Documenting and Evaluating an Electronics Project

You cannot evaluate backwards

The companion module makes one point: the four report sections hang on the design specification, and a vague specification cannot be evaluated. This module takes the next step, which is where the other half of the marks is lost. Even with a perfect, testable specification, your evaluation can only use the measurements you actually took. You cannot go back in time and measure a stage you have already moved past. That is the rule the whole build has to serve: you cannot evaluate backwards. So realisation and testing are not just about making it work. They are about capturing evidence. Every stage should leave behind a measurement that a later section can point to. If a stage is never measured, the evaluation has nothing to say about it, however well it went. That is why the build climbs a ladder of permanence, and why you test at every rung. A breadboard prototype is quick and easy to change, so faults are cheap to fix there. Stripboard is soldered and harder to alter. A printed circuit board, a PCB, is permanent and reliable but very hard to change. A fault caught on the breadboard costs minutes; the same fault buried in a finished PCB can cost the project. And it is why the instrument has to match what you are measuring. A multimeter reads steady values, voltage, current and resistance. An oscilloscope shows how a signal changes over time. Reaching for the wrong one leaves you with no usable evidence. Test each sub-system before combining them, so that when the whole system misbehaves you already know which part is sound. One honest note: the building itself is coursework and cannot be done here. What can be practised is the thinking that decides whether the build leaves you something to evaluate. Carry one question through everything that follows: will this stage leave me a measurement I can evaluate against later?

Words for building and testing

Five terms, each defined by what it is. They are the tools of realisation and testing.

Tap the two that capture evidence

Tap the TWO statements that leave a measurement the evaluation could later use, rather than an opinion.

• The output was measured at each stage with a multimeter and the readings recorded against the expected values
• The changing signal was captured on an oscilloscope and compared with the shape the design predicted
• The circuit seemed to work quite well when it was switched on
• The finished project looked neat and tidy

Which instrument for this job

You need to check the shape of a signal that is changing rapidly over time. Which instrument should you reach for, and why?

• An oscilloscope, because it shows how a signal changes over time, which a steady-reading meter cannot capture
• A multimeter, because it can measure anything electrical
• Just look at the output device and judge by eye
• No instrument is needed if the circuit seems to work

The quick build against the permanent one

The two ends of the realisation ladder, and why you test before climbing.

Match each tool to its job

Two that follow from no going back

Select the TWO statements that follow from not being able to evaluate backwards.

• You should plan, before building, what you will measure at each stage and with what instrument
• You should record readings as you go, because a stage you did not measure cannot be evaluated later
• It is fine to leave all the testing until the finished PCB
• An evaluation can rest on how well the project seemed to work

Planning tests that feed the evaluation

A method for turning the build into evidence. Work back from each specification point. For every measurable thing the system must do, decide now what you will measure to prove it, and what result would count as meeting it. That is the test the evaluation will later report. Choose the instrument for the measurement. Steady voltage, current or resistance calls for a multimeter. A signal that changes over time calls for an oscilloscope. Pick before you build, so the right kit is to hand. Test each sub-system before combining. Prove each part on its own, and record the reading, so a later fault in the whole system can be traced to the part that changed. Test at each rung of the ladder. Confirm it on the breadboard before soldering to stripboard, and on stripboard before committing to a PCB, because a fault gets more expensive to fix at every step. Write a risk statement, not a warning label. For each practical risk, name the specific hazard and what will be done about it. The detail of safe practice belongs with your teacher and centre, not a revision card. Then the evaluation writes itself: set each recorded measurement against the specification, say whether it was met, and propose an improvement grounded in what the measurement showed. Two habits cost marks. The first is building first and wondering what to measure afterwards. The second is an evaluation of opinions, with no recorded measurement behind them.

Order the build and test ladder

Put the stages of realising and testing a system into a sensible order.

• Plan what to measure for each specification point, and with what instrument
• Build each sub-system on a breadboard and test it
• Combine the tested sub-systems and test the whole system
• Move the working design onto stripboard and test again
• Commit the proven design to a PCB and test the finished system

Build the evidence rule

This is the idea the whole module rests on. Assemble it.

The realisation quick-fire

Five questions on building, testing and evaluating. Three lives.

One brief, planned to be evaluated

Here is the thinking on one project, described in general terms. The two numbers below are given values inside an invented brief, not real standards. Suppose the brief is a light-controlled alarm that must sound when a beam is broken, and the specification says it must respond within a stated short time and must run for a stated period without resetting. Work back from those points before building anything. To prove the response, you will time how long the alarm takes to sound after the beam is broken, and you will use an oscilloscope to see the moment the signal switches, because that is a change over time. To prove the running period, you will run the system for the stated time and record whether it resets, which needs only a clock and careful watching. Both tests are now decided, and so is what counts as a pass. Now build up the ladder. Put each sub-system, the sensor, the processing and the alarm, on a breadboard and test each one against its own expected reading with a multimeter. Combine them and test the whole system against the two specification points. Only when it passes on the breadboard move it to stripboard and test again, then to a PCB and test the finished system. At each stage you recorded a measurement. So when you reach the evaluation, you are not reaching for words like reliable or effective. You have the response time you measured and the running test you ran, you set each against the specification, and you propose an improvement based on what the numbers showed. Nothing had to be measured backwards, because everything was measured on the way.

Complete the realisation facts

A first build that is quick to change while ideas are still being tested, usually on a breadboard, is a _____. A permanent, reliable build that can be reproduced but is hard to change is a _____. An instrument that shows how a signal changes over time is an _____. Because you cannot evaluate backwards, each stage must leave a measurement the evaluation can use.

Three projects to keep evaluable

Three students are partway through a project. In each case pick the move that protects the evaluation.

• A student wants to solder straight to a PCB to save time, skipping the breadboard. What is the risk?
• A student built and combined everything, then found the whole system misbehaves, but tested no sub-system on its own. What went wrong?
• At the evaluation a student writes only that the project worked well and was reliable, with no readings. How could this have been avoided?

Explain how to build so it can be evaluated

A student in the year below plans to build their whole project first and work out the evaluation at the end. Explain why that loses marks and how to build so the project can be evaluated well.

• Explain what it means that you cannot evaluate backwards
• Explain why each stage should leave a measurement the evaluation can use
• Explain the breadboard, stripboard and PCB ladder and why you test at each rung
• Explain how to choose between a multimeter and an oscilloscope
• Explain why each sub-system is tested before combining them
• Finish with the one question to ask of every stage of the build