Objective

Explain why building a competitive robot is a loop — design, prototype, test, revise — rather than a waterfall, and defend building your first version to throw it away.

Concept

New builders treat robot construction like a waterfall: pick a design, CAD it, build it, compete. This almost always goes badly. Real mechanisms fail in ways your imagination cannot simulate. The chain skips at a range of motion you did not anticipate. The roller leaves a dead zone. The C-channel twists because the bracing geometry looked fine in CAD but the torque path is real.

The correct model is a loop. You design, you prototype, you test, you learn something that invalidates the design, and you go back. The first version of any mechanism is built to be thrown away. If you fall in love with your first iteration, you will defend its flaws instead of fixing them.

📐 Engineering tip. The loop has a strategic cost: you need time. A team that starts prototyping the week before a competition will ship their first version and hope. A team that started in the pre-season is on iteration four, and iteration four is better than iteration one in ways iteration one's builder could not predict.

There is one legitimate shortcut: if your previous robot was the same design, the previous robot is the prototype. You already paid the iteration cost. You can skip straight to the refined build. This only applies when the design is genuinely the same — not "same category of mechanism," but same geometry, same motor layout, same drivetrain.

Guided practice

Pick a mechanism you plan to build for the current season. On paper, write three columns:

1. What I think it will do. The behaviours you are assuming.
2. What could go wrong. Every failure mode you can imagine.
3. What I will only know once I have built it. The unknowns you cannot simulate.

The third column is your prototype budget. Every item in that column is a question only a physical build can answer. If the third column is empty, you are overconfident — add three more things to it.

Now plan the iteration count. A simple mechanism needs two iterations minimum. A complex mechanism (lift, stacked intake, pneumatic endgame) needs three to five. Write the iterations on your team calendar with real dates. If iteration three lands the week of your qualifier, you are going to be fixing problems instead of tuning.

Independent exercise

Take a mechanism your team has already built this season. Write a one-paragraph post-mortem: what did you learn in the first build that you could not have learned in CAD? This is proof-of-concept for the loop — if the list is long, the loop worked. If the list is short, you probably did not iterate hard enough.

Common pitfalls

• Treating the first version as "the real robot" and defending its problems.
• Starting the prototype loop too late in the season to actually complete iterations.
• Skipping prototyping because "we CADed it carefully." CAD cannot predict friction, chain slap, or how a game object actually behaves in contact with your rollers.
• Prototyping something you already know works. If you have built it before, it is not a prototype — it is a re-build.
• Confusing "prototype" with "rough build." A prototype answers a specific question.

Where this points next

L1.1 teaches you how to actually run a prototype — what materials, what shortcuts, and what counts as a successful prototype test.

Next up: L1.1 — Prototype before you CAD

Objective

Build a disposable prototype of a new mechanism in cardboard or uncut C-channels, test it against real game objects, and extract specific answers before opening CAD.

Concept

A prototype is not a "rough version of the real thing." A prototype is an experiment that answers a specific question you cannot answer any other way. "Will a 2.5-inch spacing between these rollers leave a dead zone?" is a prototype question. "How does this look?" is not.

The right material for a prototype is whatever is fast. Uncut C-channels let you test geometry without cutting metal you may not reuse. Cardboard lets you test clearance and sightlines. A handful of standoffs and some rubber bands can approximate a whole intake path. The goal is minimum effort, maximum information. If you find yourself making the prototype look good, you are wasting time.

🔧 Build tip. There is a temptation to skip this step and go straight to CAD. The argument sounds reasonable: "CAD is more precise, so why waste time on cardboard?" The answer is that CAD is precise about the things you already know and blind to the things you do not. A CAD model cannot tell you that a game object will wedge between your preroller and ramp at a particular angle because it does not simulate contact, friction, or compliance. A prototype can, in about fifteen minutes.

The one case where you may skip prototyping: when your previous robot already has this exact mechanism. The previous robot is the prototype. You already paid the cost. Do not re-prototype something you have already built and understood.

Guided practice

Pick a mechanism that is new to you this season — something you have never built before. Do the following at the build table:

1. Write the question. One sentence. Example: "Does a 3.5-inch roller spacing handle the game object without jamming at any realistic approach angle?" If you cannot write the question, you do not know what you are prototyping.
2. Gather raw materials. Uncut C-channels (20-hole and 35-hole), a handful of low-strength axles, two motors if movement matters, spacers, bearings, zip ties, rubber bands, cardboard. No cutting. No drilling custom holes.
3. Build the minimum. Bolt enough structure together to hold the mechanism in roughly the right place. If a standoff will work instead of a C-channel, use the standoff. If a rubber band will work instead of a screw joint, use the rubber band.
4. Test with the real game object. Run ten attempts. Count how many succeed. Note every failure mode — not a vibe, a specific sentence like "game object stalled at the transition between preroller and second roller on three of ten runs."
5. Write the answer. If the prototype answered your question, write the answer down and tear the prototype apart. If it did not, change one variable and run the ten attempts again.

Independent exercise

For your current robot, identify one subsystem you built without prototyping. Build a quick prototype version of an alternative geometry (roller spacing, lift bar length, intake ramp angle — whatever is the critical dimension). Test it against a real game object. Compare the result to your current robot's performance. Did the alternative reveal something you would not have seen from CAD alone? Your success criterion is a specific, written finding — not an opinion.

Common pitfalls

• Prototyping without a written question, so the test produces a feeling instead of a finding.
• Making the prototype look finished, which triples the build time and tempts you to keep it.
• Testing without the real game object, or with a foam substitute that behaves differently.
• Prototyping for too long when the answer is already clear after three trials.
• Skipping the prototype on something new because "we can CAD it."

Where this points next

L1.2 teaches you how to extract the constraints a mechanism must satisfy from the problem itself — which is the input to every prototype.

Next up: L1.2 — Reading constraints from the problem

Objective

Read a problem brief — a scoring objective, a field element, a game rule — and produce a written constraint list that drives every downstream design decision.

Concept

A good robot has reasons for every dimension. Not "we liked 14 inches," but "we picked 14 inches because the scoring zone opening is 15 inches and we needed a 0.5-inch margin on each side." If you cannot produce the reason, you are guessing, and a guessed dimension will be wrong in a way you only find out at competition.

Constraints come from four places:

Physical constraints are the hard numbers. The sizing box the robot must start inside. The legal weight. The dimensions of the game objects. The height of the scoring elements. These are non-negotiable and measurable.

Rule-based constraints come from the game manual. Motor count caps. Pneumatic limits. Plastic allowances. What counts as "expansion." These change season to season — your job is to read the manual, not remember last year's numbers.

Temporal constraints are about your calendar. How many weeks until your first qualifier? How many drivers does your team have? How many hours per week at the build table? A beautiful design that needs eight iterations and you have time for two is the wrong design.

Strategic constraints are about what your robot must be good at. You cannot be good at everything. Pick the two or three things your robot will win on, and treat them as hard constraints — everything else bends around them.

📐 Engineering tip. Constraints get read, not invented. If you find yourself making up a constraint, you are probably making up a justification for a decision you already made.

Guided practice

Pick one scoring objective from your current season's game. Grab the game manual, a ruler, and a pen. Produce a constraint list using this template:

If any row reads "I don't know," go find out before moving on. A constraint list with unknowns is not a constraint list.

Independent exercise

Take a subsystem currently on your robot. Produce a retroactive constraint list using the template above. Now ask: does the subsystem as built match the constraint list, or does it match a different list that you never wrote down? If the answer is "it matches a list I never wrote," you have a design that was chosen by vibes, and you should write down what the real driving constraints actually were.

Common pitfalls

• Writing constraints to justify a design you already picked, instead of deriving the design from the constraints.
• Treating strategic constraints as optional — they are the most important kind.
• Skipping the "what we are NOT optimising for" line. A robot with no strategic trade-offs is a robot that is mediocre at everything.
• Using last season's rule numbers. Read this season's manual.
• Calling a preference a constraint. "We want it to look cool" is a preference.

Where this points next

L1.3 teaches you how to turn a constraint list into a defensible design choice using a decision matrix.

Next up: L1.3 — The trade-off grid

Objective

Build a decision matrix for any design choice on your robot, score options against criteria that actually matter, and use the grid to pick a design rather than to justify one you already picked.

Concept

A decision matrix — a trade-off grid — is a table with options along one axis and criteria along the other. Each cell is a score for that option against that criterion. You add up the scores per option and pick the highest. That is the mechanical procedure. The mechanical procedure is not what makes the grid useful.

What makes the grid useful is that it forces you to name criteria before you score options. Naming criteria is where the real thinking happens. A team that argues about which intake design is better usually disagrees because they have different implicit criteria — one person is optimising for build simplicity, another for scoring speed, another for weight — and the disagreement is invisible because nobody wrote the criteria down. The grid drags the disagreement into daylight. Once the criteria are named and agreed, the scoring is usually fast, and the answer is usually obvious.

The criteria

For a VRC design decision, five criteria cover almost every case:

1. Effectiveness. How well does the option do the thing it is designed to do?
2. Simplicity. How complex is the option to design, build, and maintain?
3. Buildability. How confident are you that your team can actually build this option in the time available?
4. Rule compliance. Does the option fit the game manual? Typically a yes/no gate.
5. Strategic value. Does the option move your robot towards the strategic role your team picked?

Scoring

Score each option against each criterion on a small scale — 1 to 5 is standard. Do not use 1-to-100; the precision is fake. A five-point scale forces you to say whether an option is clearly good or clearly bad on a criterion, which is the decision you actually need.

Weighting

Some criteria matter more than others. Unweighted is simpler and works well for most decisions. Weight only when one criterion is dramatically more important than the others.

📐 Engineering tip. The failure mode of trade-off grids is that builders decide first and then build a grid to prove they were right. You can always score to make your preferred option win. The antidote is to build the grid before you commit to an option, and to agree on the scoring as a team.

Guided practice

Worked example — picking an intake roller layout for a generic intake.

Problem: your intake needs to move a game object from the floor into the robot. You have brainstormed three options.

