Electrical Engineering Capstone Reports: A US Guide
How US electrical and computer engineering capstones are graded against ABET outcomes, with requirements traceability, test evidence and IEEE reporting conventions.
A US engineering capstone is graded against accreditation outcomes, not against how impressive the demo looks. ABET criteria drive the rubric in most accredited programmes, which means the report must evidence requirements, constraints, teamwork and ethics as explicitly as it evidences the circuit working. Teams that build well and document loosely lose marks they already earned in the lab.
Design to Requirements, Not to a Demo
The spine of the report is a requirements table. Every requirement needs an identifier, a measurable acceptance criterion, and a test that demonstrates it was met. A requirement that cannot be tested is a wish.
| ID | Requirement | Acceptance criterion | Verified by |
|---|---|---|---|
| FR-01 | System samples the sensor | Sample rate at least 1 kHz, jitter under 2 per cent | Logic analyser capture, Test 3 |
| FR-02 | Wireless link reports readings | Packet loss under 1 per cent at 10 m line of sight | Range test, Test 5 |
| NFR-01 | Battery endurance | At least 12 h continuous operation at 25 C | Discharge test, Test 7 |
| NFR-02 | Unit cost | Bill of materials under 60 USD at quantity 100 | Costed BOM, Appendix C |
| CON-01 | Regulatory | Operates in an unlicensed ISM band | Band selection rationale, section 3.2 |
The right-hand column is what converts a build into evidence. Without it, the grader has to take your word for it.
The ABET Student Outcomes
Accredited programmes assess seven student outcomes. Naming them in your report and pointing to where each is evidenced is unusual, cheap, and reads as professional.
| Outcome | Where a capstone evidences it |
|---|---|
| Identify and solve complex problems | Problem definition and design trade studies |
| Apply engineering design with constraints | Requirements table, constraint analysis |
| Communicate effectively | Report, drawings, oral defence |
| Recognise ethical and professional responsibilities | Safety, regulatory and impact analysis |
| Function effectively on a team | Roles, meeting records, contribution statement |
| Develop and conduct experimentation | Test plan and measured results |
| Acquire and apply new knowledge | Self-taught tools, cited learning sources |
Test Evidence That Counts
A screenshot of a working demo is weak evidence. What counts is a test that could be repeated.
- Instrument and settings named. Scope model, timebase, probe attenuation, trigger condition.
- Conditions stated. Supply voltage, ambient temperature, load, firmware version.
- Repeats. A single capture is an anecdote. Report the number of runs and the spread.
- Pass or fail against the criterion, quoting the requirement ID.
- Failures included. A test that failed and was fixed is stronger evidence of process than an all-green table.
On probes specifically, a great many student measurements are wrong because a 10x probe was read as 1x, or because a long ground lead added ringing that was then reported as a real signal. Say which probe and which ground scheme you used.
Trade Studies Are the Design Process
ABET language centres on "engineering design", which is selection among alternatives under constraint. A report that presents one architecture has not evidenced design, only implementation.
| Criterion | Weight | Option A: MCU + external ADC | Option B: SoC with integrated ADC | Option C: FPGA |
|---|---|---|---|---|
| Sample rate headroom | 0.30 | Good | Adequate | Excellent |
| Unit cost | 0.25 | Medium | Low | High |
| Power draw | 0.20 | Medium | Low | High |
| Team familiarity | 0.15 | High | Medium | Low |
| Toolchain maturity | 0.10 | High | High | Medium |
State the weights before you score, not after. Weights chosen once the preferred option is known are reverse-engineered justification, and graders recognise it.
Project Management Evidence
Most rubrics carry marks for managing the work, and this is the section teams write last and thinnest. What is actually wanted is modest.
- A baseline schedule and the actual one, with the variance explained rather than hidden.
- A risk register with two or three real risks, their mitigation, and what actually happened.
- Meeting records, even brief, showing decisions and who took them.
- Version control evidence, since a repository history is objective proof of contribution over time.
- A budget against actual spend.
The variance column is the valuable one. Every real project slips somewhere, and explaining a slip demonstrates management more convincingly than a plan that claims none.
IEEE Reporting Conventions
| Element | Convention |
|---|---|
| Citations | Bracketed numerals in order of appearance, [1], [2] |
| Figure captions | Below the figure, "Fig. 1." with a sentence-case caption |
| Table captions | Above the table, title in capitals |
| Units | SI, space between value and unit, except degrees |
| Equations | Numbered on the right, referenced as "(3)" |
| Abbreviations | Defined at first use, including in the abstract |
Figure and table caption placement differ from each other in IEEE, which is the detail most often missed.
