CHANGING THE PLAN

Engineers change their designs all the time. The trick is changing one thing without breaking something else.

This page explains how real engineers handle a change, shows every change made to the MS-2000 since the design was locked on Sep 16, 2026, keeps the history of every decision, and saves what we learned for the next project.

MOC-001 MOC-002 MOC-003 MOC-004 History Lessons

What is Management of Change?

Management of Change, or MOC, is a simple rule: once a design is decided, you do not just swap a part. You write down the change, check everything it touches, get a yes from the person in charge, and then update every plan and drawing. Each one gets a number, like MOC-001, and is called a change order.

Think of a birthday party. Moving it from the backyard to the park sounds easy. But now the invitations have the wrong address, the cake needs a cooler, and the piñata needs a tree. A good planner thinks it all through before telling anyone. Factories, refineries, and airplane makers use MOC because a small change they did not think through can cause a big problem.

Why it matters

Everything is connected

Moving the laser aim from 5 feet to 3 feet changed a 3D part, five print plates, the build guide, the tests, and the fair board.

You remember why

At the fair, a judge may ask why there is no light test. The change order has the answer, months later.

Everyone has the same plan

The parts list, 3D model, print files, build guide, and web pages all change together, so nobody builds from an old page.

Surprises show up early

Checking on paper found that the big backdrop would not fit the fair table. Finding that in the Parish Hall would have been too late.

The five steps

Use these for any change after the design is locked. A loose screw or a typo does not need one. A different part, size, test, or aim does.

1Say what you want and whyOne or two sentences. "The fair table is 3 feet wide, so the laser should be aimed for 3 feet, not 5."
2Check what it touchesParts, wiring, space, safety, cost, the schedule, and the experiment. Measure and do the math. Do not guess.
3Compare the choicesWrite down at least two ways to solve it, and why you picked one.
4Get a yesDad approves it. Now it is part of the locked design.
5Update everything and test it3D model, print files, parts list, build guide, and web pages. Then write the tests that prove it works.

Where it fits in our process: brainstorm, requirements, pick a design, lock the design, order the parts, build. The MS-2000 design was locked on Sep 16, 2026. Everything after that goes through these five steps before anything is ordered or printed.

Our change orders

4 changes since the lock, newest last. The next one will be MOC-005.

MOC-001ApprovedSep 16, 2026 · approved by Dad

A store-bought mosquito and a foam board backdrop

Before

Three 3D printed mosquito bodies, each made of two white halves with the light sensor and red LED eyes inside.
A printed flight zone box with a black top and a black back.

After

A stuffed mosquito from the store with a small printed pod sewn on. The pod is a white cup 40 mm across plus a cap that holds the light sensor (part P6).
A 30 by 20 inch black foam board standing on two printed feet, with a clip, a 1/4 inch dowel, and the pendulum pivot (part P7).
No printed flight zone box.

Why

Less printing, a faster build, and a mosquito that looks real to the camera.

What checking it found

The dot does not always land in the middle

The laser sits under the camera, so the dot can land about 15 mm away from where the camera is aiming. The light sensor is only about 8 by 11 mm, smaller than a fingernail. So the pod became a light collector: the dot anywhere on the 40 mm white face lights up the whole cup, and the sensor at the back sees it.

The eyes could fool the sensor

Red LED eyes inside the cup would light it up and count as a hit by themselves. The eyes moved to the outside of the cup.

Taking out the box took away its lid

The black top of the box kept the beam from going up. Now three things do that job: the black backdrop, a printed tilt stop, and limits in the code.

What had to be updated

3D parts P6, P7, P8Print platesParts list and store listBuild guideFull plan page

Tests that prove it works

Run these when the parts arrive and the MS-2000 is built.

Flick a flashlight at the pod 20 times: 20 hits.
The eyes never set off the sensor by themselves.
Every shot ends on the backdrop.

See the setup in 3Dcommit 15221fb

MOC-002ApprovedSep 16, 2026 · approved by Dad

A new question, and a size test instead of a light test

Before

Question: does a faster mosquito get hit less?
Three tests: speed, distance, and light (room lights, one lamp, lights off).

After

Engineering goal: build a robot that finds, tracks, and hits a flying mosquito by itself.
Science question: what makes the MS-2000 miss more, a faster mosquito or a smaller one?
Target size test: black paper covers over the pod with a 40, 20, or 10 mm hole. Every cover is the same size on the outside, so the camera only has to learn the mosquito once.

Why

The light test mostly measured how well the HuskyLens camera sees in the dark, not Addie’s design. Size connects to real mosquitoes, which are about 10 mm long, and the covers only cost paper.

What had to be updated

Overview pageFair guide: question, hypothesis, variables, resultsLearn pageFull plan pageBuild guide: wand program, cutting the covers, hit threshold

Tests that prove it works

Run these when the parts arrive and the MS-2000 is built.

With each cover on, a flashlight through the hole flashes the eyes.
The camera still tracks the mosquito with each cover on.

See the fair guidecommit c6fa2d5

MOC-003ApprovedSep 16, 2026 · approved by Dad

Aim for 3 feet, because the fair table is 3 feet

Before

The laser crossed the camera’s line at 5 feet. The head tipped the laser up 1.42 degrees.
Tests at 3, 5, and 7 feet, 45 runs in all.

After

The laser crosses at 3 feet (tipped up 2.37 degrees). At 2 and 4 feet the dot is 12.6 mm off, and the pod face can take about 17 mm, so it works from about 2 to 4.3 feet.
Every run is at 3 feet: 2 tests, 30 runs, 2 graphs.
The fair demo is at about 2 feet.

Why

The fair table is only 3 feet wide. Aimed for 5 feet, a demo 2 feet away would miss by about 23 mm, which is more than half the pod.

Choices we compared

Put the turret on a diagonal and keep the 5 foot aimToo tight. No room left for error.

Shift the aim in code, using how big the mosquito looks to the cameraKept as a backup idea.

Re-aim for 2 feetBreaks the longer tests.

Re-aim for 3 feet and drop the distance testPicked.

What checking it found

The big backdrop does not fit

A 30 by 20 inch board plus a 10 inch dowel is too wide for a 36 inch table. At the fair we use a 12 by 12 inch backdrop and a dowel about 4 inches long.

The beam needs a clear lane

Nothing can sit between the turret and the backdrop.

The turret needs fair limits

With a small backdrop, the code needs tighter left and right limits so the turret cannot aim past its edge.

The board has to fold

The trifold wings fold in so everything fits in 36 inches.

What had to be updated

Head part P4Both 3D viewersAll 5 print plates re-slicedHero and share picturesBuild guide steps 13 and 14 and the wand programFair guide, Learn, overview, full planNew board and table pageProject plan v9

Tests that prove it works

Run these when the parts arrive and the MS-2000 is built.

10 out of 10 shots flash the eyes at 2, 3, and 4 feet.
In the fair layout, every shot ends on the small backdrop, with the mosquito at both ends of its swing.

See the table layoutcommit 8a683b7

MOC-004ApprovedSep 16, 2026 · approved by Dad

A power bank that charges like a phone

Before

Four AA batteries in a pack that slid into a bay on the right side of the base. The pack’s switch turned the turret on and off.
A set of batteries lasts about 2 to 3 hours of tracking, and gets weaker the whole time.

After

A USB-C power bank (the same one Dusty uses) lies flat on the floor of the base, ports to the front, held by a printed cradle and a strap.
A USB-C charge port in the right wall. Plug in a phone charger and the turret charges without opening anything.
A power switch with its rocker poking through the right wall is the main on and off.
The wand keeps its 2 AAA batteries.

Why

No more swapping batteries. One charge should run about 4 to 6 hours, and the bank gives a steady 5 volts, so the servos move just as fast on run 30 as on run 1. That keeps the speed test fair.

Choices we compared

Keep regular AA batteriesThey still come out, and they get weaker during a test.

Rechargeable AA batteriesOnly 4.8 volts, and they still come out to charge.

AA batteries with their own USB-C plugThey still come out, and you need 4 cables.

A bare lithium battery with a charging boardNot sealed, needs soldering, and about 1 amp is not enough for two servos.

A store-bought power bank and a charge port in the wallPicked. Sealed and made safe by the company, gives up to 2.4 amps, and uses the same charger as Dusty.

What checking it found

The bank does not fit the battery bay

The bank is 97 mm long and the bay was 70 mm. It lies front to back on the right side of the floor instead, with its back end against the back wall.

The speaker was in the way

The speaker sat on the back wall right where the bank ends. It moved to just left of center.

The plugs need room

The cable plugs stick out of the front of the bank. The bank sits all the way back, which leaves 31 mm, and a 90 degree adapter turns the plug sideways.

The board switch is hidden inside

The Xia mi has its own switch, but it is inside the base where nobody can reach it. So a small in-line switch sits on the right wall with its rocker poking through a window.

A shelf would not print

The base prints upside down, so a flat shelf under the switch would hang in the air. The switch sits on a wall that stands up from the floor plate instead, and the part above it is a 45 degree wedge.

Wands draw too little for a power bank

Power banks turn themselves off when a gadget uses very little power. The wand uses so little that it would get switched off, so it keeps its AAA batteries. The turret uses plenty, so its bank stays on.

5 volts is the lowest the board takes

The board’s power jack takes 5 to 12 volts, and the bank gives 5. A meter test on arrival makes sure the servos still get at least 4.6 volts. If not, the bank plugs into the board’s USB input instead.

Charging while it runs can reset it

Plugging in a charger can blink the power for a moment and restart the micro:bit. Rule: switch off, then charge.

The charge lights are hidden

The bank’s lights are inside the base. Rule: charge it the night before every session.

What had to be updated

Base shell and floor plate (P1 Rev B)Plates 1 and 2 re-slicedBought-part modelsBoth 3D viewers, hero and share picturesParts list and PDF (4 stores, $224)Build guide: safety, parts, wiring map, steps 1, 6, 7, 9, 16, fixesPrint planFair guide and Learn: controlled variablesBuild your ownOverview, full plan, board pageProject plan

Tests that prove it works

Run these when the parts arrive and the MS-2000 is built.

The bank, adapter, and switch fit with the floor plate on, and a charger plugs in from outside.
The 5 volt pins read at least 4.6 volts with both servos sweeping and the laser on.
The side switch turns everything on and off.
20 fast sweeps with no micro:bit restart.
At least 3 hours of nonstop tracking on one charge.
It charges through the side port until all 4 lights stay on.
Still 10 out of 10 hits at 3 feet.

See the new base in 3Dcommit 7971e58

Build guide Build page

Project history

Every decision and change, newest first. Add a line whenever the plan changes.

Sep 16ChangeMOC-004: USB-C power bank with a charge port and power switch in the right wall replaces the 4 AA pack. Speaker moved left. Parts now $224 from 4 stores. Print time about 10 hours 45 minutes.7971e58

Sep 16PlanChanges and lessons page started. From now on every change after the design lock gets a change order and a line here.

Sep 16Design3D viewers got a card at the bottom and chips you slide sideways, like Dusty’s.cf3345e

Sep 16PlanBoard and table page: two trifold layouts that fit a 3 foot table, and 12 sheets to print.8a683b7

Sep 16ChangeMOC-003: laser aimed for 3 feet, distance test dropped. 30 runs instead of 45.8a683b7

Sep 16ChangeMOC-002: new engineering goal and science question (what makes it miss). Target size test replaces the light test.c6fa2d5

Sep 16DecisionPew or boom is picked by touching the wand’s gold logo, because the turret buttons are inside the base. The wand runs every test: A picks the test, B picks the run, A and B together start 30 seconds.b04f515, 25f1b14

Sep 16PlanBuild guide: 16 steps, wiring maps, and 7 MakeCode programs.b04f515

Sep 16PlanSt. Rose fair guide. Schedule refit to the real due date, Feb 1, replacing the spring 2027 guess.0c76d1e

Sep 16ChangeMOC-001: stuffed mosquito with a printed pod, and a foam board backdrop instead of a printed box.15221fb

Sep 16DesignAll printed parts, P1 to P8, modeled and checked for fit and movement. 5 print plates, about 10 hours 20 minutes and 320 g of plastic.5be7d84 to 99b2377

Sep 16FixA photo of the Xia mi board showed a standing micro:bit socket and 4 corner holes. The carrier plate idea was dropped; the floor plate holds the posts.f83c7d3

Sep 16DecisionDesign locked. Final parts list with main and backup stores: $192 from 2 stores.1f7f4ee

Sep 15DecisionAddie picked Design A, the pan and tilt turret. Design B is the fallback.2103f78

Sep 15DecisionNamed MS-2000. Sound effects and saving data are required.53932dd

Sep 15DecisionClass 2 laser under 1 mW kept. A 14 W laser strong enough to kill a mosquito was turned down: ours is 14,000 times weaker, and fairs only allow the safer classes.c0cd8c4

Sep 15DecisionA fake mosquito on a fishing line that the camera can learn.c0cd8c4

Sep 15ChangeLasers became a requirement instead of the air cannon.51ca7ba

Sep 15PlanProject started: brainstorm of an air cannon, a vacuum, and a light beam.fbd9372

Lessons learned so far

What we would do the same way, or differently, next time.

Check that a switch can be reached

The board already had a power switch, but it was inside the base. A switch only helps if a hand can get to it.

Think about how it prints

A shelf that looks fine in 3D can hang in the air when the part prints upside down. Slopes of 45 degrees or more print without help.

Get the school’s dates first

We guessed the fair was in the spring. The real due date is February 1, so the experiments had to move to winter break.

Measure where it will be shown

Tests at 5 and 7 feet could never run on a 3 foot fair table. Measure the table before picking test distances.

Two things side by side only line up at one distance

The camera and the laser are 38 mm apart, like your two eyes. Their lines cross at one spot. The target needs room for error, and the pod’s 40 mm glowing face is that room.

Test your design, not someone else’s part

The light test mostly tested how the camera sees in the dark. A good test changes something about your own design.

When you take something out, check what it was doing

Taking out the flight box also took away the lid that kept the beam low. That job had to go to something else.

Look at a photo of the real part

A photo of the Xia mi board showed where the socket and holes really are, and changed the base before anything was printed.

Big numbers need a source

The 14 watt number came from a science paper. That one number decided which laser we could use.

Keep one list of what a change touches

One change reached the 3D model, the print files, 5 web pages, the build guide code, and the plan. A list makes sure nothing gets missed.

Checklist for the next project

1Get the due dates, the table size, and the display rules before designing.
2Write the goal, the question, and every requirement with a number and a reason.
3Make a 3D model and check that every part fits before ordering.
4Lock the design, then use a change order for anything after that.
5Add a line to the history log the same day something changes.