• Option A. Two flex-wheel rollers stacked vertically. Simple, well-understood.
• Option B. A flex-wheel preroller plus a rubber-band-on-sprocket main roller. More compressible at the main roller, slightly more complex.
• Option C. Full rubber-band-on-sprocket all the way through. Lightest and most compressible, but no alignment capability.

Option A wins by one point. B is the second choice if prototyping reveals that A's effectiveness is an actual problem. A narrow win says "these options are close, prototype both if you have time."

Now run the same procedure on a decision your team is currently facing. Three options, five criteria, a single five-point scale. Argue about the scores with a teammate.

Independent exercise

Pick a past design decision your team made this season and build the grid retrospectively. Be honest — score as you would have scored before you knew the outcome. Does the grid still pick the option you built, or does it pick a different one? If a different one, what does that tell you about how the original decision was actually made?

Common pitfalls

• Building the grid after you have already decided, and then scoring to confirm the decision.
• Using too many criteria. Five is usually enough. Ten hides the load-bearing criterion in the noise.
• Using a 1-to-100 scale. False precision.
• Averaging disagreements instead of arguing them out. The argument is where the real learning happens.
• Treating a narrow win (one or two points) as definitive.

Where this points next

Tier 2 begins with L2.1 — the frame-first build doctrine that makes every decision in this tier possible by giving you a rigid structural reference for everything that comes after.

Next up: L2.1 — Build the frame first

Objective

Build a drivetrain frame in the correct order — frame, then blocked components, then wheels and gears, then subsystem frames, then fill — and defend that order against the temptation to bolt things piece by piece as you go.

Concept

Steady your canvas before you start painting. The frame of a robot is a reference surface: every axle, every bearing, every gear downstream of it inherits whatever alignment the frame has. If the frame is square and flat, the robot has a chance. If the frame is assembled piece by piece with components already attached to each half, the two halves never quite line up, and every downstream subsystem is fighting a misalignment nobody put in the CAD.

🔧 Build tip. New builders start by attaching wheels and gears to one side rail, then repeat on the other, then try to bolt the rails together with a crossbrace. This is the fastest known way to produce a drivetrain that will never spin free. The problem is geometric: with gears, bearings, and motors already attached to both sides, the holes on the two rails cannot physically line up unless every component was placed with micrometre precision.

The fix is structural, and it is non-negotiable. Build the frame of a subsystem — the two side rails plus the crossbraces that connect them — with nothing else attached. Square it (L2.2). Bolt it. Now you have a rigid reference. Any component you add from this point onwards is being added to a known-good alignment, not to an alignment that is still forming around the component. This is not a style preference. It is the only build order that does not propagate error.

Once the frame is rigid, add anything that would be blocked later by wheels, gears, or motors — inner braces, wire channels, bolts that pass through two rails, electronics standoffs. Then wheels, gears, and motors. Then the frames of the next subsystems up. Then you fill in.

Guided practice

Build a drivetrain frame in this order. Do not deviate.

1. Cut two drivetrain side rails to length. Confirm they are the same length by laying them side by side — a single hole out and you have already started a misalignment.
2. Cut two crossbrace C-channels.
3. Clamp the two side rails parallel to each other at the distance your crossbraces require. Use a flat table or a squared surface as reference.
4. Attach the two crossbraces to the side rails with four screws each, keps nuts finger-tight.
5. Check square with a 5-wide C-channel across a diagonal (see L2.2). Adjust and tighten.
6. Replace keps nuts with lock nuts. Torque them in an X-pattern so nothing pulls out of square as you tighten.
7. Now add any hardware that will later be blocked — inner braces, through-bolts for motor mounts, cable guides, standoffs for electronics.
8. Install bearings on every drive axle hole. Zip-tie them flat.
9. Drop axles in. Install wheels, gears, spacers, and shaft collars in the sequence the drivetrain was designed for. Do not force anything.
10. Mount the motors. Test spin by hand before powering up. The drivetrain should coast several rotations from a single push.

If step 9 requires force, stop. The frame is not square, or a bearing is missing, or a spacer pack is wrong. Do not muscle it. Muscling it is how you hide a problem that will kill a motor in L3.4.

Independent exercise

Take a drivetrain that has already been built — yours or a teammate's. Remove one crossbrace. Does the frame relax? Does the other crossbrace now seem looser? If anything visibly shifts, the drivetrain was assembled under stress. That shift is the misalignment that was hiding under torque, and it is the reason the motors have been running warm.

Common pitfalls

• Attaching wheels, gears, or motors to a side rail before the rail is part of a squared, crossbraced frame.
• Using a crossbrace as the last thing you install instead of the first.
• Treating "finger-tight, then square, then torque" as optional. Tightening before squaring locks in whatever alignment happens to exist.
• Skipping step 9's hand-spin test and going straight to powered testing.
• Adding blocked hardware after the wheels go on, then spending twenty minutes removing the wheels again.

Where this points next

L2.2 teaches the squaring procedure that step 5 assumes — specifically, how to use 5-wides or 3-wides to check and enforce true ninety-degree corners.

Next up: L2.2 — Squaring the frame

Objective

Square any drivetrain or subsystem frame using 5-wide C-channels (or 3-wides where space is tight), confirm the square visually and mechanically, and defend squaring as the single highest-leverage alignment task in robot building.

Concept

An unsquared frame is a contract your entire robot signs without reading. Every downstream subsystem — drivetrain axles, lift towers, mechanism mounts, odometry pods — inherits the corner angles of the frame they bolt to. If the frame is a rhombus instead of a rectangle, every downstream right angle is off by whatever the frame is off by, and there is no way to fix any one of them without fixing the frame first. Lateral PID tuning (Chapter II L3.3) assumes a drivetrain that goes straight when told to go straight. An unsquared drivetrain does not — it curves, the PID blames itself, and you spend a weekend chasing a code problem that is actually a geometry problem.

🔧 Build tip. A drivetrain that is square by 0.5 mm at the end of week one will still be square in week twelve. A drivetrain that is square "close enough" will drift out of square every time you unbolt a motor.

The fix is mechanical. A VRC C-channel has a known hole pitch. Five holes along one side plus five holes along a perpendicular side define a guaranteed right angle — if the two channels can both bolt flush to a 5-wide C-channel crossing their corner, the corner is ninety degrees to within the manufacturing tolerance of the metal itself. That is the squaring check. It is also the squaring tool: clamp a 5-wide across the corner, bolt the two frame channels to the 5-wide, and the corner is forced square before you tighten the frame joints. When the frame joints are tight, remove the 5-wide. The corner stays square because the frame joints now hold it.

Where space is too tight for a 5-wide, fall back to a 3-wide. A 3-wide is less forgiving of sloppy hole alignment but still guarantees a right angle if both frame channels bolt flush to it.

Guided practice

You have built the frame from L2.1 with the joints finger-tight. Square it now, before you torque anything.

1. Place the frame flat on a smooth table. The table itself is a reference surface — a twisted frame on a flat table will rock, and rocking is a fail.
2. Select a 5-wide C-channel longer than either side of the corner you are squaring. Confirm the 5-wide is itself straight by laying it on the table and pressing; if it rocks, pick another.
3. Offer the 5-wide up to one corner of the frame so it spans both the side rail and the crossbrace. The holes on the frame channels must line up with the holes on the 5-wide. If they do not, the corner is not at ninety degrees — loosen the joint and rotate the channels until they do.
4. Bolt the 5-wide to the side rail with two screws. Bolt the 5-wide to the crossbrace with two screws. Both sets finger-tight.
5. Torque down the frame's own corner joint in an X-pattern. Keep the 5-wide attached while you do this — the 5-wide is what keeps the corner square while you apply torque.
6. Remove the 5-wide. Inspect the corner. The two channels should still meet at a visually clean right angle.
7. Repeat for every corner of the frame.
8. Final visual check: squint down the length of the frame. You should see a single straight line along each rail. If you see daylight between one rail and the next because they have bowed, the frame is twisted.

Independent exercise

After squaring, measure both diagonals of your drivetrain frame with a tape. On a rectangular frame, the two diagonals should be the same length to within about a millimetre. If they differ by more than that, the frame is not square — it is a parallelogram with right-angle-looking corners, and you need to loosen one crossbrace and reseat it. Re-square and re-measure until the diagonals match.

Common pitfalls

• Tightening the frame joints before squaring. Once the joints are torqued, the angle is locked in — good or bad.
• Squaring on a bowed or uneven table. The reference surface matters.
• Using a bent 5-wide. Test-lay it on the table first.
• Forgetting to re-check square after adding heavy components.
• Treating the diagonal check as optional. Two visually-right-looking corners can still produce a parallelogram.

Where this points next

L2.3 compares drivetrain configurations so you can decide what frame you should actually be squaring.

Next up: L2.3 — Drivetrain configurations compared

Objective

Describe the trade-offs of tank, H-drive, X-drive, mecanum, and asterisk drivetrains, explain why tank is the default recommendation, and pick a configuration based on strategic requirement rather than novelty.

Concept

The drivetrain is the single most important subsystem on the robot. Everything bolts to it. Every match depends on it. Choosing a configuration is not a style decision — it is a commitment to a set of physical trade-offs that will define how the robot moves, pushes, turns, and tracks its position.

Tank

Two sides of wheels, one motor group per side, no independent sideways motion. You turn by spinning the sides in opposite directions. Tank is simple to build, simple to tune, simple to program, takes hits well, and is the most forgiving drivetrain to put odometry on because the wheels only spin when the robot is moving in the direction they point. Tank is the default for almost every team that is not advanced enough to have a clear strategic reason to pick something else. This is not a timidity statement. Tank wins at every level of play.

H-drive

Tank with a strafe wheel in the middle for sideways motion. You get limited lateral movement, but the strafe wheel eats build space, adds weight, and rarely gets used in practice. H-drive is a legacy configuration. There is almost always a better reason to pick either tank or X-drive.