Signal Integrity and Common Measurement Traps
A large share of "the circuit is behaving strangely" reports trace to measurement error rather than circuit error. Knowing the difference is itself an assessed skill.
| Symptom | Frequent cause | Check |
|---|---|---|
| Ringing on a clean digital edge | Long probe ground lead acting as an inductor | Re-measure with a spring ground tip |
| Amplitude ten times wrong | Probe attenuation mismatched to scope setting | Confirm 1x or 10x in both places |
| Noise at 50 or 60 Hz | Mains coupling or a ground loop | Check the earthing path, try battery supply |
| Aliased waveform | Sample rate below Nyquist for the content present | Raise timebase, add an anti-alias filter |
| Supply sag under load | Inadequate decoupling or thin power traces | Measure at the pin, not at the regulator |
Reporting that you diagnosed a measurement artefact, rather than redesigning a circuit that was working, is exactly the professional judgement the rubric is looking for.
Ethics, Safety and Standards
US capstones assess professional responsibility explicitly. The NSPE Code of Ethics holds paramount the safety, health and welfare of the public, and a capstone report should show that principle applied to a real decision in your project rather than quoted in the abstract.
Concretely: if your device connects to mains, say how isolation is achieved. If it transmits, name the band and why it is permitted unlicensed. If it handles personal data, say what is stored and for how long. If it has a battery chemistry with thermal risk, say what protection is in the design.
Common Mistakes and Fixes
| Mistake | Fix |
|---|---|
| No requirements traceability | Every requirement gets an ID, a criterion and a test |
| Demo screenshots as test evidence | Instruments, settings, conditions, repeats, pass or fail |
| Only successful tests reported | Include failures and the fixes. That is the design process |
| Ethics section quoted, not applied | Tie the principle to a decision you actually made |
| IEEE captions placed wrongly | Figures below, tables above |
| No individual contribution statement | Teams are graded individually too. State who did what |
| Bill of materials with no quantity pricing | Cost at a stated production quantity, with sources |
Suggested Images, Free to Reuse
Hosted on Wikimedia Commons and free to reuse with attribution. Confirm the licence on the file page before publishing.
| Subject | Source | Licence |
|---|---|---|
| Printed circuit board, component side | Wikimedia Commons | CC BY-SA 4.0 |
| Printed circuit board, solder side | Wikimedia Commons | CC BY-SA 4.0 |
| Digitising oscilloscope plug-in | Wikimedia Commons | CC BY-SA 4.0 |
Where Ethical Support Fits
Asking a specialist whether your test plan actually verifies your requirements, or whether your measurement setup introduces the artefact you are seeing, is ordinary engineering supervision. Lab staff and faculty advisers do this continuously.
What stays yours is the design and the measurements. Capstones are usually defended orally, and the questions go straight to why you chose a topology or how you know a reading is real.
Frequently Asked Questions
How long should a capstone report be?
Commonly 30 to 60 pages including appendices, but programmes vary widely. Follow the course specification rather than a general figure.
Do failures belong in the report?
Yes. Documented failure and recovery evidences the design process, which is weighted heavily. Hiding it wastes evidence you already generated.
How do I show individual contribution in a team project?
A contribution matrix mapping team members to sections, subsystems and tests, agreed by the team and signed. Most rubrics require it.
Which referencing style applies?
IEEE in most US electrical and computer engineering programmes. Confirm with the course specification.
Does the prototype have to fully work?
A partially working prototype with rigorous analysis of why the remaining function failed can score well. An undocumented working prototype often scores worse.
How do I cite a datasheet?
Manufacturer, part number, document title, revision and date, plus the access date for an online copy. Datasheets are revised, so the revision matters.
Your Next Step Today
Write your requirements table with the verification column filled in before you build anything else. Any requirement with an empty verification cell is either untestable and needs rewriting, or is a test you have not planned, and both are cheaper to fix now than in the final week.
Trusted Sources
- ABET Accreditation Criteria, including student outcomes. Accessed 12 August 2026.
- IEEE Editorial Style Manual. Accessed 12 August 2026.
- NSPE Code of Ethics for Engineers. Accessed 12 August 2026.
- IEEE. Accessed 12 August 2026.
- Wikimedia Commons, freely reusable images. Accessed 12 August 2026.
Accreditation criteria and course specifications change. Where this guide and your course specification differ, follow the specification.
Related services
Related guides
- Civil Engineering Design Reports: A UK Student Guide
- Mechanical Engineering Project Reports: Australia Guide
- IEEE Referencing for Engineering and Computer Science
All academic writing guides or browse our academic writing services.