How to Decide When a Tool Is Worn Out

Why this matters

Most shops retire a tool the day it stops working, which means every retirement lands in the middle of a job. The tool that is worn out but still running is the expensive one. It turns a forty minute task into an hour, it strips a fastener that becomes a callback, and because there is no single day it went bad, nobody ever replaces it. A worn tool also lies about the job: your labor numbers get worse and the tool never appears in the explanation. Deciding on evidence moves the call to a Tuesday morning in the shop, where it costs a purchase order and nothing else.

Step 1: Sort the four ways a tool leaves service

These are not the same decision and mixing them is why shops argue about tools instead of replacing them.

• Unsafe. The tool has a defect that can hurt someone. This is immediate and not a budget conversation.
• Worn out. It still performs the function, slower or less accurately, and the degradation is measurable.
• Broken. It stopped. That is a repair-or-replace call, covered by its own article, and it is reactive by nature.
• Obsolete. It works fine but no longer matches the work, the fastener standard, the battery platform, or the code requirement.

This article is only about the second one. If you cannot say which of the four you are in, you will apply a cost test to a safety problem, which is the worst outcome on this list.

Step 2: Pull anything that fails a safety gate before you measure anything

Wear becomes a hazard on a short list of tools, and on those the answer is removal from service, not a stopwatch.

• A ladder with a bent rail, a cracked or split side rail, a loose or missing rung, or a damaged foot gets tagged and taken out of service on the spot. Under 29 CFR 1910.23(b) (general industry; the construction equivalent is 1926.1053(b)(16), so use whichever Part governs the work), when a ladder is found to have a structural or other defect it must be immediately tagged "Dangerous: Do Not Use" or with similar language and removed from service until repaired or replaced. Tag it at the truck, do not lean it back in the rack, and do not let it ride to the next job "just for reaching the attic."
• A test instrument you use to prove a circuit is dead is a safety device, and a worn one is a safety device that lies. Before you use a voltage tester to establish an electrically safe work condition, verify it on a known live source, test the conductor, then verify it again on the known live source. That live-dead-live sequence is what NFPA 70E requires precisely because a failed tester reads zero on an energized conductor. A tester that fails either verification is out of service immediately, not at the end of the week.
• A corded tool with a damaged cord, a missing ground pin, a cracked housing, or a guard that no longer returns is unplugged and tagged before the next use. Do not test a cord fault by wiggling it while the tool is plugged in.
• Never retire a tool while it is still energized or still holding load. Take the battery out, unplug it, bleed pressure from an air tool, let a cutting head come to a full stop, and release any spring or hydraulic preload before the tool goes into the retirement bin. Tools get pulled out of that bin by curious people.

Everything past this step assumes the tool is safe and the question is only whether it is still worth using.

Step 3: Pick one measurable wear signal per tool class

You cannot measure everything, and you do not need to. Each tool class has one signal that moves first and moves clearly.

The last row is the exception that proves the rule: on anything that restrains, lifts, or carries current, condition is the signal and speed is irrelevant. A sling with a cut does not get a stopwatch.

The reason to use the tech's own baseline rather than a factory figure is that the baseline absorbs everything you cannot control: the material you actually work in, the technique of the person holding it, the altitude and the temperature. A published rate compares your tool to somebody else's job. Your own first-year number compares your tool to itself.

Step 4: Write the retirement trigger with all three parts

A trigger that is missing any of these three parts will not survive its first argument.

• The unit of analysis. Not "it takes longer." Say: benchmark task time, measured by one tech, on the same task and material, taken as the median of five timed instances. Five, because a single slow run is a bad bit or a bad morning.
• The Boolean, if there is a second gate. State whether both conditions must hold or either one triggers. "Median benchmark time at or above 1.3 times the tool's own baseline, OR any safety gate failed" is a different rule from the same sentence with AND, and the AND version would let a cracked housing ride because the tool was still fast.
• The step size, when the trigger changes a number. If crossing the trigger moves the tool from front-line to backup rather than to the scrap bin, say so, and say what backup means: available for a second crew on the same job, not issued as a primary.

