Phase 4: Detailed Design

We have achieved our goals for this phase despite the end of semester crunch time and are prepared for MSD 2 in the fall semester.

Progress Report

In efforts to improve our preliminary design and test plans, each team member has been assigned a group of tasks that will help refine our design and test plans. The following table shows a complete list of the teams efforts thus far in the Detailed Design Phase and the completion status for each task:

Prototyping, Engineering Analysis, Simulation

Iterative activities to demonstrate feasibility, including assumptions you made in your analyses or simulations. Have you completed sufficient analysis to ensure that your design will satisfy requirements? Have you included all usage scenarios in your modeling?

Mechanical Analysis:

Torch Head Tank Valve

In order to conserve on power consumption, the design implements a completely mechanical valve. We have come up with a compact design utilizing a float and inverting lever arm to shut off the flow of fuel when the fuel level reaches a certain point.

An initial model to test functionality was 3D printed, in order to see if the valve will move properly; an updated model with proper dimensions will be printed to test the seal of the valve.

Images: Valve Open (1) and Valve Closed (2)

Additionally we further refined our fluid model as seen below. We have decided to use the same model of fluid pump for our testing that is used in the real torch system to create a more realistic test environment and used that pump's operation metrics as a basis for the fluid analysis as seen below.

Electrical Analysis:

By prototyping and initial specifications, the following table of items has been chosen to fit the needs of this project by the standard set by our customer and engineering requirements. By using the power calculations done in phase three we were able to find the right pieces to power the micro-controller and its various peripherals in order to achieve the overall goal. Also by examining the various documentation of each part we found a few simple circuitry requirements needed for these items to function properly together.  The tables below indicate the two main areas of electronics for this project.  The first is the cumulative array of items needed to make the torch head.  The second table details the base station electronics.

Mechanical:

Below are two variations of the proposed design for a mechanical valve, which utilizes a inverse lever arm float and plunger to close off the flow of fuel. For now, no dimensions are finite; this is just to demonstrate and compare functionality. With the float arm distanced from the central axis, unwanted interaction with the torch wick is largely avoided.

For ease of construction, we have elected to create a replica torch tank out of 3-D printed plastic, as the model torches we were given are sealed in such a fashion that the tank cannot be accessed without causing permanent damage to the torches we received.

Electrical:

The image below captures all the pieces of the torch electronics. Examining the image from the top left to the bottom right we see: 3.3V regulator, 10uF and 0.1uF capacitors, the analog discovery module for use in debugging, 2.1mm barrel jack adapter, solar panel holder with panels inside, Nordic dongle soldered to the Pimoroni  prototype board and held in the breadboard with common headers, the solar charger, and the 1200mAh LiPO battery. Missing from this is the copper tape which will be added at a later date as it was order much after the parts seen here.

Starting with the micro-controller and sensor technology the system is very simple hardware wise. With the dev kit attached to the prototype board all the pins are broken out and can be wired via jumpers.  Power, ground, and the copper tape capacitive sensor would all we wired to in this way.  The power would come from the solar panel setup but first it will pass through the regulator circuit formed by the two capacitors and 3.3V regulator IC.  The solar panels would be wired in parallel to keep the voltage constant but increase the charging current.  They would then pass thought the 2.1mm adapter into the charger board.  The charger board would then charge the LiPO battery which would in turn power the electronics in the head. Once the copper tape sensor is built this is all that is needed to form a working prototype of the torch head electronics.  The diagram and BOM does not include wires but it can be said that 20-24 AWG is sufficient in wiring all pieces of this system. Below is what the copper sensor would look like.  As shown it is a piece of the copper tape attached to a cardboard dielectric that will function like a capacitive touch sensor when paired with the micro-controllers on-board oscillator. As for the software of the system, most of the code can be found in example projects from existing libraries in Nordic semiconductors website.  With the system configured as previously stated the code would check the sensor every 10 or so minutes and send that data over Bluetooth mesh to the base station which would relay that information to the BugTorch app over WiFi.

Example code for the capacitive sensing and Bluetooth networking can be found here.

Relevant Files

Bill of Material (BOM)

The team continued to order parts for this phase and do more prototyping. This page is about the stance of the budget and the BOM at the end of phase 3. The team spent approximately $110 compared to last phase. The main purchases was to help build the prototype for the electronics for the torch. The main purchases were the wire kit, male DC power adapter, the ESP-32 MCU, copper tape, and power supply for the base station.

Figure 1: Budget Sheet

We also created the updated BOM for the system. Fig. 2 and Fig. 3 shows the BOM for the base electronics and for the torch respectively. It shows the quantity, the manufacturer, the part #, the description of the part, the cost of the part, and the total part of the item. For the electronics it uses the ESP-32 as the gateway microcontroller, the copper tape for making capacitive sensors, a 120V to 3.3V panel power supply, and a pump for the oil.

Figure 2: BOM for Base Electronics

Fig. 3 shows the BOM for the torch. The parts include: prototype board, main MCU for torches, copper tape for capacitive sensors, solar panels, solar lithium ion charger, lipo battery, 2.1mm adapter, 10uF capacitor, 0.1uF capacitor, and a 3.3V Voltage Regulator.