X-drive

Four omni wheels at forty-five degrees to the robot frame. Every direction of travel is driven by two wheels at once, and the robot can move in any direction without turning. Very fast, very agile, but omni wheels drift under side-load — any hit from the side spins the wheels freely, which makes the robot hard to defend from and makes odometry tricky because the wheel rotations do not cleanly correspond to robot motion.

Mecanum

Four wheels with angled rollers that allow holonomic motion while letting the wheels stay aligned with the chassis. Heavier than X-drive, slower, and each wheel contributes less force to forward motion because the rollers are always partially redirecting the contact force. Mecanum is rarely the right choice in competitive VRC.

Asterisk / six-wheel omni-traction hybrids

A traction wheel or two in the middle, omnis on the corners. Gives you a centre of rotation that sits on the traction wheel and makes turning-in-place clean, while the omnis let you skid freely in forward/reverse. Worth picking if your strategy is scoring-focused and your autonomous routines demand precise turns.

📐 Engineering tip. Two physical facts decide most of this. First, omni wheels drift under side-load, which means any drivetrain that relies on omnis for traction is easy to push around. Second, holonomic drivetrains (X-drive, mecanum) are harder to put odometry on because the wheel rotations do not directly map to robot displacement. Tank is the easiest drivetrain to get a consistent, accurate position estimate out of.

The rule: pick tank unless you can name the strategic reason you need something else. "I saw a cool X-drive on a stream" is not a strategic reason.

Guided practice

Write a one-page drivetrain decision brief for your current season. Answer each of these questions in two or three sentences.

1. What is the strategic role of the robot? Scorer, defender, generalist, support.
2. How often will the robot take side-hits during a match?
3. How important is top speed versus pushing power?
4. How will you track position? Odometry with tracking wheels, IMU-only, drive encoders.
5. Given all of the above, which configuration? One sentence. Defend it with the earlier answers.

Independent exercise

Pick a configuration that is not tank — X, mecanum, or a hybrid — and write a two-paragraph argument for why that configuration should be your robot's drivetrain. Then write a one-paragraph rebuttal. Which argument did you find more convincing?

Common pitfalls

• Picking a drivetrain because it looks advanced, then discovering mid-season that the build complexity cost you prototype iterations.
• Picking X-drive without planning for defence.
• Picking mecanum because it sounds sophisticated.
• Forgetting that omni wheels drift under side-load.
• Choosing the drivetrain before you choose the strategy. Strategy picks the drivetrain, not the other way around.

Where this points next

L2.4 picks the C-channel layout for your chosen drivetrain so the frame gives you access to motors, gears, and wheels without wasting space.

Next up: L2.4 — C-channel layouts

Objective

Pick between the two valid drivetrain C-channel layouts — ][][ and [[]] — explain why ]][[ and [][] are not options, and build the frame in your chosen layout with motors and gears actually accessible.

Concept

A drivetrain C-channel has an open side and a flanged side. The open side gives you clearance for gears, sprockets, and motor shafts. The flanged side gives you a sandwich surface for bearings and a solid outer wall. Where you put the opening matters, because it determines what you can access later without disassembly and how much width the motors steal from your usable build space.

Two layouts are valid: ][][ and [[]]. Two are not: ]][[ and [][].

][][ has the openings facing inward on the outer rails and outward on the inner rails. Motors mount between the outer and inner rails, in the gap. Gears live in that same gap, sandwiched. This is the layout most teams pick when space between the rails is tight.

[[]] has both outer rails facing inward and no inner rails at all. It is the "simple" layout — two C-channels, openings in. [[]] has more interior space but the drive axles need their own sandwich strategy because the rails alone will not provide one.

]][[ is wrong because the motors have to mount on the outside of the drivetrain, stealing width from whatever sizing constraint the robot is operating under.

[][] is wrong because nothing is accessible. Every motor, gear, and wheel is trapped between the flanged sides of the channels.

🔧 Build tip. Pick ][][ or [[]]. The decision between the two is about motor placement and drivetrain width. ][][ wins when the drivetrain has inner rails carrying structural load and the motors want to sit between the outer and inner rails. [[]] wins when the drivetrain is a clean two-rail design and the motors sit inside the interior space.

Guided practice

Look at your drivetrain frame from L2.1 and L2.2. Check which layout you built.

1. Flip the robot upside down so you can see the C-channel openings directly.
2. Identify every C-channel and which direction its opening faces.
3. Write down the layout. If it is ][][ or [[]], continue.
4. If it is ]][[, your motors are mounting outside the drivetrain and stealing width. Plan a rebuild.
5. If it is [][], nothing will be serviceable without disassembly. Plan a rebuild now.

Now check motor access. With the drivetrain assembled, can you reach every motor screw with a wrench without removing a wheel? Can you reach every gear axle end to install a shaft collar? If not, the layout is wrong for the motor placement you picked.

Independent exercise

Sketch both ][][ and [[]] drivetrains to the same overall width constraint — for example, seventeen inches outside-to-outside. For each layout, calculate the usable interior width between the innermost metal surfaces. Then mark on each sketch where a green-cartridge motor plus gear plus bearing will physically fit. Which layout gives you more interior space? Which layout gives you cleaner motor mounting? For your robot's strategic role, which wins?

Common pitfalls

• Mixing layouts along the same drivetrain. Pick one and commit for every rail.
• Ignoring motor access until after the drivetrain is built.
• Assuming [[]] is always simpler. It is only simpler if you have planned the axle sandwich.
• Copying a layout from a photograph without understanding which orientation the openings face.
• Building ]][[ because it looked correct in a thumbnail. Check the openings directly.

Where this points next

L2.5 covers bracing the frame you just laid out — how many crossbraces, where, and what physics tells you an underbraced chassis is about to do.

Next up: L2.5 — Bracing strategy

Objective

Derive why an underbraced chassis twists under drivetrain torque, place crossbraces and triangle braces where they actually resist that torque, and avoid the overbracing trap of adding weight that no longer helps.

Concept

A drivetrain applies force to the floor through its wheels. The floor pushes back, and that reaction force has to go somewhere. It goes into the frame. When the frame is rigid, the reaction force becomes the forward motion of the robot. When the frame is flexible, part of the reaction force becomes flex — the frame distorts, and some of the work the motor does turns into bending metal instead of moving the robot.

Each drive wheel applies a forward (or backward) force at its contact patch. That force times the distance from the contact patch to the opposite side of the drivetrain is a torque applied to the frame. If the crossbraces are rigid and well-distributed, the reaction is spread across the whole frame and the frame stays flat. If there are too few crossbraces — or if the crossbraces are clustered at one end — the far end of the frame twists around the braced end like a diving board bending under load. C-channels under that twist do not stay straight. They bow. Bowed C-channels carry bent axles. Bent axles rub on bearings. Rubbing axles eat current. Current kills motors. This is how a bracing mistake becomes a motor failure two weeks later.

📐 Engineering tip. The torque gets bigger when the gearing gets more aggressive. L2.6 will show you that a gearbox multiplying torque to push harder is also multiplying the reaction torque the frame has to absorb. Whenever you increase torque at the wheels, you must increase the bracing that resists the reaction.

The doctrine

Two well-distributed full crossbraces are usually enough. "Well-distributed" means one towards the front and one towards the back of the drivetrain, not two clustered on one end. The space between them matters as much as the number — two braces a couple of holes apart is mechanically almost the same as one brace.

Triangle bracing for towers

A tower C-channel sticking up off the chassis will flex at its base under any side-load. A triangle brace running from the top of the tower down to a point on the chassis several holes away converts the side-load into compression and tension along the triangle's legs, which metal handles much better than bending. The triangle should be close to isosceles and should connect at least two holes from the bottom of the tower.

The overbracing trap

Once the frame is rigid, a third or fourth crossbrace does not make it more rigid — it just adds weight. Weight costs acceleration, costs motor current, costs battery life. If you cannot explain what load a brace is carrying, it is a candidate for removal.

Guided practice

Take your drivetrain frame. Without motors, wheels, or gears attached, set it on a flat table and press down on one corner with the heel of your hand while the opposite corner rests on a pencil or a thin spacer.

1. Observe how much the frame flexes. A single full crossbrace will flex noticeably. Two well-distributed crossbraces will flex only slightly.
2. Now push one side rail forward and the other side rail backward — a shear load. This is closer to what the drivetrain actually experiences under drive torque.
3. Install the bracing configuration you plan to run. Repeat both tests. If the frame still visibly flexes under either test, add bracing. If it does not, stop.

Now install a tower on the frame. Push the top sideways. Add a triangle brace running from two holes up the tower down to a point four or five holes away on the chassis. Push again. The flex should drop dramatically.

Independent exercise

Push your drivetrain sideways with both hands while watching the opposite end from a low angle. You are looking for frame racking — one rail shifting relative to the other by more than a couple of millimetres. Your target is under 3 mm of racking under hand-pressure side-load. If the frame racks more than that, you are under-braced. If it racks less than a millimetre, you are probably overbraced and carrying weight you do not need.

Common pitfalls

• Clustering two crossbraces on the same end of the drivetrain. Distribution is half the doctrine.
• Going below two full crossbraces without boxing the connections and using structural screw joints.
• Adding a third crossbrace "just in case." Overbracing is weight, and weight costs you in every match.
• Using inner braces as the primary structural member. Inner braces are for mounting; outer crossbraces are for structure.
• Forgetting triangle bracing on towers.
• Building a triangle brace that connects less than two holes from the base of the tower.

Where this points next

L2.6 derives the speed-torque-power trade-off that picks your gearing, and ties it directly back to the bracing doctrine — because more torque means more reaction, and more reaction means more demand on the frame.

Next up: L2.6 — Gear ratios, speed, torque & power

Objective

Derive the trade-off between speed and torque from the fixed power budget of a V5 Smart Motor, calculate the actual linear speed of a drivetrain from gear ratio and wheel circumference, and pick a cartridge and gearing that will not cook a motor under match load.

Concept

A motor has a fixed power budget. That single sentence is the whole of this tutorial. Everything else is consequence.