A workable default trigger for a shop of three to fifteen: a tool is worn out when its median benchmark time on five timed instances reaches 1.3 times its own first-year baseline, OR when it fails any safety gate in step 2. Tune the 1.3 later. Commit to it now, because a shop without a number retires tools by whoever complains loudest.

Step 5: Run the trigger on a real tool

A rotary hammer, shared across two techs, used for anchor sets.

Baseline. In its first year, the anchor set on a standard job type ran 0.6 hours. That number came off five jobs of the same type, timed by the tech who does most of them.

Now. The same tech, same job type, same anchor size and same bit spec, times five more instances. The median is 1.0 hours.

Apply the rule. 1.0 divided by 0.6 is 1.67, so the tool is running at 1.67 times its own baseline. The trigger is 1.3. It is crossed, and it is not close.

Check the confounder before you buy anything. The bit is the consumable, and a worn bit produces exactly this symptom. Fit a new bit of the same spec and re-time the full five instances the trigger calls for, not a quick three: the confounder check has to clear the same bar as the reading it is challenging, or you have replaced a measured number with a hunch. Say the median comes back at 0.8 hours. That is 1.33 times baseline, still above the 1.3 trigger, but now the picture is different: the bit was carrying about 0.2 of the 0.4 hours of loss, which is half of the degradation. The tool itself is responsible for 0.8 against 0.6, and the honest read is that the tool is marginal rather than dead. If instead the new bit had brought it back to 0.65 hours, you would have found a consumable problem and a tool with plenty of life, and condemning it would have been a wasted purchase plus a bit you still had not replaced.

Size the loss so the decision is not a matter of opinion. Take the confirmed figure with a fresh bit, 0.8 hours against a 0.6 baseline, which is 0.2 hours lost per anchor set. If the crew runs three of these job types a week, that is 0.6 hours a week. Across a 48 week working year that is about 29 hours, and every one of those hours is technician time you paid for and did not sell.

The call. At 1.33 times baseline with a fresh consumable, the tool has crossed the trigger and gets replaced, but the replacement goes in the next scheduled purchase rather than the same afternoon. The worn tool is demoted to backup, meaning it goes on the shelf for a second crew that needs a hammer on the same job, and it never gets issued as somebody's primary. Demotion is the honest disposition here because the tool works, just slowly. If the same tool had failed a safety gate, demotion would not have been available and it would have gone straight out.

Step 6: Handle the disposition so the tool does not come back

A retired tool that stays in the building comes back, usually on the worst day, usually in the hands of the newest tech. Three dispositions and nothing else:

• Scrap. Remove the battery or cut the cord before it goes in the bin. A tool that looks fine gets rescued out of a bin. One that visibly cannot be powered does not.
• Backup. Mark it, put it in a designated backup location, and record that it is a backup in whatever list you keep. A backup tool with no marking is a primary tool with a bad reputation.
• Sell or trade. Fine for a working tool. Never for one that failed a safety gate, and never for a test instrument whose accuracy you could not confirm. Passing on a lying meter is a different kind of problem than passing on a slow drill.

How to verify you got this right

Six weeks after the replacement, re-time the benchmark task five times with the new tool and the same tech. You are looking for the median to land at or below the original baseline. If it comes in at 0.6 hours, the tool was the cause and your trigger works. If it comes in at 0.9 hours, the tool was not the cause and you have a technique, material, or fastener-spec problem you just spent money to not fix. That second outcome is the one worth catching, because a shop that never re-times will replace the same tool class every two years and never notice the real driver.

The second check is on the trigger itself. After a year, look at how many tools it caught and how many were replaced outside it. If most replacements happened outside the trigger, the number is wrong or nobody is timing anything. Lower the multiple before you conclude the method failed.

References

• 29 CFR 1910.23, OSHA ladder requirements including defective-ladder tagging and removal from service
• NFPA 70E, Standard for Electrical Safety in the Workplace, test instrument verification before and after absence-of-voltage testing
• Manufacturer documentation for service intervals, consumable specifications, and rated duty cycles
• See related: Broken Tool Replace vs Repair Decision Tree, Equipment Replacement Cycles, The Real Cost of a Cheap Tool Over Its Life