

Mini Hull Forms $ 500 Prototyping

Paint $ 700 Hull

Automated Steering System $ 1,200 Steering

Propulsion $ 500 Propulsion

Emergency Systems $ 200 Emergency

Scuba Gear $ 3,000 Logistics

Transport Cart $ 100 Logistics

Pool Testing $ 150 Prototyping

Misc Supplies $ 250 Logistics

$11,250 Spend Plan

$1,600 Tech Kits

$1,400 Prototyping

$7,000 Sub -specific

$1,250 Race Fee

Budget

38

Appendix A: VO2 Max Testing

39

40

41

Appendix B: Ping Sonar Altimeter and Echosounder

Specifications