P15029: Air Muscle Underwater Fish Phase 2 - Buoyancy

Week 11

Table of Contents


The original PVC Pump Tank resulted in large leaks during run times and when sitting idle in the water. These leaks would eventually be detrimental to the overall performance of the pump. For example, the pump could stop working, produce less pressure, or corrode after sitting in the water for extended periods of time.

Taking this problem into consideration, a new pump was purchased along with a new waterproof box. The waterproof box was tested for durability and leaks by being submerged under water for 35 mins. After 35 minutes, no leaks were present in the box and the build moved forward to integration.

The pump and the waterproof box were integrated to create a whole pump system, with the tubing being tightly secured and siliconed through holes in the waterproof box. This integrated system was then tested by running the pump for 10 minutes while being completely submerged. This test resulted in minimal leakage. The hole and tubing openings were sealed with extra silicon to prevent the minimal leaks and a note in the instruction manual will be made to instruct user to open and empty the box after 20 minutes of run time.

This system has been extremely more successful than the pump and PVC system, having minimal leaks and making the pump component easier to access. This system will be the final pump product attached to and utilized on the fish.

New Pump Subsystem Set-Up

Communication System

Air Test

The communication system testing set-up is comprised of the transmitter, receiver, and an oscilloscope. The transmitter sent out a set signal, and the oscilloscope was used to see if the receiver was receiving the same signal that the transmitter was transmitting.

The air testing was done inside the lobby of Institute Hall (see map below). The receiver was held stationary at one point as the transmitter was walked away at a steady pace until the signal received by the receiver began to distort.

The receiver was able to pick up the signal from a distance of around ~300-340ft (see map below). This test has proven that the system functions properly.

Map of Air Range Test

Underwater Test

The communication system was tested in the Tow Tank. The same set-up with the oscilloscope as the air testing was used. It ultimately tested with the transmitter submerged at the bottom of the tank (~1ft depth) and at the opposite end of the tank as the receiver (~10ft distance). The receiver was in air at the other end of the tow tank. There was no noticeable distortion in the signal at the receiver with this configuration. This test has proven that the communication test can function properly when submerged underwater. Future testing will be performed at the pool to ensure that the communication system will continue to work while being submerged under 3ft of water.

Communication System Tow Tank Testing
Oscilloscope: Benchmark with the transmitter and receiver sitting next to each other in air. Oscilloscope: Signal when the transmitter was submerged and at the opposite end of the Tow Tank as the receiver.

Swimming Algorithm

In order to mimic the body caudal fin motion swimming motion of a fish, it was decided to base the timing of actuating the valves and air muscles on a square wave. The tail will actuate right or left depending on if the square wave generated by the code is high or low at the current time. However, in order to mimic the swimming motion of a fish better, the second tail joint needs to have a modified sinusoidal signal that it follows. This is because if both tail joints actuate at the same time, the entire tail will simply move back and forth without the curving motion a fish has when it swims. In order to achieve this, a phase offset was calculated for the second joint based on the time it took the first joint to fully actuate. Below is a picture of basic calculations and timing for this code, as well as a video of the swimming code being implemented while the fish is in air.

Here is a snip-it of forward swimming code, where Segment 1 and 2 are the two different joints, and A and B stand for which direction:

Segment1A = sgn(cos( PI*(1.0/t1)*t)); Segment2A = sgn( sgn(cos( PI*(1.0/t1)*t)) - sgn(cos( PI*(1.0/t1)*t + (t2/t1)*PI)) ); Segment1B = sgn(cos( PI*(1.0/t1)*t+PI)); Segment2B = sgn( sgn(cos( PI*(1.0/t1)*t + PI)) - sgn(cos( PI*(1.0/t1)*t + (t2/t1)*PI + PI)) );

The current draft versions of the code can be downloaded and then opened with Notepad here.

In addition, LEDs were added to the MOSFETs controlling the valves, so that the user can see what valve is being actuated and when by looking at the board. The second video below shows the LEDs working.

Swimming Algorithm
Swimming Motion of Tail in Air
LEDs for Valves

Turning Capabilities

When testing swimming motion in the tow tank, the fish showed a great mechanical ability to turn within a radius of one fish length or around 18”, while restraining one segment of the fish tail structure.

Turning Video

Forward Motion

The fish also demonstrated the ability to move forward using just the second segment of the tail at varying speeds. The first video shows the fish swimming forward at a slower timed out speed while, the second video shows the fish swimming forward at a faster speed. Please note, this movement was achieved only through the movement of the second tail segment, which shows great promise for future testing of both segments.

Swimming Forward Slowly Video
Swimming Forward Quickly Video

Electronics Hardware

Currently, all of the electronics on-board the fish have been condensed onto one perforated board, including the electronics for communications, the microcontroller, voltage regulator, all of the MOSFETS for the valves and pump and passive components. Below are pictures of the electronics board. The bottom of the board has been plasti-dipped to prevent shorts.


Air Muscles

Considering that there was an air muscle blow-out during the last review, re-evaluation of the mechanical robustness of the air muscles was needed. After reviewing the set-up and build of the current air muscles with Dr. Lamkin-Kennard, new improved air muscles were built to make them more mechanically robust. The orange mesh was burned at the ends to prevent fraying and the crimps were clamped and tightened with extra emphasis on sealing the ends.

Air Muscles
New Air Muscle
Building of New Air Muscles
Tool used to Burn the Edges of the Mesh

Tensioning System

In addition to air muscle robustness, a tensioning system was designed and implemented to address the loosening of the cables after prolonged testing. As seen by the photo below, the tensioning system consists of screws with holes placed within the tail. The cables are tied through the hole multiple times to create an initial hold point. When the cables become loose, the screws can simply be turned to wrap the cable in order to tighten. This system eliminates the need to re-wire and manually tighten the cables after every test.

Tensioning System

Buoyancy System

Push-to-Connect Problem

The ballast tank from the last review was tested further and found to leak air through one or more of the push-to-connects. This was deemed a major issue because the air must stay in the ballast tank for proper functionality. The air in the ballast tank is compressed as water enters and expands when passively draining to push the water out of the tank. Therefore, when air leaks through the push to connects, the ballast tank fails.

To address this problem, multiple root causes were considered and tested.

Push-to-Connect Test

The first element of the ballast system to be tested were the push-to-connects. The push-to-connects were tested using air and a gauge from the lab. The push-to-connects leaked air from the ballast tank after only 2 psi of air pressure. After reviewing the push-to-connect rating (See technical specs below), which was around 230 psi, it was determined that push-to-connects must be failing due to improper implementation or originally damaged from the manufacturer.

Technical Specs of Push-to-Connect

Push-to-Connect Glue Test

Seeing that the push-to-connects were leaking, glue was used to seal them to see if this would alleviate the problem (See photo). As seen by the video below, the glue did not affect the rate of air leakages through the bad push-to-connect. Furthermore, the non-glued push-to-connect did not leak air until very high pressures (which is what was expected). Therefore, new push-to-connects must be built in and glue must not be used to seal.

Glue Test
Circled Push-to-Connect is glued with pipe cement and plugged with
red plug. The other push to connect is sealed with silicon and
functioning properly.
Push-to-Connect Glue Test Video

Push-to-Connect Bottle Test

To prove the push-to-connect damage theory, an Arizona Water Bottle was connected to the pump with a new push-to-connect and filled to capacity before blowing. The video, (snapshot pictures shown below) shows that the new push-to-connect successfully stop air leakages and the ballast system worked as expected. It is noted that the pus-to-connect should remain below the water line to prevent any air leakages, as a precaution.

This method will be used moving forward.

New Ballast Tanks

New Ballast Tank Build

As a result of the testing mentioned above, two new ballast tanks were built with new push-to-connects. These ballast tanks have been tested for any air leakages and successfully passed with no leaks present. These two ballast tanks will be used moving forward.

New Ballast Tanks
New Ballast Tank Build Testing

Back-Up Ballast System

When encountering the ballast issues mentioned above, other options were considered for the design of the tank. An automatic blood pressure monitor and cuff were purchased and tested. The blood pressure cuff was able to pull 2 lb weights off the bottom of the test tub when inflated. Additionally, when placed in a makeshift ballast tank, the blood pressure cuff was able to push water out of the ballast tank to float, while the water was able to compress and push the air out of the blood pressure cuff to sink. This system allows for an open water system and closed air system ballast tank, eliminating the concern for air leakages.

As a result, based on the further testing of the current ballast tanks, this system will serve as a back up in case any other issues present themselves.