Speed is how fast a point on the robot moves through space. Torque is the rotational force applied at an axle. Power is the rate at which work happens. A motor that outputs one watt can do one joule of work every second, no more.

Power = Torque × Rotational Speed. Rearrange and you get the design principle that runs a drivetrain: Torque = Power / Rotational Speed. Cut the speed in half and you get twice the torque. Double the speed and you halve the torque. The motor does not care which side of the trade-off you want — it gives you a total power envelope and nothing more.

Gear ratios

A gear ratio is the machine that lets you move along that envelope. When a driving gear with twelve teeth meshes with a driven gear that has thirty-six teeth, the driven gear turns once for every three turns of the driving gear. The driven axle spins at one-third the speed — and carries three times the torque. Count the teeth on the driven gear and divide by the teeth on the driving gear. That number is the torque multiplier and the speed divider.

At the wheel, torque becomes linear push: Force = Torque / Wheel Radius. A fatter wheel turns the same torque into less push on the tiles, and at the same RPM carries the robot faster. Wheel diameter is a gear ratio you cannot change without a trip to the parts bin.

Cartridge colours

Red, green, blue — these are pre-built ratios inside a motor housing, giving the same motor different free speeds and stall torques. Red cartridges are torque-heavy and slow. Green cartridges are the balanced default. Blue cartridges are fast and weak at the wheel. A cartridge is not a performance tier. It is a point on the speed-torque curve.

📐 Engineering tip. Here is where physics stops being abstract and starts breaking robots.

Consequence one — drivetrain linear speed

Take the motor's free speed in RPM, multiply by the external gear ratio (driving over driven), multiply by wheel circumference, convert units. That is the theoretical top speed. Real drivetrains hit about seventy to eighty-five percent of that under load.

Consequence two — chassis twist under high-torque ratios

Gearing for push multiplies the torque reacting against the frame. A well-braced chassis absorbs that reaction. An underbraced chassis flexes, the C-channels bow, and the wheels walk out of plane. The torque you multiplied with your gearbox is the same torque that twists your frame.

Consequence three — friction, current, and dead motors

This is the one that ends seasons. A motor spinning against a load draws current proportional to torque. Cross the motor's thermal limit and the controller cuts power to protect the silicon. A gear ratio that is too aggressive for the weight of the robot puts the motor at elevated current constantly. Add the friction of a bound drivetrain and a gear mesh that is too tight, and the motor does not wait for a pushing match to thermal out. It disconnects in driving skills, in a lead autonomous, on a clean field.

Guided practice

Pick the drivetrain you just built in L2.1 through L2.5. Write down three numbers:

1. Motor free speed for your cartridge, in RPM.
2. External gear ratio, driving-teeth over driven-teeth.
3. Wheel diameter in inches.

Now calculate theoretical free speed in inches per second. RPM × ratio gives wheel RPM. Wheel RPM × wheel circumference (π × diameter) gives inches per minute. Divide by 60 for inches per second.

Build the gearbox. Use a sandwich (L3.1 preview) on both sides of the intermediate axle. Mesh the gears with a thin gap — you should feel the tiniest amount of backlash when you wiggle the driven gear by hand. Spin the drivetrain in the air, motors disconnected, and feel for any point where it stiffens.

Independent exercise

Calculate your theoretical free speed. Put the robot on tiles. Drive it a measured ten feet at full throttle and time it with a stopwatch. Compute actual speed. If the actual number is less than seventy percent of the theoretical number, you have a friction problem or a power problem. Go to L3.4.

Common pitfalls

• Picking a cartridge because other teams picked it, not because the numbers said to.
• Gearing for pushing without bracing for the reaction torque.
• Meshing gears too tight because "no backlash looks professional." Tight gears are high-friction gears.
• Ignoring wheel diameter as a gear ratio.
• Treating free speed as actual speed.

Where this points next

L3.1 begins the screw-theory and friction doctrine that makes sure the power budget you just calculated actually reaches the wheels.

Next up: L3.1 — Sandwich everything load-bearing

Objective

Identify every load-bearing axle and joint on your robot, decide whether it must be sandwiched or may be cantilevered, and build the sandwich correctly so the axle stays straight under load.

Concept

A sandwich is an axle or joint supported on both sides by metal or plastic. A cantilever is the same axle or joint supported on only one side. Under load, a cantilevered axle bends. Under repeated load, it bends more. A bent axle wallows in its bearing, the bearing eats into the C-channel, the gear on the axle walks out of mesh, and everything downstream of that axle inherits the slop.

The doctrine is short. Sandwich anything load-bearing. That is every drivetrain axle, every lift pivot, every intake roller shaft that carries torque, every arm joint that holds weight, every chain or sprocket under tension. If the axle has a force on it that would try to bend it, it needs support on both sides.

🔧 Build tip. The sandwich does not have to be a full C-channel on each side. It can be a C-channel on one side and a cut plastic plate on the other. It can be a C-channel on one side and a standoff with a bearing block on the other. What matters is that both ends of the axle are constrained against the direction of load.

The exceptions are real but narrow. A very short, low-load screw joint — for example, a floating intake preroller that only carries the weight of a light roller — can get away with being cantilevered as long as it is mounted on a bearing. "Short" means a hole or two. "Low-load" means the forces on it are small compared to the structural capacity of the axle. If you are unsure, sandwich it. Sandwiches cost almost nothing. Bent axles cost the season.

Guided practice

Walk around your robot and identify every load-bearing axle and joint. For each one, write down:

1. The location (drivetrain axle, lift pivot, intake roller, etc.).
2. The type of load — rotation under torque, static weight, side impact.
3. The current support — sandwiched on both sides, cantilevered with a bearing, cantilevered without a bearing.

Anything cantilevered without a bearing is a problem. Anything cantilevered with a bearing is a problem unless it is short and low-load. Anything sandwiched is fine.

Now pick one cantilevered high-load axle and fix it. The fix is usually one of: add a second C-channel or plate on the unsupported side, add a standoff-mounted bearing block, or add a cut piece of plastic with a bearing hole.

Independent exercise

Grip the end of every drivetrain axle on your robot and try to wiggle it perpendicular to its length. If the axle moves visibly in its support, it is either cantilevered or sandwiched with a loose bearing. Both are fails. Fix the looser of the two.

Common pitfalls

• Treating a bearing as a substitute for a sandwich. A bearing centres an axle in a single support — it does not hold the axle against bending.
• Sandwiching with a plate that is not rigidly attached to the frame.
• Putting a gear far out past the outer support.
• Declaring a cantilever "short and low-load" when it is not.
• Forgetting that lift pivots are load-bearing.

Where this points next

L3.2 teaches screw theory — how to keep each screw centred in its hole so the sandwiches you just built actually hold the axle where CAD put it.

Next up: L3.2 — Screw theory

Objective

Centre any screw in its hole using shoulder screws, bearings and nut retainers, or cut plastic centring pieces, and explain why off-centre screws propagate misalignment through the robot until the drivetrain eats itself.

Concept

A VRC C-channel hole is larger than the shaft of a standard screw. That clearance is not a bug — it is how the screw gets through the hole. But it also means that every screw has a little room to wander inside its hole, and wherever the screw actually sits when you tighten the nut is the position the joint locks into.

Half a millimetre does not sound like much on one screw. It is a lot across twenty screws. Two drivetrain rails each attached to two crossbraces with four screws per joint means thirty-two screws, and if they are each a half-millimetre off in a random direction, the cumulative misalignment is measured in centimetres. A centimetre of misalignment in a drivetrain is enough to turn a straight drive into a curve, pinch a bearing into its housing, and multiply the friction of every gear mesh.

📐 Engineering tip. The doctrine: every load-bearing screw should be centred in its hole. There are three ways to enforce it.

Shoulder screws

A shoulder screw has a precision-machined cylindrical shoulder between its head and its threads. The shoulder diameter matches the C-channel hole tightly, so when you install the screw the shoulder physically forces the screw into the centre of the hole. This is the strongest and cleanest centring method, and it should be your default for every high-load structural joint.

Bearings and nut retainers

Where a shoulder screw cannot reach, you use the bearing itself as the centring surface. The screw passes through the bearing bore (or through a nut retainer with a precision bore), and the bearing bore forces the screw into position.

Cut plastic centring pieces

In specialty situations, a custom-cut plastic piece with a precision hole can sit between the metal layers and centre the screw. This is the exceptional case.

The "push into the same corner" trick

When none of the above work, you can still force alignment by tightening the two pieces of metal while pushing both of them hard into the same corner of their respective holes. This works only for parallel or perpendicular joints. Use it as a fallback, not a first choice.

Guided practice

Audit your drivetrain frame. For every screw joining two pieces of metal, decide which centring method applies.

1. Structural drivetrain joints — side rail to crossbrace — should be shoulder screws. If they are standard screws, swap them.
2. Drivetrain axle screws — where a screw is doubling as an axle for a wheel or a gear — should be centred by the bearing bore.
3. Any joint where you cannot install a shoulder screw and cannot use a bearing — use the push-into-the-same-corner trick.

After the audit, spin the drivetrain by hand. If the feel improved — smoother rotation, less drag — the audit caught real misalignment.

Independent exercise

Pick a single drivetrain joint that currently uses a standard screw. Measure the distance from the screw shaft to the edge of the hole on both pieces of metal. Replace with a shoulder screw of the correct length. Measure again. Report the before-and-after.

Common pitfalls

• Using a shoulder screw that is too short. The shoulder has to pass through both metal layers.
• Using a shoulder screw that is too long. The head will not seat.
• Assuming a tight nut compensates for a loose screw in its hole.
• Skipping screw theory on "small" joints. Small joints are where the misalignment hides.
• Forgetting to re-check centring after hitting a wall in a practice match.

Where this points next

L3.3 compares screw joints to traditional axles and tells you which joint type to reach for in which situation.

Next up: L3.3 — Screw joints vs. axles

Objective

Decide whether a given joint on your robot should be a screw joint or a traditional axle, build structural screw joints correctly for drivetrains, and know when to reach for a long screw joint made from a sawed-off standoff.

