Patch the Symptom or Fix the Design: A Decision Tree
Why this matters
Some systems were never right. They were undersized, wired wrong, piped with the wrong slope, or built for a load they never actually carry, and no amount of component diagnosis will ever "find" the fault, because nothing broke. A tech who treats a design flaw like a fault chases parts forever and eats the callbacks. The skill is telling the two apart fast, then making an honest call: patch the symptom to live with the design, or correct the design so the symptom stops for good.
Start here: did it ever work right?
One question splits this whole problem. Ask the customer, and check the evidence: has this system ever performed the way it should, or has it always done this?
- Used to work, now does not: something changed. That is a fault. Go find what changed.
- Never worked right from day one: nothing changed because nothing was ever correct. That is a design or original-install problem, and the fix is not a part.
Do not skip this. Half of the "impossible" diagnostics on the board are techs hunting a fault in a system that never had one.
Safety gate: does the design defect create a hazard?
Before any patch-versus-correct logic, one branch overrides everything. If the flawed design endangers people, you do not get to patch it and walk away. Undersized conductors that run hot, a fuel appliance with no combustion air or a missing vent, missing backflow protection on a potable water line, a pool bonding grid that was never installed, a pressure or structural part carrying more than it was built for: these fail toward danger. De-energize, shut off, or take the system out of service and make it safe, then tell the customer plainly that it cannot run as built. There is no legitimate band-aid on a safety-critical design defect.
If it used to work: treat it as a fault
A system that performed correctly and then degraded has a real fault to find. Something wore, loosened, leaked, drifted, or got fouled. Run your normal what-changed diagnosis: read the failed part for the cause, check what was touched or added recently, and follow the symptom to its origin. The design is not on trial here; a component is.
If it never worked right: read the design
Signs you are looking at a design problem, not a fault:
- The complaint is chronic and load-related: cannot keep up on the hottest or coldest day, runs out under peak demand, trips only when everything is on at once.
- Every component tests within spec. Nothing is broken. The system is simply asked to do more than it was built for.
- The symptom fits a sizing or configuration error: short-cycling from an oversized unit, a drain that never drained because the fall is wrong, a circuit that nuisance-trips because it is loaded past its rating, a pump that cavitates because the suction line is too small.
Confirm by checking the design against the actual load or code: the sizing calculation, the pipe diameter, the wire gauge versus the connected load, the vent versus the appliance input. If the design does not match the load, you found it, and no replacement part will change that.
The patch-versus-correct decision
| What you found | Patch the symptom | Correct the design |
|---|---|---|
| Hazardous design defect | Never | Only real option, or take out of service |
| Chronic non-hazard, correction is a big job the customer must defer | Legitimate stopgap, fully disclosed | Recommend, schedule when they are ready |
| Symptom recurs a few times a season, correction is modest | Poor value, you are renting a fix | Do it once |
| A one-time peak the system will rarely see again | Reasonable to live with | Optional |
A patch is fair when the correction is genuinely out of reach today and the patched state is stable and safe. It stops being fair the moment you let the customer believe the patch was the fix.
Recap
- Ask whether it ever worked right. Used-to-work is a fault; never-worked is a design problem.
- If the design defect is a hazard, correct it or take the system out of service. No patch.
- If it is a comfort or reliability design flaw, weigh the patch treadmill against the one-time correction and let the informed customer choose.
- Document that it is a design limit, not a broken part, so the next tech does not re-diagnose a fault that was never there.
References
- Trade-standard root-cause practice on distinguishing original-design deficiencies from in-service faults
- Manufacturer sizing and application guidance (equipment matched to the actual load)
- Applicable installation codes for conductor, vent, drain, and backflow sizing
- See related: The Band-Aid Repair Versus Correcting the Root Design; Root Cause vs Symptom Fix (Decision Tree)