A Sewing Machine Maintenance Schedule a Supervisor Can Actually Run

Every factory I have visited has a mechanic who is genuinely good at his job and genuinely overloaded, because he spends his entire day running to machines that have already stopped. That is breakdown maintenance, and it is what most CMT factories including mine have defaulted to at some point. It feels efficient because nothing is being done to machines that are working. It is not efficient, because the bill arrives in a form nobody adds up: an operator sitting idle for forty minutes, a line unbalanced for half a shift, and a week of skipped stitches that got blamed on the girl at station nine.

This article is the schedule I would hand to a supervisor, written so it can be executed by people who have a line to run and no interest in a maintenance philosophy.

Start With What Breakdown Maintenance Actually Costs

The cost has three parts and only the first one is visible.

The visible part is the repair — a part, an hour of the mechanic's time. The second part is lost line hours: the operator waiting, and, on a balanced line, everyone downstream of that operation slowing down too. The third and largest part is quality. A machine drifting out of tension or timing does not stop. It keeps running and keeps producing garments that a checker will reject later, and by the time the DHU number moves, a few hundred pieces have gone through.

That third cost is why preventive maintenance is a quality programme as much as a machine programme, and it is the argument that gets an owner to fund it when "the machines are running fine" is the counter-argument.

The Four-Level Schedule

Treat manufacturer instructions for your specific models as authoritative, especially on oiling, since oil type and points vary and over-oiling a machine that does not need it is its own defect source. What follows is the general shape most CMT floors use.

Daily — operator, 5 minutes at start and end of shift

• Clear lint from the throat plate area, bobbin case and feed dog with a brush. Not compressed air blown into the machine, which pushes lint deeper into the hook area and into the motor.
• Check the oil sight glass where the machine has one, and report a low level rather than topping it up from whatever tin is nearest.
• Wipe the machine bed and the table. A clean bed is how you notice an oil leak on the day it starts.
• Check the needle: bent, blunt or burred needles come out now, not at break time.
• Sew a test piece on the actual fabric and look at the stitch before the first bundle of the day.
• Report anything abnormal — noise, heat, vibration, thread breaking more than usual — to the supervisor at that moment.

Weekly — mechanic, roughly 10 minutes per machine

• Open the throat plate and bobbin area, clean thoroughly, check for lint packed under the feed dog.
• Oil according to the manual's weekly points. Check the oil level and the return.
• Check belt tension and condition. A glazed or cracked belt is replaced now, not when it snaps mid-shift.
• Check thread path for burrs: tension discs, thread guides, take-up lever, needle plate hole. A burr anywhere on that path causes thread breaks that get blamed on thread quality.
• Check presser foot pressure and feed dog height against the fabric being run.
• Sew a sample and inspect the stitch on both sides: balanced tension, no skipping, no puckering.

Monthly — mechanic, 30 to 45 minutes per machine

• Strip the throat plate, feed dog and hook area. Clean and inspect the hook point for wear.
• Check needle-to-hook clearance and hook timing. This is the adjustment that most often explains persistent skipped stitches.
• Inspect bobbin case for scoring; replace rather than polish where scoring is deep.
• Check and clean the motor or servo unit and its ventilation. Lint accumulation around a motor is a fire risk as well as a cooling problem.
• Check electrical: cable condition, plug, earthing, switch function, foot pedal.
• Check the table, stand, and machine mounting for looseness. Vibration destroys settings faster than use does.

Quarterly — mechanic, half a day per machine, scheduled off-line

• Full strip-down clean, including the oil pan and filter where fitted, and an oil change if the manual calls for one.
• Complete timing verification: needle bar height, hook timing, feed timing.
• Replace consumable wear parts on a planned basis rather than on failure — bobbin cases, tension springs, worn presser feet.
• Motor or servo inspection: brushes where applicable, bearings, speed calibration.
• Verify the machine against a sample of each fabric type it is expected to run, and record the settings that worked.

The Needle Policy Buyers Audit

Needle change and broken-needle control are two different things, and factories that export need both in writing.

Change policy. Needles wear continuously and produce skipped stitches and fabric damage well before they break, so most factories change on a schedule rather than on failure. A workable policy: a fixed change interval measured in machine-running hours, plus a mandatory change on any change of fabric type, on visible damage, and after any needle strike. Needle manufacturers publish selection and condition guidance for different fabrics and applications — Groz-Beckert's sewing product documentation is a reasonable starting reference — but the interval that suits your fabric mix is found by setting one, watching what the checkers find, and adjusting it.

Broken-needle control. This is the one auditors ask for, because a needle fragment inside a garment is a safety recall. The standard shape of the control:

• New needles are issued only against the return of the old one, one for one, from a single controlled store point.
• A broken needle must be recovered in all fragments and taped to a needle log sheet, dated, machine number recorded, signed by operator and supervisor.
• If a fragment cannot be found, the machine, the work in process and the surrounding area are searched, and the garment in process is quarantined or destroyed rather than passed.
• Broken needle logs are retained and produced on request.
• The same control extends to other sharp tools: scissors, trimmers, pins.

Sharp-tool and needle control is a routine line item in factory audits. It appears in the WRAP principles under health and safety expectations, and comes up in the workplace-standards sections of audits such as SMETA. Individual buyers publish their own detailed requirements on top of those, and their version is the one you should build the policy against. The point for a supervisor is simpler: the log has to be real, contemporaneous and complete, because a log that was written up the night before an audit is obvious to anyone who has read a few.

