Milestone 4

Milestone 4.1 - Optimization

SWE - To improve our software, we focused on adding more customizability to account for more cases of tinnitus. We added the ability to change the loudness of the sound per ear to account for differences in hearing loss and we added more tunability for the harmonic techniques employed so that each person can have a noise that is more comfortable and more similar to their tinnitus. Now, our generated sound is extremely custom to the person and gives an audiologist much more control over how to employ our software. This stages us very nicely for any potential FDA regulations we can face.


We've tested our changes as we tested the methodology ourselves, and have been able to make the necessary changes on the fly to properly bracket the person's tinnitus and play the appropriate sounds that the person finds comfortable and effective. We have spent much less time on fiddling with the generation since we made these changes and have had much smoother testing experiences as a result.


ME - For the mechanical side of the project, the group need the prototype casing that would house the electronics that were developed. The components were developed in Solidworks, where the group had to make a Right, Left, and Central Casing along with the Right, Left, and Central Covers to go along with it. Additionally, the group had to test what is the best material to use, between TPU and PLA, based on the properties and what we wanted the prototype to accomplish. After using the vibrational analysis tool in Solidworks, the group was able to determine that PLA was the best material and as result, printed the parts out in that specific material. This leads to TinnX having a good prototype and proof of concept to show the work that has been done in the past couple of months. 


CPE - So far, TinnX has been using off-the-shelf bone conduction headphones. Milestone 1 outlines the design requirements set out for the project, and one of the most important was that the device shall last the entire day. The team was also advised to explore and compare the effectiveness of bone conduction via the temporal bone and bone conduction via the mastoid process. This prompted the development of a TinnX bone conduction headset, with the entire goal being optimization of currently available technology. When consulting with an experienced audiologist, the main concern was that the previous models of bone conduction via the mastoid process included a headband that users found uncomfortable. 

With this information, the team designed a headset without Bluetooth®, microphones, or LEDs. The headset would wrap around the back of the head instead of over the head. Components were purchased and inspected on arrival (See Costs to view all of the components, where they were sourced, the price of each component, and the total cost of the TinnX TNX-1). An electronics schematic was generated to guide the breadboard phase:

Fig. 1. Electronics schematic for the TNX - 1.

From there, the design was translated to a breadboard to test the functionality and system performance. It is important to note that the team decided to tether this prototype to a laptop instead of increase the size and weight of the device by implementing the battery. 

Fig. 2. Breadboard phase of the TNX - 1.

During testing, the voltage dividers were recalculated for better performance with the Arduino. From there, all of the header pins were de-soldered to allow for final construction. The logic behind the circuit is that a signal is generated on the main laptop (audiologist's laptop), and printed to a .wav file that is then saved onto a microSD card. The Arduino reads the .wav file and sends it via I2S to the two amplifier boards (each with a gain of 9dB). The boards are connected using the same pins and are therefore sent identical signals. Using appropriate voltage dividers allows the boards to select the channel (left or right) it sends to the bone conductors. There is also an inline volume control for each ear to address situations in which a user has an imbalance of tinnitus between both ears. The drawback of the system is that the .wav file has to meet very specific criteria to be read by the Arduino in terms of bitrate and format.

Fig. 3. Multiple angles showing the electronics inside the TNX - 1.

Note: The PCBs and wires of the bone conduction transducers that came from the factory were replaced as they did not perform well under mechanical stress. To replace them, thicker wires were used and new PCBs were cut out to match the shape of the OEM boards. Once everything was resoldered, JB Weld was used to adhere the PCB to the transducers.

Fig. 4. The factory PCB (left) replaced by new PCBs (middle and right).

As will be explained further in the coming sections, this design was ultimately too challenging to achieve a snug fit for efficient conduction. Our test participants also preferred the comfort of the off-the-shelf bone conduction headset compared to our mastoid version. These disadvantages outweighed any incremental effectiveness of bone conduction via the mastoid process. Furthermore, the team decided being compatible with off-the-shelf bone conduction headphones would better position our company. Should TinnX opt to streamline the customer experience, the production of TinnX-branded bone conduction headphones would be delegated to an existing manufacturer with a strategic partnership that would mutually benefit both parties and drive innovation of the product with miniaturization and further optimization. Therefore, the optimal design is the updated software with commercially available bone conduction headphones.


Challenges Faced:

The biggest obstacle to overcome with developing the TNX-1 electronics was the looping of the .wav file being unpredictable. Instead of a perfect loop, there are random cuts and overall inconsistency. This could be due to the libraries used being incompatible with the method. It could also be a hardware error due to hardware incompatibility. As a proof of concept, the code was adjusted to achieve a functional loop, although this is an area for significant future improvement. Apart from that setback, the volume and audio quality rival that of the off-the-shelf headphones. 

Milestone 4.2 - Delivery

SWE - Since the last milestone, we worked with an audiologist to better understand how to fine tune the software on a person to person basis. We formulated a way to bracket people's tinnitus and work with them to create the best tuned sound for that particular case. As a result, we've seen a much larger degree of success than when we created a handful of generalized noises. We have determined that the best way forward is for an audiologist to evaluate a person and work with them to create the sound with our software (a UI has been made to make it non-developer friendly), and have the person download a companion app to activate / deactivate the noise or change it within certain bounds set by the audiologist.


The UI works to generate the sound as requested, but we are not including it here as it includes most of our harmonic techniques. However, anything that can be changed in the code can be changed via this UI so that every audiologist can use it.


ME - After Milestone 3, the group had to work on developing and printing out the parts inorder to house the electronics that was created. This took multiple iterations with the help of trial and error, but ultimately, the group was able to create the housing for the electronics. The electronics was able to fit in its respective housing unit, and after assembling the various parts, led to the finalized prototype that can be seen below. 


Fig. 5. Finalized prototype of the TNX-1.

The electronics are housed inside the casing, with an ear loop inserted, which allows for the person to wear the headphones. 

Milestone 4.3 - Management

Despite all of the challenges and pivots the team faced, all of the goals set for the project at the beginning of the year were achieved and executed well before the Innovation Expo. This was accomplished as a result of excellent team management and organization. Our team structure allowed all of the members to take the lead at different points throughout the year. Each member would manage the project toward a specific goal or stage. Responsibility and accountability were important values held during the year. Below are the "specializations" each team member leaned into by the end of the project beyond their specialization.

Rohit Jayas - Compliance and Regulation

Nicholas DiMeglio - Business Strategy 

Shady Kamel - Infrastructure & Operations


Team Responsibilities for Milestone 4:

Rohit Jayas - Engineered the housing/casing for the electronics of the TNX-1 prototype. Used Solidworks to draft the models, printed out the models in PLA and TPU, assessed the fit of the electronics inside the housing, and then assembled the prototype.  

Nicholas DiMeglio - Develop and maintain software for both bracketing person's tinnitus and for creating appropriate harmonically modified sound files to deliver relief for that particular person. 

Shady Kamel - Design and develop electronics schematics for the TNX-1 prototype. Construct, solder, and code the TNX-1. Ensure deadlines are met on time. Print TPU components needed for the housing. Facilitate purchases of supplies and components. Help conduct tests and obtain data. Document meetings and processes.


All team members contributed to deliverables and all Expo materials generated.