Concept

A traditional axle is a steel or plastic shaft that passes through bearings on either side of a sandwich and carries a gear, wheel, or sprocket in the middle. A screw joint is a screw doing the same job: passing through two pieces of metal, carrying something rotating in the middle, with a nut on the far end. The screw is the shaft.

Screw joints are your friend for short, low-load rotating connections. Compared to an axle, a screw joint does not need to be sandwiched if the joint is short, low-load, and mounted on a bearing — the screw itself provides rigidity that a standalone axle cannot.

For structural applications — drivetrain wheels, lift pivots, anything carrying real torque or weight — you use structural screw joints. A structural screw joint has nuts securing the screw on both C-channels of the drivetrain, not just one. The screw is now tying the two rails together as a load path, and the joint is simultaneously bracing the drivetrain and carrying the wheel.

🔧 Build tip. For longer joints, you can build a long screw joint from a sawed-off standoff. Thread a long screw into one end, and saw off the excess to length. This gives you a custom-length shaft with threaded ends on both sides. Install sprockets, gears, and lock nuts on the assembly before you saw anything — otherwise you cannot get the parts on past the sawn ends.

The decision tree

1. Short, low-load, rotating? Screw joint on a bearing. No sandwich needed.
2. Drive wheel on a drivetrain? Structural screw joint with nuts on both rails.
3. Long span, high load, weight-sensitive? Sawed-off standoff screw joint.
4. Long span, continuous high-load torque, exposed? High-strength axle.
5. Otherwise? Low-strength axle, sandwiched, with bearings.

Guided practice

Walk your drivetrain. For every axle or screw joint, confirm it is the right kind.

1. Drive wheel mounts. Should be structural screw joints. If any drive wheels are on a screw with a nut on only one rail, that is non-structural and needs an extra nut.
2. Idler or free-spinning gears. Can be screw joints or axles. Screw joints are lighter.
3. Lift pivots. Usually sawed-off standoff screw joints or high-strength axles.
4. Roller shafts in mechanisms. Almost always screw joints if short, axles if long.

Build one sawed-off standoff screw joint end-to-end. Pick a span, select a standoff slightly longer, thread a long screw through, install all internal parts, saw to length, dress the cut end with a file.

Independent exercise

Count the non-structural screw joints on your drivetrain wheels. For every one, ask: would making this structural (adding a nut on the second rail) improve drivetrain rigidity without costing buildability? If yes, upgrade. Document the reasoning in your notebook.

Common pitfalls

• Leaving drive wheels on non-structural screw joints.
• Building a sawed-off standoff joint without installing the internal parts first.
• Treating every axle as needing to be high-strength. Most do not; they just need to be sandwiched.
• Using a screw joint for a long, high-load span without the standoff trick.
• Forgetting that a screw joint on a bearing still needs the bearing to be flat and zip-tied.

Where this points next

L3.4 is the physics-and-procedure friction audit that will catch any joint — screw or axle — that is adding drag to your drivetrain.

Next up: L3.4 — The friction audit

Objective

Derive the physics of friction from first principles, run a ten-minute binary-search audit on a drivetrain, find the source of any drag, and explain exactly how excess friction destroys motors mid-match.

Concept

Friction is a force that opposes motion. It shows up anywhere two surfaces press against each other and try to move, and it comes in two flavours. Static friction is the force that resists the start of motion. It has a maximum — once the push exceeds that maximum, the surfaces break free. Kinetic friction is the force that resists motion once the surfaces are sliding. Static friction is always larger than kinetic friction for the same pair of surfaces. That is why a bound drivetrain feels stuck when you start pushing and then lurches forward.

The maximum friction force between two surfaces is given by a short, brutal equation:

F_friction = μ × F_normal

μ (mu) is the coefficient of friction — a dimensionless number that describes how grippy the pair of surfaces is against each other. F_normal is the force pressing the two surfaces perpendicularly together.

📐 Engineering tip. Two consequences fall straight out of that equation. First, every unnecessary source of normal force — a spacer pack crammed too tight, a bent bearing, metal rubbing on metal because a washer is missing — multiplies its own μ by that normal force, and the result is drag you have to overcome every single time the drivetrain turns. Second, because μ_static is larger than μ_kinetic, a drivetrain under high static friction re-pays the static-friction tax on every command.

Good drivetrains maximise μ at the wheels and minimise μ × N everywhere else.

The consequence that ends seasons

A motor drawing current is fighting a load. Current is proportional to torque. Torque, at a drive wheel, is proportional to the force the motor has to apply against the floor plus the force it has to apply against internal friction inside the drivetrain. A clean drivetrain pays almost nothing for internal friction. A fouled drivetrain pays for every missing bearing, every over-tight gear mesh, every chain link bound against a sprocket tooth. The motor draws current for that drag whether or not the robot is being pushed. Current turns into heat. The thermal limiter cuts power. The cartridge brakes. The match is over.

This is the most common reason a well-built robot loses matches. Not design. Not code. Friction.

Friction sources (binary-search order)

1. Bearings. Missing bearings. Cracked bearings. Bearings not zip-tied flat against the C-channel.
2. Spacers. Too many spacers crammed onto an axle push gears and shaft collars into hard contact.
3. Gears. Over-tight mesh is worse than loose. You want the faintest backlash.
4. Sprockets and chains. A slightly loose chain is better than a slightly tight chain. A correctly tensioned chain should deflect a few millimetres under thumb pressure.
5. Washers. Plastic washers between any two pieces of metal that would otherwise rub against each other.

Guided practice

Lift the robot so all four drive wheels are off the tiles. Disconnect the motors if you can. With the motors disconnected and the drivetrain in neutral, grip one drive wheel and spin it by hand.

1. A good drivetrain spins freely for two to three full turns before coasting to a stop on a single push. Spin each wheel in turn. Compare them.
2. If one wheel resists or coasts for half as long, you have found the quadrant of the problem.
3. Remove the chain or disengage the gears feeding that wheel. Spin the wheel again. If it now spins freely, the friction was downstream. If it still resists, the friction is at the wheel itself.
4. Re-engage the power train one section at a time and spin after each step. The section that makes the wheel resist again is the culprit.

Put the robot back on the tiles. Push it with one hand — a gentle push. A friction-clean drivetrain rolls at least a full tile length on a gentle push.

Independent exercise

Print the friction audit checklist below, tape it to your build table, and run it cold on a drivetrain that is not yours. Time yourself. Your target is under ten minutes.

Common pitfalls

• Running the audit with the motors still engaged. Motor inertia and gearing mask small drag sources.
• Comparing a wheel to itself instead of to the other three.
• Fixing the first thing you find and declaring victory. Friction stacks.
• Tightening a chain to "fix skipping." Tight chains are high-friction chains. Fix the geometry instead.
• Skipping the floor test. A drivetrain that spins well in the air but refuses to roll on tiles has a load-path friction source.

Where this points next

L3.5 teaches you to box the C-channels so the structural flex that causes alignment-driven friction does not come back the next time the drivetrain takes a hit.

Next up: L3.5 — Boxing C-channels

Objective

Identify where a drivetrain needs boxing, install boxing correctly using inner metal or screws through bearings, and explain why the flanges of a C-channel bend under hit loads that the web of the channel would shrug off.

Concept

A C-channel is strong along its length and stiff along its web, but the open side — the two flanges — can fold in or splay out when a load hits them edge-on. Under a bump from another robot, under the reaction torque from a gear mesh, under repeated impact, those flanges distort. The distortion is tiny per hit but it is permanent and it accumulates. Over a competition weekend a drivetrain that started square has flanges that now tilt slightly, the bearings no longer sit flat, and the axles are riding on tilted supports. Friction goes up. Current draw goes up. Motors get warm.

Boxing is fitting a second piece of metal inside a C-channel so the flanges cannot bend inward — or attaching a second C-channel along the open side so the overall cross-section becomes a closed box. The closed box is dramatically stiffer against twist and impact than the open C.

Where to always box

1. The ends of every drivetrain C-channel. The ends are where wheels attach, bearings sit, and hits land. This is the single highest-leverage boxing location.
2. Anything exposed to hits. Outer faces of the drivetrain, the leading edge of an intake.
3. All structural connection points. Anywhere a crossbrace or triangle brace connects to a rail.

🔧 Build tip. Boxing through a bearing: you cannot slide a solid piece of metal into a C-channel across a bearing. What you can do is use screws with keps nuts passing through both the bearing and the opposing flange. The screw-and-nut pair pins the flange against the bearing block and stops the flange from splaying. This is one of the three legitimate uses of keps nuts (L3.7).

Guided practice

Box the ends of your drivetrain C-channels. For each end of each rail:

1. Measure the span between the outermost bearing and the end of the channel.
2. Cut a piece of C-channel or L-channel to fit inside the rail, with hole spacing matching the rail.
3. Slide or fit the cut piece into the open side of the rail.
4. Install screws through the rail's web, through the boxing piece, with lock nuts or keps-through-bearing as appropriate.
5. Tighten. The flanges should now be constrained.

Independent exercise

Push the outer flange of an unboxed section of your drivetrain rail with your thumb. You should be able to see it flex slightly. Now do the same on a section you have boxed. The boxed flange should not move at all. If it still flexes, the boxing is not long enough, not tight enough, or not installed where the load is actually arriving.

Common pitfalls

• Boxing an entire rail end-to-end. Excess boxing is weight for no structural benefit.
• Boxing with a bent or bowed piece.
• Skipping the keps-through-bearing trick and leaving bearings in unboxed flanges.
• Using keps nuts outside the boxing-through-bearings use case. Keps are prototype nuts — lock nuts are the competition standard.
• Assuming that screws in a rail constitute boxing. Screws hold things together; they do not stiffen the cross-section.

Where this points next

L3.6 covers plastic choice and when to reach for polycarb or delrin instead of metal.

Next up: L3.6 — Plastic choices

Objective

Pick between polycarbonate and delrin for any plastic part on your robot, know when plastic is the right answer versus metal, and understand that your game manual has a limit on how much plastic you are allowed to use.

Concept

Metal is the default building material — strong, rigid, well-tolerated in the parts list. Plastic is the weight-saving alternative for locations where metal is overkill: sandwiches that only carry alignment loads, non-structural bracing, mounts, funnels, and protective plates where impact is light.

The decision comes from three questions:

1. Is the part load-bearing in a structural sense? If yes, stay with metal.
2. Is the part exposed to impact? If yes and you are going with plastic, use polycarb.
3. Is the part weight-sensitive and in a protected location? If yes, plastic is the right call.

🔧 Build tip. Your game manual imposes a limit on the total amount of plastic a robot may carry. The exact number changes from season to season — check the current manual and do not guess. Budget plastic like any other finite resource.

Guided practice

Walk your robot and identify every part that is currently metal. For each, ask: could this be plastic instead without costing rigidity where rigidity matters?

1. Intake funnels. Almost always plastic.
2. Minor bracing between already-rigid subsystems. Plastic.
3. Sandwich plates on non-drivetrain axles. Plastic is often fine.
4. Drivetrain crossbraces. Metal. Do not replace.
5. Lift pivot plates. Metal unless the load is known small.

For one candidate, make the swap. Cut the plastic piece, install it, and confirm the robot still meets its rigidity check. Track the total plastic on your robot as you go.

Independent exercise

Pick one plastic part currently on your robot and write a three-sentence defence of the material choice — polycarb or delrin — against the two questions above: is it load-bearing, is it exposed to impact. If your defence is "I used delrin because we had delrin sitting around," replace it with polycarb unless you can make an impact-free argument.

Common pitfalls

• Using delrin in an impact zone and watching it crack in the first pushing match.
• Replacing drivetrain bracing with plastic because it saves weight. It saves weight and costs the match.
• Forgetting the game manual plastic limit until inspection.
• Heat-bending polycarb because someone said you had to. Cold-bending in a vice works for angles under about ninety degrees.
• Cold-bending delrin. It cracks.

Where this points next

L3.7 finishes the hardware doctrine with nuts and screws — what to use where and why keps nuts are a prototype-only tool outside of three specific legitimate uses.

Next up: L3.7 — Nuts and screws

Objective

Pick the right nut and screw for every joint on your robot, defend lock nuts as the competition default, and name the three legitimate uses for keps nuts.

Concept

Nuts and screws are not interchangeable. The wrong nut on a critical joint can loosen during a match, and the robot that worked in the pit is no longer the robot on the field.

Nuts

• Lock nuts (nylon-insert) are the competition default. They resist loosening under vibration and impact. Aluminium lock nuts save weight compared to steel.
• Keps nuts are prototype nuts. They have an attached toothed washer that bites into the metal surface. They loosen under vibration over time.
• Plastic nuts are for low-strength, weight-sensitive applications only.

Screws

• Steel screws are the default. Strong, predictable, available in every useful length.
• Aluminium screws save significant weight but reserve them for competition-final weight optimisation on a known-good design.
• Plastic screws are for low-strength applications — mounting bearings, mounting light plastic pieces.

The three legitimate uses of keps nuts

1. Mounting radios with rubber links. The rubber link isolates vibration, and the joint is low-load enough.
2. Drivetrain wheel screw joints. A structural screw joint on a drive wheel is short, the load path is direct, and keps are acceptable. Lock nuts are still better.
3. Boxing through bearings. Using a screw and keps nut to pin a flange against a bearing (L3.5).

🔧 Build tip. Everywhere else on the robot, use lock nuts. Every keps nut you leave on a structural joint is a failure waiting to happen in the match you cannot afford to lose.

Guided practice

Audit every nut on your robot. Walk the robot with a pen and paper and categorise each joint:

1. Lock nuts on structural joints. Correct. Leave alone.
2. Lock nuts on low-load joints. Correct, though plastic nuts would save weight if it matters.
3. Keps nuts on a structural joint. Swap to lock nuts.
4. Keps nuts in a legitimate use. Leave alone and document why in your notebook.
5. Plastic nuts on a structural joint. Swap to lock nuts immediately.

Independent exercise

Count the keps nuts on your robot before the audit. Run the audit. Count the keps nuts after. The after number should equal the number of legitimate uses — which is small.

Common pitfalls

• Leaving keps nuts on a drivetrain crossbrace joint because "they have been fine so far."
• Using aluminium screws early in a season.
• Using plastic screws on drivetrain bearings.
• Mixing nut types randomly across the robot.
• Forgetting that the "three legitimate uses" list is short on purpose.

Where this points next

Tier 4 begins with L4.1 — the mechanism design pattern that uses everything in Tiers 2 and 3 as a reference for what can and cannot be built reliably.

Next up: L4.1 — Mechanism design as a problem-solving pattern

Objective

Apply a single five-step pattern — identify, brainstorm, decide, prototype, refine — to the design of any mechanism on your robot, and explain why every tutorial in this tier follows the same shape.

Concept

Every mechanism on a competition robot is a response to a problem. The mistake new builders make is reaching for a mechanism they have seen before and trying to fit the problem to the mechanism. The mistake advanced builders make is reaching for a mechanism they think is clever and spending weeks proving it is not.

The fix is to treat every mechanism design as a repeatable process.

1. Identify. State the problem in one sentence, specific to the game objects, field elements, and strategic role. "Pick up the game object from the floor and move it into the robot" is a mechanism problem. "We need an intake" is not — it is a solution pretending to be a problem.
2. Brainstorm. Generate at least three mechanisms. Some should be ideas you know you will not pick. The point is to build a set of alternatives.
3. Decide. Use a trade-off grid (L1.3) to score the options. Write the decision down with the reasoning.
4. Prototype. Build the option to throw away (L0.1, L1.1). Every prototype should have a named question it is answering.
5. Refine. Take what the prototype taught you and either refine or go back to step three.

📐 Engineering tip. Every mechanism tutorial in this tier — intake, roller, chain tensioner, pneumatics, lift, tracking wheel — is a specific application of this pattern. The structure is the same. What changes is the physics, the geometry, and the parts list.

Guided practice

Pick a mechanism on your robot that is already in its second or third iteration and run the five steps on paper, retrospectively.

1. Identify. What was the problem? Write it in one sentence.
2. Brainstorm. List every alternative you considered. If the list has only one item, add at least two you could have considered.
3. Decide. Score each option against criteria that mattered. Does the option you actually built still score highest?
4. Prototype. What specific questions did your first prototype answer?
5. Refine. What did the refinement step change?

Independent exercise

Pick a mechanism you have not built yet and run all five steps on paper before doing any CAD or building. Turn in a one-page design brief at the end. This brief is a template you will re-use for every mechanism tutorial in this tier.

Common pitfalls

• Skipping the brainstorm step by going straight from problem to a single option.
• Running the decision step without a trade-off grid.
• Treating the prototype as a rough draft of the final build instead of as an experiment.
• Refining a mechanism that the prototype invalidated.
• Skipping the pattern entirely because the mechanism is "simple."

Where this points next

L4.2 applies the pattern to intake design — rollers, compression, dead zones, and the floating preroller.

Next up: L4.2 — Intake design

Objective

Design a roller-based intake that moves a generic game object from the floor into the robot without dead zones, with correct compression, and with a floating preroller that hinges around the changing angle of the object during pickup.

Concept

An intake is a sequence of rollers that move a game object from outside the robot to inside it. Reliability comes from three things: even roller spacing, correct compression, and a preroller that can follow the object's angle change during pickup.

Roller spacing

Rollers must be close enough together that the game object is always in contact with the next roller before it leaves the current one. The failure mode is a dead zone — a gap between rollers where the game object stalls. Plan roller positions with sketches. Use circles to represent the game object and confirm that the object is always touching at least one roller along the path.

Compression

Compression is how much the rollers squeeze the game object as it passes. Too little and the rollers slip. Too much and they deform the object, bind against the ramp, or stall the motor. The distance between the roller and the ramp should equal the natural dimension of the game object minus a few millimetres of grip margin.

The floating preroller

The preroller is the first roller the game object touches. As the preroller picks up the object, the object rotates — the angle of contact changes. A fixed preroller has to handle every angle at once. The solution is a floating intake — the preroller is mounted on a hinge (a screw joint, L3.3) so it can pivot to follow the object. When the object has been taken in, the preroller returns to its resting position.

🔧 Build tip. Use a hardstop to cap the upper range of the floating preroller. Set the upper limit just slightly below the height of the game object. If the limit is too high, the preroller has no contact with small objects and they roll underneath.

Guided practice

You are designing an intake for a generic game object. Measure the object first.

1. Sketch the roller path. Draw the intake's interior as seen from the side. Place circles along the path representing the game object.
2. Compute compression. For each roller, measure the gap between the roller surface and the opposing surface.
3. Build the main roller sequence. Mount each roller on a sandwiched axle or screw joint (L3.1, L3.3).
4. Build the floating preroller. Mount on a screw-joint hinge. The hinge should allow the preroller to swing up and return under light rubber-band tension or gravity.
5. Install the hardstop. Limit the upward range to just below game-object height.
6. Test with game objects. Feed objects one at a time and observe.

Independent exercise

Your intake must move ten game objects in a row from the floor into the robot without jamming or slipping. If it fails on any of the ten, the test restarts. Count attempts to ten clean passes.

Common pitfalls

• Skipping the sketch step and building by eye.
• Setting compression by feel and declaring it "close enough." Measure the gap.
• Building a fixed preroller and hoping the compression is forgiving enough.
• Setting the preroller hardstop too high.
• Mounting the preroller hinge on a non-bearing joint.
• Assuming the intake is done after one successful pickup. Run the ten-in-a-row test.

Where this points next

L4.3 picks the roller type — flex wheel or rubber-band-on-sprocket — for the intake you just designed.

Next up: L4.3 — Roller types and selection

Objective

Pick between flex wheels and rubber-band-on-sprocket rollers for any intake or transport mechanism, choose the right flex-wheel durometer, and decide whether a rubber-band roller needs a mesh cover.