Who Does What

The line that matters most is the one about adjustment. Operators adjusting their own tension is how a floor ends up with every machine set differently and nobody able to say what changed. The rule should be that operators report and mechanics adjust, and the rule only holds if the mechanic actually arrives quickly. If the response time is forty minutes, operators will adjust the machine themselves and they will be right to.

Telling a Bad Machine From a Bad Operator

The single most useful maintenance question a supervisor can answer is which machine is generating defects, and it is normally answered wrongly, because the person standing at the machine is easier to blame than the machine.

The method is comparison across people. If your production records identify both the operator and the machine for each piece of work, then:

• A defect that follows the machine across three different operators is a mechanical problem. Send the mechanic.
• A defect that follows the operator across three different machines is a training problem. Send twenty minutes of instruction.
• A defect that appears only for one operator on one machine is usually a fit problem — machine speed, seat height, or a setting that suits someone else's way of working.

That comparison needs the pairing to exist in the record. In our own system every scan records the operator, the machine, the operation and the timestamp together, which turns the question into a query rather than a debate at month-end; any system that tracks work at machine level supports the same reasoning, and a tally sheet naming only the operator makes the machine invisible by construction. The corresponding defect side of the record is covered in our note on what floor QC software needs to do — the pairing only pays off if the defect carries a type, so that "skipped stitch on machine 47" is a sentence the data can produce.

Machine History Without a CMMS

You do not need software to start. You need one card per machine that never leaves the machine, and a rule that nothing gets done to the machine without a line on the card.

Each line carries the date, what was done, by whom, which parts were replaced, and how long the machine was down. That is all. Six months of those cards tells you which machines eat your mechanic's time, which model in your fleet is a false economy, and which machines are due for replacement rather than another repair. It also gives you something to show an auditor, which a mechanic's memory does not. Keep the cards at the machine, not in an office file: a filed card gets written up weekly from memory. When the cards outgrow themselves, that is the signal to look at software.

When a CMMS Is Worth It

A CMMS is a computerised maintenance management system: assets, preventive-maintenance schedules that generate their own reminders, work orders, spare-parts inventory, and cost history per machine. It is a real category with real products, and at a few hundred machines it earns its fee, mostly by making the schedule self-driving rather than dependent on somebody remembering.

Two honest notes. First, a CMMS will not fix a floor where the mechanic has no time; it will document the backlog precisely. Buy the discipline before the software. Second, and to be explicit about my own product: Scan ERP is not a CMMS. It has no maintenance scheduling, no work orders and no parts inventory. What it contributes is the attribution described above — production and defect records that carry the machine, so you can tell which machines deserve the maintenance budget. The work orders belong in a different tool. I have written that argument out at length in CMMS for garment factories, including the realistic options by factory size.

If you take one thing from this article, take the daily five minutes. A brush in every operator's drawer and a supervisor who checks that it has been used costs nothing and removes a surprising share of the mechanic's queue.

Frequently Asked Questions

How often should a sewing machine be serviced?

The common practice in CMT factories is a four-level schedule rather than a single service interval: cleaning and lint removal daily by the operator, a mechanic check weekly covering oil level, belt tension and stitch quality, a deeper monthly service with the throat plate and hook area opened up, and a quarterly strip-down including timing, motor or servo, and wiring. Machine manufacturers publish their own intervals for each model and those should take precedence over any general rule, particularly for oiling.

How often should sewing needles be changed?

Most factories set a fixed change interval rather than waiting for a needle to break, because a worn needle produces skipped stitches and fabric damage long before it fails. A common policy is a scheduled change measured in machine-running hours, with an immediate change on any change of fabric type, on visible damage, and after any needle strike. The exact interval depends on fabric and needle type, so the workable approach is to set an interval, watch the defect record, and adjust it rather than copy a number from another factory.

What is a broken needle policy and why do buyers audit it?

It is a written control that accounts for every needle in the building so that no needle fragment reaches a garment. The usual shape is that new needles are issued only against the return of the old one, a broken needle must be recovered in all its pieces and taped to a log sheet, the machine and any garment in process are searched before work resumes, and the log is signed and kept. Buyers audit it because a needle fragment in a garment is a safety recall, and sharp-tool control is a standard line item in factory social and quality audits.

How can you tell whether defects come from the machine or the operator?

By comparing across people. If your records identify the machine as well as the operator for each piece of work, then a machine that produces the same defect for three different operators is a machine problem, and an operator who produces that defect on three different machines is a training problem. Without that pairing you are guessing, and the usual guess blames the operator because the operator is the person standing there. Our own scans record operator, machine, operation and timestamp together, which makes the comparison a query rather than an argument.

A maintenance schedule fails for one of two reasons: it was written by somebody who does not have to run it, or it was never checked. Print the daily list, tape it to the wall at the end of each line, and walk the line at start of shift for two weeks. The schedule that survives that fortnight is the one you have.

Santosh Rijal is the founder of Scan ERP, a garment manufacturing ERP system designed for factory floor operations. He works directly with sewing lines, cutting rooms, and production supervisors across Nepal's garment manufacturing sector.