Back-Up Ballast System Set-Up and Test
Inflating Blood Pressure Cuff
Back-Up Ballast Tank Testing - Sinking
Back-Up Ballast Tank Testing - Rising

Neutral Buoyancy

Another concern regarding buoyancy relates to the neutral buoyancy of the overall fish. To address this concern, the fish was taken to the tow tank, placed in water to where it sunk to the bottom. The fish was then filled with foam pieces until it reached the surface. This test proved the ability to utilize foam pieces to make the fish neutrally buoyant. Once the subsystems are completely integrated and finalized, foam will be cut and placed strategically to achieve the necessary buoyancy.

Neutral Buoyancy with Foam Pieces

Rolling Test

When initially putting the fish in the water, the weighting affected the stability of the fish causing the fish to roll. In order to address this issue the buoyancy team worked with different materials, like bubble wrap and foam pieces, and placed them strategically around the fish to prevent roll. They were able to prevent roll by taping a piece of bubble wrap to the front top part of the fish (See photos below).

The next step was to test this set-up with the ballast tanks. However, the ballast tanks filled unevenly causing rolling, despite the bubble wrap. At this point, the buoyancy team decided to go ahead and test the bubble wrap set up with the back-up buoyancy system.

Rolling Test Photos

Back-Up Buoyancy Testing

Since the double original buoyancy tanks were inducing roll, as mentioned in the last section, the back-up buoyancy tank was tested. With this setup there was a compressed air line running from the fish outside of the tow tank. This back-up test proved successful, where the fish went vertically up and down. Further testing in a larger/deeper pool is needed to fully understand buoyancy control. See photos and video below for testing and set-up.

Back-Up System Photos and Video
Back-Up Ballast Test

Finalizing Buoyancy System

Although the back-up ballast tank proved to work better than the two smaller ballast tanks during testing, the usage of a blood pressure cuff would have introduced some risk in regards to customer requirements. If the blood pressure cuff was to be utilized, an air compressor would either need to be present on board the fish, introducing new electrical risk, or the fish would have to be tethered to a compressor on the surface of the water, thus preventing the remote control/autonomous requirement set forth by the customer.

Taking this into consideration, it was determined that one more ballast tank should be built and tested. This final ballast tank was the size of the back-up ballast tank, 3” PVC, but, did not include a blood pressure cuff, reverting back to the original design.

In an initial test, this ballast system proved to bring the fish up and down quite well. In subsequent tests, the fish was weighed and analyzed in order to achieve neutral buoyancy for final testing. To achieve correct neutral buoyancy, the fish was suspended from a string to find the center of mass. The fish, in this set-up, was then slowly placed into the tow tank. Weights and Styrofoam were added in the necessary areas (making sure to add/subtract weight equally on either side of the center of mass) to achieve neutral buoyancy underwater. This method also helped alleviate an issue of rolling when the tail moved. Please reference photos below.

With neutral buoyancy achieved and rolling eliminated, the final ballast tank was tested. The video below shows one test, where the fish was able to go up and down successfully.

Final Buoyancy System Testing and Analysis
Final Buoyancy Test Video
This photo shows the fish successfully suspended in the water
with the usage of foam and weights.
These photos show how the successful weighting of the fish
eliminated any previous rolling due to tail movement.
This photo shows the fish out of water - to better view
the foam/weight placement.

Eco-Flex Skin

The Eco-Flex molding has been completed and fitted to the fish, as seen by the picture below. The Eco-Flex will be secured around the fish utilizing Velcro along its bottom edge. The Eco-Flex has the desired stretch and look to accommodate our requirements, while giving an interesting look into the internal mechanisms of the tail structure.

Eco-Flex Tail Molding

Problem Tracking


Schedule to Imagine RIT

Here is a link for the Schedule to Imagine RIT.

Table of Contents MSD II

Week 5 Week 8 Week 11 Week 14

Ballast Tank Testing

Skin Color Testing

Tail Motion Testing


Waterproofing Pump

Composite Head

Electronics Prototype

Tail Motion Testing

Problem Tracking

Updated Project Plan


Communication System

Swimming Algorithm

Electronics Hardware

Air Muscles

Buoyancy System

Eco-Flex Skin

Problem Tracking

Schedule to Imagine


Performance vs Requirements

Imagine RIT

Final Videos

Home | Planning & Execution