Seeing Past a Stopgap to the Underlying Problem

Why this matters

A stopgap does not just hide a fault, it corrupts your data. Every reading you take, every symptom you observe, is the reading of a system with a band-aid on it, not the reading of the real fault. If you diagnose what you measure at face value, the stopgap leads you by the nose. The skill here is mental subtraction: picturing what the system would be doing if the band-aid were not there, so you chase the fault instead of the compensation. This is different from spotting a stopgap. It is what you do once you have found one.

The stopgap moves your baseline

Every diagnosis is a comparison against "normal," and a stopgap redefines normal. A start-assist makes a weak compressor start like a healthy one, so your start-up reading looks fine. A booster pump makes low supply pressure read as adequate at the fixture. A dumped dose of chemicals makes bad pool water test in range for a day. In each case the instrument tells the truth about the compensated system and lies about the fault. Until you account for the stopgap, your baseline is fiction.

Ask what it compensates for, then reverse it

The fastest route past a stopgap is to name what it does and invert it:

• A device that adds pressure implies the real pressure is low. The fault is upstream of the booster.
• A device that adds starting torque implies the thing being started is weak. The fault is the motor or its capacitor, not the assist.
• A dose that suppresses a symptom implies the symptom is being generated faster than normal. The fault is whatever generates it.

Every compensation is an arrow pointing at its cause, in the opposite direction. Follow the arrow backward.

Your instruments are reading a compensated system

Take your normal readings, but annotate each one: is this value real, or is the stopgap propping it up?

• A pressure, temperature, or flow that reads normal only because a rig is active is not a passing reading. It is a masked failing reading.
• Where you can, measure on both sides of the stopgap. The difference across it tells you how hard it is working to hide the fault, which tells you roughly how bad the fault is, though a partially failed stopgap can understate the gap, so treat the differential as a floor on the fault, not an exact measure.

Two ways a stopgap fools you

• Masking hides a fault that is present: the clamp holds the leak, so you see no water. Absence of the symptom is not absence of the fault.
• Mimicking creates a symptom that looks like a different fault: a bypass jumper makes the control behave as if a sensor failed, sending you to replace a good sensor. The stopgap generates a false lead.

Knowing which one you face changes the hunt. Masking means look harder for the hidden fault. Mimicking means distrust the obvious lead.

Remove and remeasure, on purpose

Sometimes the only way to read the real system is to take the band-aid off, deliberately and safely, and measure again.

• If the stopgap was holding back pressure, a leak, or stored energy, relieve it and control the release before you remove it. Pulling a cap off a pressurized line to "see what happens" is how you get hurt.
• Remove, remeasure, and now you have the true reading of the fault, which is the reading you should have had all along.

The mental model

Diagnose the system, not the band-aid on it. Before you trust any reading, ask: would this number look this way if the rig were gone? If the answer is no, you are reading the compensation, and the fault is still out there, exactly as big as it was before someone hid it.

References

• Trade-standard practice: establish true baseline readings, accounting for prior repairs
• Measurement discipline: differential readings across a compensating device
• See related: The Temporary Fix Someone Left in Place; A Prior Patch Is Masking the Symptom (decision tree)