EE 517 Spring 2024: Instrumentation and Sensing

Note: To keep things organized, most materials are posted on Canvas.

Design Project

Overview

The design project in EE517 is an opportunity to conceptualize a feasible wireless sensor system that can be deployed in the field and serve some purpose. In your design project final writeup and presentation, you need to answer three big questions:

1. What data will it produce, to what degree of accuracy (e.g., the temperature will be accurate to with 0.002 °C)?
2. What will be its SWaP-C? (Size, Weight, and Power & Cost)
3. Why is this data source important/valuable/useful given this data rate, wireless range & datarate, SWaP-C, and battery lifetime?

Note: The answer to question 3 doesn't have to be ironclad -- you don't have to identify a market, etc., like in a startup funding pitch -- but should get you thinking about whether or not a given object is useful. For example, a temperature sensor that weighs 10 kg, lasts two hours on battery, can only transfer 1 bit every minute, and is only accurate to 10ºC is probably only useful as a doorstop.

Your project must include elements from the four major components that complement a microcontroller-based embedded system:

1. Sensor (analog -- nothing that already speaks a digital protocol!)
2. Interface circuit
3. Wireless communication
4. Energy characterization

Finally, you must show that your design is feasible. This can be demonstrated by describing that your device can be made from real (commercial, academic, rare, expensive, discontinued...) components. Include a complete bill of materials (BOM) in depth, referencing datasheets or other publications for performance calculations, and include cost estimates for major components. Minor components, such as wires and resistors, should be mentioned but specifications and costs do not need to be included. Don't forget to include an estimate about construction/fabrication/manufacturing costs and packaging size/weight/cost.

You are free to select components depending on your future career goals. Students leaning toward a career in industry might prefer to design using components that are already commercial available. Students leaning toward a career in research might prefer to design using components that have not been mass-produced but have only been written up in journal and conference papers.

If you design includes commercial components, it is expected that those components are generally available from vendors such as Digi-Key, Mouser, Arrow, Newark/Farnell/Element41, or even eBay and AliExpress. It's OK if they are out of stock or discontinued as long as you can access datasheets about their capabilities to support your performance claims. If some components are only described in academic literature, cite publications and extract relevant parameters to your analysis such as sensor sensitivity, leakage current, energy per bit, etc.

The design project should include results from hand-calculates based on values sourced from datasheets, published papers, etc., along with simulations developed in appropriate tools such as LTspice for sensor interfaces and Matlab for energy consumption estimates. Building your design is offered as extra credit, and is not required. We want to avoid limitations like supply chain availability, shipping times, costs, and hair-tearing troubleshooting to focus on what's possible with readily-available, and/or cutting-edge, technology. For those who want to build and demonstrate their design, resources such as EPL store credit may be available.

This project is limited to the simulation domain, so it is OK to end up with an incorrect analysis (e.g., noise level would actually be much higher if the system was built) as long as it's well-motivated and well-documented!

A good way to get started on the design project might be:

1. Think of an application you know about that could be served by a sensor that periodically samples/measures some feature of its environment, wirelessly communicates and could be installed/deployed for months or longer.
2. Identify an analog sensor that is sensitive to the environmental feature of interest. Suggested ways to find appropriate devices are to search for, e.g., "humidity sensor" on digikey.com. (If you look on sites like Adafruit and Sparkfun, be sure to keep digging into their designs until you get to the actual components they include in their boards.)
3. Identify the rate at which the sensor needs to sample. Once per month is probably too slow and every millisecond is probably too fast, but what's the right number or range of numbers?
4. From here, other parameters like wireless communication power and range, intended battery life, weight, might start to be more clear.

The three-stage Design Project lifecycle in EE517 is lightly modeled on Department of Defense design reviews. Those are documented in more detail in such places as:

https://swehb.nasa.gov/display/7150/7.9+-+Entrance+and+Exit+Criteria https://en.wikipedia.org/wiki/Design_review_(U.S._government) https://resources.sei.cmu.edu/asset_files/TechnicalNote/2006_004_001_14708.pdf

The design project is an individual assignment.