Concept

Flex wheels are solid elastomer wheels that compress slightly under load. They come in different durometers — 30A (softest), 45A (medium), and 60A (hardest). A 30A flex wheel squishes easily against a game object and grips well even at imperfect compression. A 60A barely deforms and holds a consistent roller diameter. Flex wheels take up less space than rubber-band rollers and are the right choice when alignment matters.

Rubber bands on sprockets wrap a circle of rubber bands around a sprocket, using the bands as the contact surface. They are lighter and very compressible — the bands stretch freely when an oversized object comes through. Two variants exist: bare bands (maximum compressibility, exposed axle) and mesh-covered (slightly stiffer, protected, preferred by default).

🔧 Build tip. Both roller types lose grip when dirty. Tile dust, rubber debris, and bits of game object accumulate on roller surfaces. Clean rollers are high-grip rollers. This is a match-day habit.

Guided practice

Take the intake you built in L4.2. For each roller in the sequence, decide which type is right and document the decision. Swap the rollers and re-test with the ten-in-a-row test.

Independent exercise

For one roller, change only the durometer (flex wheel) or the band count (rubber-band roller) and re-run the ten-in-a-row test. Record the before and after.

Common pitfalls

• Defaulting to 60A flex wheels because they look "high-performance." 60A is a specialty roller. 45A is the default.
• Using bare-band rollers in exposed locations.
• Cleaning rollers only when they visibly fail. Clean on a schedule.
• Assuming a soft flex wheel will fix a compression problem. It will mask it.
• Mixing roller types arbitrarily without a reason.

Where this points next

L4.4 solves the chain-tensioning problem for intakes where the driving sprocket is not co-located with the hinge.

Next up: L4.4 — Chain tensioners

Objective

Identify when a floating mechanism needs a chain tensioner, build either a banded tensioner arm or a four-bar tensioner, and decide when double-chaining is sufficient insurance against chain skip.

Concept

A chain drive connects two sprockets at a fixed distance apart. When that distance is fixed, the chain stays at whatever tension you set. When one sprocket is on a floating mechanism that pivots relative to the other, the distance between sprockets changes as the mechanism moves, and the chain tension changes with it. A loose chain skips teeth; a skipped chain is a mechanism that stops working mid-match.

There is a special case where this is not a problem: when the driving sprocket is mounted exactly at the hinge of the floating mechanism. In that geometry, the centre of the drive sprocket and the pivot of the arm are the same point, and the chain tension is constant. Pick this geometry whenever possible.

Option 1: Banded tensioner arm

A small arm hinged on the frame, with an idler sprocket at its end, and a rubber band pulling the arm against the chain. As the chain goes slack, the band pulls the tensioner further into the chain. The key constraint: band tension must be just enough to keep the chain taut when the mechanism is lifted — not so much that it adds friction when the mechanism is down.

Option 2: Four-bar tensioner

A four-bar linkage that keeps the tensioner arm at a constant angle as the mechanism moves. Bulkier but more consistent — the tension is determined by geometry rather than band behaviour.

Option 3: Double-chaining

If the mechanism has a small range of motion (under thirty degrees of pivot), two chains running the same path on duplicate sprockets provide belt-and-braces insurance. Default to double-chaining on any mechanism that is even slightly suspicious.

🔧 Build tip. Even with a tensioner, add a second parallel chain as insurance. Two chains, two independent failure modes, much longer mean time to skip.

Guided practice

Pick a floating mechanism on your robot.

1. Measure the chain length difference across the range of motion.
2. If the difference is small, double-chain and move on.
3. If significant, pick between banded (space-tight, moderate range) and four-bar (large range, consistent tension).
4. Build the tensioner.
5. Tune band tension.
6. Double-chain anyway.

Independent exercise

Cycle the floating mechanism through its full range twenty times at full speed. No chain skip, no visible slack beyond what the tensioner is designed to absorb.

Common pitfalls

• Ignoring the tension change because the mechanism "does not move that much."
• Over-tightening a band tensioner. The motor feels fine in the workshop and thermals out in the match.
• Building a banded tensioner with insufficient range.
• Skipping double-chaining because the tensioner "should handle it."
• Using a four-bar tensioner in a space too tight for its geometry.

Where this points next

L4.5 covers pneumatic actuator geometry — stroke length, leverage, and the distances that decide whether a piston can move the arm you want it to move.

Next up: L4.5 — Pneumatics geometry

Objective

Lay out a pneumatic piston that moves a pivoting arm through the range of motion you want, with the leverage you need, using stroke length, distance a, and distance b as your design levers.

Concept

A pneumatic cylinder is a piston inside a sealed tube. Air pressure pushes the piston through a fixed stroke length. On a VRC robot the most common use is to rotate a pivoting arm through a fixed angular range.

The geometry is a triangle. One vertex is the pivot of the arm. A second is the anchor point where the base of the piston is mounted to the frame. The third is the rod-end attachment on the arm. The piston's stroke is one side of the triangle — its length changes as the piston extends.

Three relationships drive every pneumatic layout

Stroke length and range of motion. Longer stroke = more angular range for a given a and b.

Distance a and leverage. a is the distance from the pivot to the rod-end attachment on the arm. A small a means more range but less leverage (less torque at the arm). Halving a halves the torque.

Distance b and arm angle. b sets the starting and ending angles of the arm's motion. Adjust b to place the motion in the angular window you want.

📐 Engineering tip. A piston's effective leverage depends on the angle between the piston and the arm at each instant. When the piston is perpendicular to the arm, all of its force becomes torque — maximum leverage. When the piston is close to parallel, most of its force acts along the arm — minimum leverage. Design the geometry so the piston is as close to perpendicular as possible across the range where the mechanism does work against a load.

Guided practice

1. State the range of motion. From what starting angle to what ending angle?
2. Pick a stroke length. Use whatever cylinders are available.
3. Sketch the triangle at both extremes. Check the swept angle matches the requirement.
4. Check the leverage angles. If either extreme has the piston nearly parallel to the arm, move the anchor point.
5. Build the layout. Mount the pivot first. Mount the anchor second. Mount the rod-end attachment last.
6. Test the motion. Actuate the piston and watch the arm's travel.

Independent exercise

Build a pneumatic arm. Measure the actual start and end angles with a protractor. Compare to the target range. If off by more than a few degrees, the a and b values did not match the stroke length — redo the maths and rebuild.

Common pitfalls

• Picking a too large because "more leverage is better." Large a means small angular range.
• Picking a too small because "more range is better." Small a means tiny moment arm.
• Ignoring the piston-to-arm angle at the extremes.
• Adjusting b by moving the pivot instead of the anchor.
• Building before sketching. Pneumatics are geometry problems; solve them on paper first.

Where this points next

L4.6 surveys lift mechanisms — four-bar, six-bar, DR4B, lady brown — and tells you when to pick each.

Next up: L4.6 — Lift mechanisms

Objective

Pick between four-bar, six-bar, double-reverse four-bar, and single-joint arm lift mechanisms for a given lifting task, and defend the choice against the alternatives.

Concept

A lift moves a payload from one height to another. Four families cover most VRC designs.

Four-bar

Two parallel arms connected by a rigid end plate, pivoting around two fixed points on the chassis. The payload stays at the same orientation throughout the lift. Simple, reliable, easy to tune. The default lift for payloads that need to stay horizontal and do not need extreme height.

Six-bar

A four-bar with additional linked arms extending the payload outward as it rises. Greater horizontal reach at the top. The cost is more joints, more alignment, more places for slop.

Double-reverse four-bar (DR4B)

Two four-bars stacked, with the second inverted so the payload rises vertically. Gives pure vertical motion. The most complex of the four families. Use a DR4B only when the vertical motion is strategically required.

Single-joint arm (lady-brown variant)

A single arm pivoting around one point. The simplest lift — one joint, one motor group, one moment arm. Wins when the payload does not need to stay at a constant orientation.

Picking a lift

1. Does the payload need to stay at a constant orientation? If yes → four-bar, six-bar, or DR4B.
2. Does the payload need pure vertical motion? If yes → DR4B.
3. Does the payload need to reach further than a four-bar can sweep? If yes → six-bar.

The tie-breaker is build complexity. A simpler lift is a more reliable lift.

🔧 Build tip. Every pivot on a lift is load-bearing. Every pivot is sandwiched (L3.1). Every pivot uses a structural screw joint or a sandwiched axle (L3.3). Every arm is braced against twisting (L2.5). Lifts carry compounded torque reactions — under-built lifts sag, and sagging lifts lose alignment every cycle.

Guided practice

Pick a lifting task on your robot. Run the three questions to pick a family. Then:

1. Sketch the lift geometry. Side view. Both start and end positions.
2. Check the swept arc for collisions.
3. Size the arm members for rigidity under payload weight.
4. Plan the pivots — all sandwiched, all structural.
5. Plan the power transmission. Reference L2.6 for the torque calculation.
6. Build the frame of the lift first (L2.1 applies to subsystems too).

Independent exercise

Build a four-bar lift for a test payload. Measure lift time, maximum lift angle, and end-plate deflection under load. The end plate should be horizontal at rest and should not visibly deflect.

Common pitfalls

• Picking a DR4B because it looks advanced.
• Cantilevering lift pivots. Lifts have high pivot loads and cantilevered pivots sag within one match.
• Under-bracing the arm members.
• Under-sizing the motor torque for the payload.
• Building the lift without a frame-first approach.
• Forgetting that the end plate carries the payload's full weight at the end of a long moment arm.

Where this points next

L4.7 is the structural build procedure for tracking wheels — the odometry pods that Chapter II L4.4 assumes are mounted correctly on the robot.

Next up: L4.7 — Tracking wheel mounting

Objective

Build a tracking wheel assembly — the physical hardware that the position-tracking code in Chapter II assumes — with the correct wheel size, the right mounting material, a downward motion limit, and enough down-force to keep the wheel in contact with the tiles.

