Records, logs and working with inspectors

Every Friday afternoon Ray spent an hour at the Northcrest shop with his route's records: the week's logs, the callback reports, the open repair items and any letters or reports from Technical Safety BC. In June Hugh asked Tamsin to sit in, and then to take over the hour for a month while Ray watched. She found it harder than any test. “The equipment tells you what is wrong with it,” Ray said. “The records tell you what is wrong with your maintenance.” This lesson covers the records a route keeps, how callback logs and test records are analysed to improve maintenance, and how to work with the inspectors and safety officers who read those records.

The records on a route

Lesson 2 listed the records a maintenance control program must keep: maintenance records, repair and replacement records, callback logs, oil usage records and periodic test records, all for at least five years. On a route these exist in two places. On site, the primary record is kept where the mechanic and the inspector can see it, as Technical Safety BC's directive requires, with instructions posted at the controller. At the contractor's office, the electronic system holds the same information in a form that can be searched, sorted and analysed across many units. The two must agree, and keeping them in step is a weekly job. A log that says a task was done on site, while the system shows it outstanding, or the reverse, is an audit finding even if the work was done.

Good records are specific. “Adjusted doors” tells the next mechanic nothing; “car door closing time at floor 3 was 2.9 s against a plate minimum of 3.2 s, operator speed reset, now 3.4 s” tells them what was found, what was done and what the result was. Ray's rule was that a record should let a stranger repeat the work and check the result.

Callback logs

Each callback is recorded with the date and time of the call, who called and what they reported, when the mechanic arrived, what was found, the cause, the corrective action, when the unit returned to service, and whether anyone was trapped. Over a year, those records show which units fail most, which components fail most and how long units are out of service. Two simple measures are used across the industry: callbacks per unit per year, which compares units and routes, and mean time between callbacks, which is the running time divided by the number of callbacks. The causes matter more than the counts, because the causes tell the mechanic what to change in the program.

# Route 7 callback analysis for the last 12 months
from collections import Counter
calls = [  # (unit, cause category, entrapment?)
    ("Seaview 1", "doors", False), ("Seaview 1", "doors", False), ("Seaview 2", "doors", False),
    ("Seaview 2", "doors", False), ("Seaview 1", "doors", False), ("Seaview 2", "vandalism", False),
    ("Tillicum 1", "controller", False), ("Tillicum 1", "controller", True), ("Tillicum 1", "doors", False),
    ("Tillicum 1", "power", False), ("Birchwood 2", "doors", True), ("Birchwood 1", "valve", False),
    ("Pemberton 1", "water in pit", False), ("Pemberton 1", "doors", False), ("Holloway 1", "doors", False),
    ("Holloway 2", "valve", False), ("Marlowe 1", "power unit", False), ("Larchmont 1", "belt monitor", False),
    ("Larchmont 1", "doors", True), ("Dunmore 1", "doors", False), ("Ashgrove 1", "doors", False),
]
UNITS = 14
HOURS_IN_SERVICE = 8760            # hours per unit per year, running or available
print(f"callbacks: {len(calls)}, per unit per year: {len(calls) / UNITS:.1f}")
print(f"mean time between callbacks per unit: {UNITS * HOURS_IN_SERVICE / len(calls):,.0f} hours")
print(f"entrapments: {sum(1 for c in calls if c[2])}")
by_cause = Counter(c[1] for c in calls)
print("by cause:", ", ".join(f"{k} {v}" for k, v in by_cause.most_common()))
by_unit = Counter(c[0] for c in calls)
print("worst units:", ", ".join(f"{k} {v}" for k, v in by_unit.most_common(3)))
share = by_cause["doors"] / len(calls)
print(f"doors are {100 * share:.0f} % of callbacks: review door tasks first")

callbacks: 21, per unit per year: 1.5
mean time between callbacks per unit: 5,840 hours
entrapments: 3
by cause: doors 12, controller 2, valve 2, vandalism 1, power 1, water in pit 1, power unit 1, belt monitor 1
worst units: Tillicum 1 4, Seaview 1 3, Seaview 2 3
doors are 57 % of callbacks: review door tasks first

The result surprised no one who had worked in the trade: doors caused most of the callbacks, as they do on almost every route, and Seaview's heavily used cars and Tillicum's old car caused the most. Seaview's door tasks had already been shortened after lesson 2's review, and the callbacks since then had dropped. Tillicum car 1's controller faults would end with the modernization. The two door entrapments at Birchwood and Larchmont had different causes, a broken hall door roller and a cracked gib, but both were in hall door hardware, so Tamsin proposed adding an examination of hall door rollers and gibs at every landing every three months on all the route's cars. Hugh approved it, and the program review was recorded.