Stage 1: Pitch (Typically during Week 4)

(15 points presentation, 5 points peer evaluations, 5 points self-evaluation)

Prepare a 3-4 minute "pitch" conceptual presentation to the class about your project. Three slides maximum. Slides must be added to one class-wide Google Slides presentation prepared by Dr. Burnett (link posted to Canvas). We are strictly limited on time and need to fit our presentations in the first half of a lecture period, so presentations can not last longer than 4 minutes.

In your pitch, include the following (one item per slide is probably a good idea):

• Problem to be solved. What will the device do and why is it useful to someone or for some application?
• Block diagram with major components (e.g., sensor, preamplifier, microcontroller, power control...). Include as much detail as you can illustrate cleanly in a schematic. Identify exactly which components you intend to use (e.g., a Honeywell SS466A Hall Effect sensor), if you can at this stage. Most students' diagrams will probably look very similar, which is good!
• An initial guess about battery lifetime, size, and weight with calculations to show how you got to those estimates. It's fine to have numbers that turn out wrong, but you need numbers to start with.

We won't have made it through a lot of the necessary course material by the time you need to prepare for this pitch. That's OK, because you will better identify how later course material fits in to the design project as it is introduced later on.

All students are also required to provide brief feedback to each presenter about their pitch. A form to provide feedback will be provided by Dr. Burnett.

Stage 2: Intermediate Update (Typically during Week 8)

(8 points presentation, 2 points peer evaluations, 15 points written document)

The intermediate update will focus on quantitative sensor performance calculations. The primary deliverables are:

• A detailed schematic captured in a schematic editor of your choice.
• A companion document showing:
  • Calculations of the following sensor performance metrics: dynamic range, sensitivity, detection limit, resolution, and signal to noise ratio.
  • Qualitative statements about response times, drift/stability, selectivity, and cross-sensitivity. In other words, how do you expect your sensor to perform with respect to these metrics?
  • Updated calculations about expected current draw and battery life.

The schematic should include all components involved in obtaining an environmental reading, including ADC, microcontroller, amplifiers, filters, and power sources to deliver needed voltages using your chosen power sources. If a standard schematic symbol for your sensor does not exist, include an electrical model. For example, if your sensor transduces received light into current, your schematic could represent your sensor as a varying current source or a current source with value that is swept during a simulation.

The schematic should include part numbers for each active component (i.e., all components that are not resistors, capacitors, or inductors), or citations to published literature for any component not in production. Ancillary components that do not contribute to the sensor signal, such as decoupling capacitors on power supplies, do not need to be included.

This written update will be complemented by an in-class presentation (again, aim for 3-4 minutes so all the presentations fit in the first half of our lecture period) covering two topics:

1. How does your sensor work? What is the method of transducing an environmental signal into an electrical representation?
2. What is the signal path from environmental signal to bits coming out of an ADC?

Students will be asked to answer Question 1 for each presentation of their classmates.

Stage 3: Final Project (Typically during Finals Week)

(20 points presentation, 5 points peer evaluations, 25 points final report)

This is the the culmination of your design to wrap the three major sensor system components around a core embedded control system: power, sensor, and data. The goal is to produce a design documented in enough detail you could convince someone this will really work, and could be used as the basis for a future research project, venture capital pitch, etc.

Report

Your report should include the following sections:

1. Introduction & Overview (Update & expansion of Pitch)

   • Motivation of project. What is the use case and/or who will it benefit? This should include a summary table including estimated size, weight, cost, battery life, wireless range, and rate of data production.
   • Schematic of complete system, revised from Intermediate Update to be as complete as is practical. A suggested method of developing a complete schematic for your microcontroller (or other complex component) of choice is to consult its datasheet for a reference design. For example, page 15 of the CC2538 datasheet shows Figure 2, an application circuit, illustrating the typical set of external off-chip components required to support the CC2538 System-on-Chip.
   • Any conceptual images you may have generated during the project

2. Sensor (Update & expansion of Intermediate Update as necessary)

   • Section on theory of operation of sensor. How does the sensor/transducer work? Use this section to describe how you end up with an electrical-domain quantity to later process and digitize.
   • Sensor performance metrics updated from Intermediate Update as necessary. Battery life calculations will be broken out into a separate section described later.

