Demonstration deployment — catalogue, prices and customers are sample data. Payments run in labelled sandbox mode.What this means
projectsplanning

A final-year hardware project that survives exam season

Nabil Mustaque ·

Integrate early, order long-lead parts in week three, and keep one page of truth.

Most final-year hardware projects do not fail on the hard technical problem. They fail in week eleven, when three subsystems that each worked on the bench meet for the first time and exams start the week after.

Order long-lead parts in week three

Anything with a lead time over a week — custom boards, specific sensors, enclosures — gets ordered before you are sure of the design. A part you did not use costs less than a month you did not have.

Integrate badly, early

Aim for an end-to-end prototype by week nine, even if every stage is crude. Integration is where the surprises are, and you want them while there is still time.

Keep one page of truth

A milestone map, a budget, a risk list and a weekly check-in. Our Semester Project Planner is that page, printed.

Book review time before you need it

Two hours with someone who has shipped boards before — a consultation or a code and circuit review — catches the decoupling capacitor you forgot while it is still cheap to fix.

More like this

Choosing your first microcontroller

An 8-bit AVR board or a Wi-Fi board? A decision that takes two minutes if you ask the right question.

Why your line follower oscillates (and how to fix it)

The wobble is almost always one of three things.

How to read a datasheet without reading all of it

Four sections answer almost every question you actually have.