Test records and trends

Periodic test records are kept so that each result can be compared with the acceptance baseline and with earlier tests, as lessons 4, 7 and 9 showed for wear, governors and escalator brakes. A test record that cannot be compared, because it uses different units, a different method or no values, is almost useless. When Tamsin went through Route 7's five-year test records she found that the Pemberton's last safety test had recorded the stopping distance in inches and the one before it in centimetres, without saying so on the sheet; it took a phone call to the mechanic who had retired to sort out which was which. The program's test procedures were amended to state the units on every form.

Deficiencies and orders

Inspectors record what they find. Depending on the jurisdiction and the seriousness, a finding may be a deficiency to be corrected by a date, an order requiring action, or a notice that the equipment must be removed from service. Technical Safety BC's safety officers issue inspection reports and, where needed, compliance orders, and the contractor and owner must respond by the date given; Technical Safety BC also requires contractors to report each year on the devices they maintain. Whatever the form, the contractor tracks every open deficiency to its closure, with the date, the action and the evidence, because an uncorrected finding at the next inspection is a much more serious matter than the original one.

# Open deficiency tracker from the last round of inspections, as of 30 June 2027
from datetime import date
TODAY = date(2027, 6, 30)
items = [  # (unit, finding, due, closed on, evidence)
    ("Ashgrove 1", "emergency light battery not dated", date(2027, 7, 8), date(2027, 6, 18), "photo and note sent"),
    ("Seaview 2", "pit light guard missing", date(2027, 6, 15), None, None),
    ("Marlowe 1", "relief valve seal broken", date(2027, 5, 20), date(2027, 4, 22), "resealed, test record"),
    ("Birchwood 2", "hall door gap floor 4 over limit", date(2027, 7, 30), None, None),
    ("Pemberton 1", "machine room storage", date(2027, 6, 1), date(2027, 5, 28), "photo"),
]
for unit, finding, due, closed, evidence in items:
    if closed:
        state = f"closed {closed}" + (" (late)" if closed > due else "") + f", evidence: {evidence}"
    else:
        state = "OVERDUE" if TODAY > due else f"open, {(due - TODAY).days} days left"
    print(f"{unit:12s} {finding:36s} due {due}  {state}")
late = [i for i in items if i[3] is None and TODAY > i[2]]
print(f"overdue and open: {len(late)} - tell the owner and the authority today, with a date for correction")

Ashgrove 1   emergency light battery not dated    due 2027-07-08  closed 2027-06-18, evidence: photo and note sent
Seaview 2    pit light guard missing              due 2027-06-15  OVERDUE
Marlowe 1    relief valve seal broken             due 2027-05-20  closed 2027-04-22, evidence: resealed, test record
Birchwood 2  hall door gap floor 4 over limit     due 2027-07-30  open, 30 days left
Pemberton 1  machine room storage                 due 2027-06-01  closed 2027-05-28, evidence: photo
overdue and open: 1 - tell the owner and the authority today, with a date for correction

Seaview car 2's missing pit light guard had been ordered, but the part had not arrived and nobody had told the inspector. Hugh's rule was that a correction that will be late is reported before the due date, not after, with the reason and a new date; inspectors almost always accept a late correction that is explained in advance, and almost never accept silence. Tamsin called the inspector's office that afternoon.

Working with inspectors

Inspectors and safety officers are trained specialists, and in many jurisdictions they are certified under ASME QEI-1, the standard for qualified elevator inspectors, or an equivalent program, which requires experience in the trade and an examination. They have seen far more elevators than most mechanics, and most of them want the same thing the mechanic wants: equipment that is safe and records that are true. Ray's habits with inspectors were simple. Be on time, with the equipment ready and the records open. Answer the question asked, honestly, and say “I don't know, I will find out” rather than guess. Show the reading rather than describe it. Never hide a problem, because inspectors find most of them and remember the ones that were hidden. And keep the relationship professional in both directions: an inspector is not a customer to be pleased, and not an opponent to be beaten.

Inspectors also make mistakes, and a mechanic who believes a finding is wrong should say so politely, with the clause and the edition, and ask for the decision in writing if the inspector maintains it. As ELV-410 lesson 3 explained, interpretations can be requested formally from the authority, and a written interpretation is part of the record. What a mechanic must never do is ignore a finding because they disagree with it.

Records as evidence

Records are also evidence. After an incident, a fall, an entrapment injury or a death, the maintenance and test records are among the first things investigators ask for, along with the program and the callback log. They show whether the equipment was maintained as the program required, whether known problems were acted on, and whether tests were done and passed. Records written at the time, specifically and honestly, protect the mechanic who did the work properly. Records written afterwards, altered or vague, do the opposite. Hugh put it plainly to every new mechanic: “Write every entry as if a coroner will read it, because one day one might.”