3. Data

   • Data rate calculations. How much data does your device produce per unit time and how does that compare with the abilities of your chosen wireless communication system?
   • Wireless range calculations.

4. Power

   • Energy consumption and battery life calculations.

5. Conclusion

   • This is an optional section to reflect on the development arc of your design project and speculate about its future.

6. Appendices

   • Bill of materials: a table of components with quantity, part numbers, estimated costs (if available), and a description of all significant components. Insignificant components include resistors, capacitors, inductors, etc., which are generally interchangeable. These components to not need part number or cost, and should be summarized at the end of the table using quantity and description only, e.g., "five 4.7 kΩ resistors" etc. For all components, form factor (e.g., package type, surface mount resistor size, etc.) is not needed.
   • All slides you presented during the term (pitch, update, and final) in PDF form.
   • References

Presentation

Your presentation should conform to the following specifications:

Final presentation will last approximately 8-9 minutes with 1-2 minutes for questions. Audience members should plan to write a few sentences evaluating the presentation. (Details in the following paragraphs.) Walk us through your proposed design and show that its theoretical specifications will meet the goals of the application (e.g., if handheld, is it light enough to be carried by the user? If it will be left in place for a long time, will the battery last enough time between maintenance cycles?). The presentation should include the following elements:

• Brief overview to remind the audience what your project's goals are
• Brief review of sensor theory of operation
• In-depth analysis of schematic and major electronic components. One incidental goal of this presentation is to make us all more aware of useful electronics, so it's good to spend time on the details of this section!
• Estimated size, weight, cost, battery life, wireless range, rate of data production, and sensor performance metrics
• Comment on how your design changed over time. What misconceptions did you have at the start of the project that had to be rectified at the end?
• Any plans for the project after the end of the course?

A good way to think about your final presentation is to imagine your project is being considered for an allocation of resources (money, engineer-hours, lab space, convincing others to help you, etc.). Pretend that members of the class have been asked on behalf of the resources' stewards to provide a critical evaluation of your project. Assume the need, customers, etc., are well-established and the evaluation is limited to determining whether your proposed design meets the need's requirements and how likely it is to be successful. In other words, are the performance specifications appropriate for the application and is the design thought through in enough detail that it will probably work?

Note that this sort of third-party evaluation is a common freelance job for engineering experts! Venture capital firms, government entities, philanthropic organizations, etc. will frequently hire experts for a few days or weeks to help them steer their resources toward the most efficacious destination. Dr. Burnett has served in such capacities several times for the Department of Defense, National Science Foundation, and several private companies.

Peer Evaluation

Students in the audience will be asked to write a few sentences about each project in the mindset of a critical technical evaluator. Would you recommend a given project be allocated resources necessary to build a prototype, and why? (Assume there are enough resources for all the class's projects, but we don't want to waste them on concepts that still need development.) In this fictitious role, you won't make any more or less money regardless of your recommendation. You can only maintain your reputation by providing a fair and unbiased evaluation. You can also compromise your reputation if you recommend supporting, say, a friend or family member's idea against your better judgment. But this would be a situation in which you have a conflict of interest, and you shouldn't be in that situation in the first place! You may have friends taking this class at the same time, so EE517 is a little different in that regard.

You will not be graded based on how well you do or do not convince the audience (or Dr. Burnett) of the merit of your project, nor how well you do or do not provide an unbiased evaluation. The evaluator mindset is just a tool to help you conceptualize how to structure the information you are presenting, and how to think about a project presentation you are viewing.

Thinking about how your ideas will be critically evaluated by a panel of knowledgeable experts is also a good way to think about work you do for the rest of your career. Sooner or later you will be in a position to promote an idea for development, first by convincing your company or a funding agency that your idea is worth the money, time, etc. You may need to do this several times as your idea is allocated increasing amounts of resources, such as in a Phase I, II, and III Small Business Innovation Research (SBIR) grant or the angel, seed, Series A, etc. rounds involved with startup company fundraising. Being able to package your ideas in a form that accommodates critical evaluation will more clearly communicate your concept and help you move your ideas forward.

Good luck!