Facilitator Guide

The pages carry the teaching. A learner working alone, with a kit and a laptop, can complete every band without asking anybody anything. That was the design goal and it is the standard every page is written to.

But one of the twelve pedagogy principles this course is built on cannot be delivered by a page, and that one is working together. Nor can the thing a person does that no page can: asking a learner a question at the moment they are about to go wrong. Those two need you.

Every activity that needs another person also has a written alternative for a learner who genuinely has nobody. Nothing in the course is impossible alone. It is just harder, and slower.

Four kinds of evidence, and all four are needed. There is no half a band. Band 8, the capstone, asks for a fifth: an honest evaluation against criteria the learner wrote before building.

The table below is the shape, not the list. Every band ends with its own Before you move on panel naming the specific artefacts it wants under each heading, and those are what you collect: the calibration walk in Band 4, the ruler-accuracy table in Band 7, the six PWM pin numbers in Band 2, the resistor values read off the board on a module build. Collect the band's list rather than this table, or you will miss the measurement work that is the whole point of Bands 4 and 7.

There is also no failing. The only two results are passed and not yet, and "not yet" always comes with the one specific thing that would change it.

Moving faster or slower

• Already knows it. A learner may attempt any band's check and Make task cold. Pass both and they are awarded that band and everything below it.
• Missed a few weeks. Nothing is lost. Bands are passed on evidence, not attendance, so they resume where they stopped.
• The kit. If you run the loan model, Bands 0 to 7 use a loaned kit and completing Band 8 converts it to ownership. That is a completion incentive and an equipment recovery mechanism in one move.

One entry per band, written as they go, not afterwards from memory.

The most important part is what went wrong. A log with nothing broken in it is a log from somebody who copied, and it should be marked that way.

Bands 0 to 2 are almost entirely physical faults, visible in a photograph. From Band 3 the balance shifts and more of them are thinking faults, where the wiring is right and the code compiles and the thing still does the wrong job.

This is the thing a page cannot do and you can. Check these six, in this order, before you reply. Resist naming the fault until you have been through all six, because boards often have two.

1. Does every LED have its long leg on the pin side?
2. Does any component have both legs in the same column of five?
3. Does the − rail actually reach a GND pin?
4. Is anything sitting across the middle channel that should not be, or not across it when it should be?
5. Are any bare legs touching each other?
6. Is every leg pushed fully in, or is one resting on top of a hole?

Ask before you tell

From Band 3 onwards a learner starts being able to find their own faults, and that ability is worth more than any single fixed circuit. Three questions are usually enough:

• What did you expect to happen, and what happened instead?
• What does the serial monitor say? If the answer is "I did not look", that is the next step, not your diagnosis.
• What was the last thing that worked?

• Work through all nine bands yourself as a learner, handing in the same evidence you will ask of them. It is the cheapest quality control available and it produces your first set of worked examples. Band 5 especially: the delay lesson is far easier to teach once you have watched it fail on your own bench.
• Check the kits you actually bought. Band 3 needs an active buzzer, Band 5 needs a passive one, and Band 5 needs a full-size breadboard or two small ones. Every one of those has a documented fallback, but find out before a cohort starts, not during.
• Check the four late-band parts before Band 5 ends. Band 6 needs an SG90 servo and a four-cell AA battery holder; Bands 7 and 8 need an HC-SR04 ultrasonic sensor; Band 7 needs an LCD1602. The servo has no substitute and the band cannot run without it. The screen has two fallbacks: the four-digit TM1637 display project, which is a complete route through the band, or the serial monitor. Either one costs the band’s best design exercise, the sixteen-character writing, so ask for that on paper either way. Turn every LCD over and check which kind it is before Band 7 starts: those with a four-pin board on the back need the LiquidCrystal_I2C library installed, which needs internet once, and those without take twelve more wires. A cohort with a mix of both is normal and the band covers both paths in full.
• Set up the proof side before the first cohort, not after. Every band ends with a sixty-second video and a five-line post, and that artefact is both the band's evidence and the learner's portfolio piece. Decide three things now: where posts go, what the tag is, and who reviews a first post before it is public. The Show your work page carries the template and the consent checks. Get the first post out of every learner in week one, on whatever they have built by then, because the first one is the hard one and it gets harder the longer it waits.
• Do the part identification once, for the whole hub, and label the bags. Two components in these kits come in two kinds that look alike and behave completely differently: the buzzer (active or passive) and the RGB light (common cathode or common anode, and modules with or without their own resistors). The Which part have I got? page has two sketches that answer all of it in about twenty-five minutes. Doing this once saves every learner in the cohort a wasted evening, and it is the single highest-value hour of preparation on this list. Write the answers and the date on a slip of paper and put it in each box.
• Decide which alternative builds you are offering, and buy for those only. Most bands now have a second route: a keypad lock for Band 3, a rain alarm or a room monitor for Band 4, a stepper dial or a transistor fan for Band 6, a four-digit display for Band 7, and module versions of Bands 1 and 2. None is required. They exist so that six people and two keypads is a timetable rather than a queue, so that a kit without a servo does not close Band 6, and so that somebody working alone can do a band twice by two roads and find out what they actually understood. Each project page lists its own parts.
• If anybody is building the transistor fan, check there are diodes in the box. That build switches a motor about a thousand times a second and every switch-off throws a spike at the transistor. There is no visible symptom, so nobody will report it: parts simply wear out. If you have no diodes, direct those learners to the ULN2003 route on the same page, which has the diodes built in. This is the only build in the whole course where a missing part causes silent damage rather than a build that does not work.
• Budget consumables. LEDs and servos die, and jumper wires fail more often than anything else in the box, producing faults that look exactly like code bugs. Assume around 15% annual replacement.
• Test the cable. Charge-only USB cables cost more first-week hours than any other single thing. Check every cable before it goes in a kit.

The course is built on the Raspberry Pi Foundation's twelve principles of computing pedagogy. Rather than list them, here is where each one is actually enforced, so you can hold the course to them.