Figure 3: BOM for Torch

Fig. 4 and Fig. 5 shows the BOM as Fig. 2 and Fig. 3 but in quantities of 100. The price either stayed the same or did not change much unfortunately. It is noted that if we had contacted the manufacturers directly the price could be possible lower. The base cost for the electronics would be $27.84 which didn't change from Fig. 2. However, the base cost for Fig. 5 is a little less than the base cost for the torch in Fig. 3.

Figure 4: BOM for Base Electronics QTY: 100

Figure 5: BOM for Torch QTY: 100

Test Plans

create tests that will show how our design can meet the customer requirements. It doesn't have to be anything fancy. -Owen

Using the table below we have already designed testable values for the engineering requirements. Below here are some tables explaining at a high level, some of the tests we will perform to ensure the customer requirements are met to the best of our abilities.

Mechanical Tests:

Electrical Tests:

Based upon the customer and engineering requirements this design needs to meet certain criteria. The following table describes each requirement and how they will be tested.

Relevant Files

System Design and Flowcharts/System Block Diagram

Since the preliminary design, the design team has solidified our system block diagram. Only small changes were made to the diagram from the preliminary detailed design phase. Changes included the addition of a power supply for the gateway, directing the voltage from the solar panel to the microcontroller instead of the fuel sensor, changing voltage values to match realistic values, and updating the color code for better viewing.

It should be noted that prototyping during this phase and consulting subject matter experts has made the team consider alternative components. Some of blocks in the system block diagram below are represented by different components. For instance, the fuel level sensor in the individual torch has been exchanged to be a custom sensor made of copper tape. Now that the fuel sensor is no longer an off the shelf component and an adaptive capacitive touch sensor using the oscillator pins of the Nordic microcontroller, power is no longer needed for the fuel sensor. Instead, the microcontroller handles the feedback from the sensor and adjusts the oscillation frequency of the oscillator accordingly.

To eliminate the need for more components than necessary and part failure risks, a simple supply converting 120VAC from the home to 3.3VDC for the ESP32 is utilized. This addition can be viewed in the blue box in the bottom left of the diagram. A connection from the supply is made to the gateway microcontroller providing 3.3 V of DC voltage. It has also been confirmed that the team plans to use a one-way valve near the fuel pump to maintain fuel pressure when the pump stops. Thus, there is no need for additional power to be supplied here and only to the gateway.

Moving into MSD II, there may be slight revisions in order to integrate the system together, but ultimately will be the same.

Risk Assessment

For phase 4, the team has re-evaluated the risk assessment chart and made some changes. The team has found new risks to add for this phase but the overall importance went down. The team added 2 new risks. Those risks are 20 & 21. There could be issues with trying to use citronella oil and as substitution water might be used. This in turn could not give us the most accurate data as water and citronella are very different and the flow of each fluid could be drastically different. To minimize this risk the team is going through proper guidelines to get permission to use citronella oil. Along with this, the team is trying to find any substitute substance that be could be used instead of citronella oil. The latest risk item that was added linked in Bluetooth mesh failing. This could happen by something happening to the MCU like not getting power or a part is damaged. This in turn could cause the communication between the mesh network to fail. To minimize this risk we have to make sure each link in the mesh has multiple connections so it can find its way back to the master.

The team has also decided that some risks could get a lower likelihood or severity. Those risks are 4, 5, 6, 10, and 13. For risk 4 during the previous phases, we were worried about dev-kit costs and how that could be really expensive. As the team chose out some dev-boards to use we were able to find them at a cheap cost, reducing the likelihood and severity of spending a ridiculous amount on dev-kits. The next risk that the team thought could be reduced is the market for the product. After seeing the torches and focusing on keeping the aesthetics of the torches we felt confident in this product idea and thought that it would sell well. The next risk that the team thought could be reduced was about covid-19. As vaccines are rolling out and after working for a semester together mostly online the team decided that the severity could be bumped down as it didn't affect our productivity. This leads to our next risk which was the team not being able to meet in person. The main worry for this was about covid-19, but as more things are opening up and the team also meeting a few times in person we felt that it wouldn't be a big risk. The last risk that the team reviewed was the limited knowledge of app development and fluids. The team worked towards learning more about each topic and has a plan on getting the resources to learn about it and felt that it wouldn't cause as big of a problem as we initially thought it would.

A chart with the importance vs the number of risks is shown at the bottom. It shows how the importance and the number of risks changed over each phase.

Design Review Materials

• Pre-read: P21389_DDR_Preread.pdf
• Presentation: Detailed Design Review.pdf
• Presentation Playback: Phase4_playback.m3u

Plans for next phase

At the time of the MSD 2 Phase 1 Review, we would ideally have a finalized test plan, review the events of MSD 1, ensure long lead items are processing as expected, and form a plan for the remainder of the project. It is highly likely there will be some inefficiency at the start of MSD2 as people return to the college routine from summer break

MSD 2 Phase 1 Three Week Plans can be found here: