Unit 08 · lesson
Reliability Wins Matches Nobody Remembers
A robot cannot score from the pit.
Reliability is not glamorous, which is one reason teams underinvest in it. But competition rewards machines that boot, connect, move, survive impacts, recover quickly, and return to the field.
Think in failure modes
For every major subsystem, ask:
How can this fail, how would we notice, and how quickly can we recover?
Example:
| Subsystem | Failure mode | Detection | Prevention | Recovery |
|---|---|---|---|---|
| intake | roller chain slips | motor runs, roller does not | alignment/tension checks | retension/replace |
| CAN | connector loosens | multiple device faults | retention + routing | reseat/repair |
| elevator | sensor loses zero | position becomes implausible | homing/check | safe re-home |
| software | wrong build deployed | behavior mismatch | release/version process | redeploy known-good |
This is a lightweight failure-mode analysis.
Pre-match checks are not superstition
A pit checklist should be short enough to use and specific enough to matter.
Possible categories:
- battery installed and secured;
- main connections inspected;
- critical fasteners/retainers checked;
- CAN/network/device status normal;
- mechanisms move freely;
- sensors show plausible values;
- correct code version deployed;
- driver controls verified;
- pneumatics/stored energy handled correctly if used;
- bumpers/inspection-sensitive items ready.
The current rules decide exact legal requirements. Your checklist supports readiness.
Triage under time pressure
Suppose you have eight minutes before the next match and the elevator stops.
The pit does not need a philosophical discussion. It needs a bounded diagnostic flow.
- Confirm symptom.
- Decide whether the issue is mechanical, electrical, software, or uncertain.
- Check the highest-information observations first.
- Choose a repair with a known rollback.
- Run a controlled functional test.
- Record what changed.
If the repair cannot be validated safely, the strategy may need to adapt around a disabled subsystem.
Spares and serviceability
Reliability includes the ability to repair.
A design with inaccessible fasteners, permanent wire tangles, unlabeled controllers, or no spare critical part creates longer downtime.
Ask during design:
- Can this component be reached?
- Can it be replaced without removing three unrelated systems?
- Are connectors identifiable?
- Is there enough cable service?
- Do we have the tool and spare part?
Create your reliability card
For three subsystems, identify:
- top failure mode;
- earliest detectable symptom;
- preventive check;
- recovery action;
- estimated recovery complexity;
- evidence required before returning to play.
Finish with a known-good baseline list: the minimum checks that tell you the robot is ready to leave the pit.
Unit evidence
Combine your game strategy, scouting decision, and reliability card into a one-page match brief.
A strong brief does not assume the robot is perfect. It explains what the alliance can rely on, what is uncertain, and how the plan changes if a subsystem degrades.
That is competition engineering.