Concept

A tracking wheel is a small free-spinning omni wheel mounted on the robot purely to measure distance travelled. It does not drive the robot; it rolls along the tiles under the robot's motion and a rotation sensor on its axle reports how far it has rotated. The code turns that rotation into a linear distance and combines two or three tracking wheels into a full position estimate.

Wheel size

Use 2.75-inch wheels. Smaller wheels (2-inch omnis) can rotate freely when the robot moves perpendicular to the wheel's direction — a false rotation that feeds phantom motion into the position estimate. A 2.75-inch wheel has bigger rollers and better geometry and reports cleaner data.

Mounting material

Plastic (polycarb) mounting is preferred. The tracking-wheel assembly is not structural in the drivetrain sense, and the weight savings from plastic are real.

Downward motion limit

Limit the downward motion with a hardstop so the wheel cannot flip over when the robot goes over a field element.

Down-force for contact

Band the tracking wheel assembly downward with a rubber band — one per side — so there is constant light force pressing the wheel onto the tiles. Strong enough to recover contact after a bump, light enough not to add noticeable rolling resistance.

📐 Engineering tip. The "screw joint odom pod" designs you may see online are not a special category. They are axles in precise mounts. Understand the terminology so you are not fooled into thinking there is a secret technique.

Guided practice

Build a single tracking wheel assembly end-to-end.

1. Select the wheel. 2.75-inch omni. Confirm it spins freely.
2. Cut the mounting plates. Two pieces of polycarb with axle holes and mounting holes.
3. Install bearings. One per plate, in the axle hole. Zip-tied flat.
4. Install the axle. Through bearing, through wheel, through bearing. Shaft collars to retain.
5. Mount the assembly to the frame. The mount should allow slight upward and downward pivot.
6. Install the hardstop. Position it to block the wheel from dropping below neutral.
7. Install the down-force band. One rubber band per side.
8. Mount the rotary sensor. On the axle, with no slop in the coupling.
9. Verify rotation. Push the robot on the tiles by hand. The sensor reading should update smoothly.

Independent exercise

Push the robot in a straight line for one full tile. Record the sensor reading. Convert to distance using the wheel circumference. The measured distance should match one tile length to within a few millimetres.

Common pitfalls

• Using 2-inch omnis and watching the position estimate drift sideways during hard turns.
• Skipping the down-force band and losing wheel contact over bumps.
• Skipping the downward motion limit and watching the wheel flip over.
• Mounting the tracking wheel in metal for no structural benefit.
• Coupling the rotary sensor through a joint with any slop. Slop is pure noise in the position estimate.
• Assuming the wheel is frictionless. Run the spin-by-hand test (L3.4) on the tracking wheel axle.

Where this points next

Tier 5 begins with L5.1 — electronics placement as a structural decision, including where the IMU has to sit for the sensor reading in Chapter II to work.

Next up: L5.1 — Electronics placement

Objective

Place every piece of electronics on your robot — battery, V5 Brain, radios, IMU, pneumatic reservoirs — with explicit structural reasoning, and defend the centre of gravity that placement produces.

Concept

Electronics are not decoration. Every component has mass, and every mass shifts the centre of gravity. Where you put the heavy components decides whether the robot tips on aggressive turns, whether the drivetrain gets enough weight on each wheel for traction, and whether the IMU can read the robot's orientation cleanly.

Battery

Mount it low — as close to the tiles as the frame allows — to drop the centre of gravity and reduce tipping. Mount it accessible but protected: accessible so you can swap batteries at queue without removing half the robot, protected so it does not take a direct hit.

Brain

Same logic — low, accessible, protected. Place it where the wire runs are short to the majority of motors and sensors.

Radios

Two radios, always. If one disconnects mid-match the robot has dead time while the radio re-links. Mount them high, away from motors, because motors generate electrical noise that can interfere with radio link quality. The correct mounting method is rubber links — one of the three legitimate uses of keps nuts (L3.7).

IMU

Mount it low, on a crossbrace, as close to a dead-level flat surface as you can achieve. Vibration from the drivetrain and lift motors is the IMU's worst enemy. Flat, low, crossbrace, minimal vibration.

Pneumatic reservoirs

Mount them low, near the battery. Do not mount a heavy reservoir high on the robot.

⚡ Competition tip. The electronics placement is the last chance to fix centre of gravity cheaply. Moving a battery a few inches is easier than relocating a lift tower. Do the maths while you still have time.

Guided practice

Walk your robot and make an electronics placement report.

1. Battery. Location, height above tiles, retention method, accessibility rating (can you swap in under 10 seconds?), protection rating.
2. Brain. Same four points.
3. Radios. Are there two? How high? Distance from nearest motor? Rubber links?
4. IMU. How flat? How low? How far from the nearest vibrating motor?
5. Pneumatic reservoirs. If applicable — how low?

Compute the centre of gravity by inspection. Set the robot on a narrow pivot and see which way it leans. Shift electronics to correct the error.

Independent exercise

Swap your battery in under ten seconds. Start a stopwatch, unplug, remove, insert, plug in, secure retention. If you cannot do it in ten seconds, the battery placement is not accessible enough.

Common pitfalls

• Mounting the IMU on a mechanism that vibrates.
• Running a single radio.
• Mounting radios near motors.
• Mounting the battery inaccessible.
• Mounting pneumatic reservoirs high.
• Not checking centre of gravity at all.

Where this points next

L5.2 turns the wiring those electronics need into a set of habits that make the robot serviceable instead of a rat's nest.

Next up: L5.2 — Wire management

Objective

Route every wire on your robot with custom-cut lengths, organised paths, and strain relief, so the robot is serviceable at queue and passes inspection without a fight.

Concept

A wire mess is not cosmetic. It is a time penalty at every service visit, a failure point on every maintenance cycle, and a disqualification risk at inspection.

Custom wire lengths

A stock wire is almost always too long. Long excess wire has to be coiled somewhere, and the coil lives in the middle of the robot. Invest in a wire-cutting and crimping kit and cut every wire to the length it actually needs, plus a small service margin.

Organised routing

Every wire has a path. The path runs along structural members, not across the middle of the robot. It crosses motors and gears at right angles. Wires of similar destination group together.

Strain relief

Secure every wire close to both ends with a zip tie or a clip so that any pull is absorbed by the zip tie rather than by the crimp. Especially important on wires that cross a moving joint.

⚡ Competition tip. Three specific failure modes make wire management match-critical. (1) Inspection fails on tangled or exposed wiring. (2) Mid-competition motor swaps take twice as long on a messy loom. (3) Intermittent shorts from wires rubbing against sharp metal edges cause "something electrical" failures that no one can diagnose quickly.

Guided practice

Re-wire one subsystem from scratch.

1. Remove all existing wires. Save the connectors; cut the wires.
2. Plan the routing on paper. Sketch the subsystem and draw the wire paths. Mark moving joints and sharp edges.
3. Cut wires to length. Measure the planned path, add a small service margin, cut, crimp.
4. Install wires along the planned paths. Bundle with light zip ties every few inches.
5. Strain-relief every connector. A zip tie within a centimetre of each connector.
6. Test every connection. Power on, check every motor responds, check every sensor reads.

Independent exercise

Take a photo of your robot's wiring before you start re-wiring, and another after. Time yourself swapping a motor on the before-version and on the after-version. The time difference is the real return on the investment.

Common pitfalls

• Using stock-length wires and coiling the excess.
• Running wires across the interior instead of along structural members.
• Skipping strain relief because the crimps feel tight.
• Bundling power and signal wires tightly together. Power wires can induce noise in signal wires.
• Waiting until inspection to fix wiring.

Where this points next

L5.3 is the pre-competition audit — the final checklist that uses every tutorial in this strand to confirm the robot is actually ready.

Next up: L5.3 — The pre-competition build audit

Objective

Run a pre-competition build audit on any robot in under an hour, catch every structural, friction, wiring, and electronics issue that would cost matches, and produce a printable checklist that travels with the team.

Concept

A pre-competition audit is the last defence between a broken robot and a ruined weekend. It is not a style review. It is a mechanical check that looks for specific, known failure modes — the ones that every strand in this curriculum has been warning you about — and confirms that none of them are present on the robot you are about to put on the field.

⚡ Competition tip. Run it the night before the event, not the morning of. If the audit finds problems, you need time to fix them. Run with a witness — one person runs, one person checks off.

The six audit categories

1. Friction. Lift the robot, spin every drive wheel by hand, confirm clean rolling and equal coast distances (L3.4). Push across a tile with one finger.
2. Bracing. Push the frame sideways with both hands. Under 3 mm of racking (L2.5). Inspect every crossbrace joint.
3. Wires. Strain relief at every connector. No exposed conductor. No wires snagging a moving mechanism. Pull-test every connector.
4. Battery and brain. Battery accessible in under 10 seconds. Brain cover in place. Port assignments match the code.
5. Subsystem range of motion. Cycle every mechanism through its full range. Watch for binding, chain skip, wire snag, or interference.
6. Sensors. IMU flat and low. Tracking wheels in contact and reporting. All other sensors powered on.

A fail on any category stops the audit. Fix the fail, restart the category, and continue.

Guided practice

Run the audit on your robot tonight. Work with a teammate. One runs the checks, one ticks the boxes. Do not skip steps because "we did that last week."

Independent exercise

Print the checklist. Run it cold. Time yourself. Your target is under one hour for a full pass on a robot you know well.

Common pitfalls

• Running the audit the morning of the event. Too late to fix anything.
• Running the audit solo. The witness is the point.
• Skipping a category because the robot "felt fine yesterday."
• Treating a near-fail as a pass. Marginal items become failures at the worst possible moment.
• Not writing down the results. If you do not record the audit, you cannot track degradation.

Where this points next

The audit is the exit criterion for this strand. A robot that passes this audit is ready to receive the programming work from Chapter II and the strategic work from Chapter V without structural surprises getting in the way.

Coming up next: Chapter IV — CAD — sketches-as-planning, part studios, assemblies with motion, and the CAD-to-build handoff.