5S on a fabrication floor: what it changes in practice

A press tool comes back from a 250 T press at 14:20 and goes into the wrong bay of the die rack, because two tools for the same part family are labelled by drawing number only. The next setter finds it at 06:10. Nothing is broken and no part is scrap, and the line still loses most of a shift because a die sits 12 m from where the setup sheet says it lives.

5S on a fabrication floor is judged by three measurements: how long it takes to find a tool, how often material of the wrong grade reaches a machine, and how quickly anyone can tell that a station has drifted from its standard condition. Tidiness is what the photographs show. The measurements are what the customer pays for.

The plant, quality and sustainability hub collects the rest of this cluster, from incoming inspection to the effluent treatment plant.

What 5S on a fabrication floor actually changes

It changes search time, material segregation and the time it takes to see a deviation, and only the first of those can be timed with a stopwatch on the day the racks are rebuilt.

Every missing shim, every unlabelled clamp and every offcut in the wrong bin adds a walk. A setter who leaves a 250 T press to fetch a 0.2 mm shim from the tool crib walks about 28 m each way. At a normal pace of 1.2 m/s that is 47 seconds of walking, and the journey is never only walking: the crib is searched, the size is checked against the setup sheet, and the trip is often repeated for a second item. Four to six minutes for one retrieval is realistic.

Now multiply it. A cell with six tool changes a shift and one missing item per change loses 24 to 36 minutes a shift to search, which is 72 to 108 minutes a day across three shifts. A shadow board at the press, holding the six shim sizes, two clamp types and the die-height gauge the setup sheet names, removes the walk and leaves about 30 seconds to take the shim from the board. That is the entire case for the second S, and it can be checked in an afternoon.

The arithmetic applies to the laser as well. A 12 kW fibre laser with a 40 ft x 16 ft bed holds a sheet 12.2 m by 4.9 m, or 59.8 m² of material, and a 300 mm offcut left on the skeleton is invisible from the console. It becomes a part cut from the wrong grade at the next load, or a sliver inside a coated enclosure three weeks later.

One thing needs saying plainly, because suppliers and buyers both get it wrong: 5S is not a standard and cannot be certified. There is no clause number to quote and no certificate to file. What a buyer can test is the outcome the clauses describe. ISO 9001:2015 clause 7.1.4 requires the environment for the operation of processes to be suitable and maintained, and clause 8.5.1 requires production to run under controlled conditions. ISO 45001:2018 clause 8.1.2 requires hazards to be eliminated where possible and risks reduced, which is what a clear walkway and a guarded, uncluttered press station evidence. ISO 14001:2026 clause 8.1 asks for operational control of processes, and a segregated scrap bin is one of the cheapest examples of it.

A shop that treats 5S as an initiative will run it for four months. A shop that treats it as the physical evidence behind clauses 7.1.4, 8.5.1 and 8.1 gives it an owner and a review date, the same as a calibration schedule.

The five elements, translated to a sheet metal shop

The generic version of 5S, written for offices, is unusable on a fabrication floor. The version below is the one that answers to a press shop, a laser cell and a coating line, with the target that makes each element auditable rather than aspirational.

Read the table as a set of failure modes rather than virtues. Each row exists because something specific goes wrong without it: a tool that cannot be found, a shim taken from the wrong kit, a stainless offcut welding into a mild steel frame, or a fastener rattling inside a finished cabinet.

The failure that costs the most is a mixed bin

Scrap segregation looks like a sustainability measure and functions as a quality control. A 50 mm offcut of 304 stainless dropped into the CRCA bin is not a recycling problem; it is a piece of the wrong alloy sitting within reach of a setter who needs a spacer, and stainless does not weld to mild steel with a MIG torch set for CRCA. The part that gets the spacer is discovered at the weld cell, or worse, at the customer's corrosion test after the coating has cured.

The same logic runs the other way. Aluminium swarf on a steel floor, in the presence of moisture, produces a galvanic couple that pits the steel, and a wet paint line is a wet environment. Copper and brass are the worst offenders because they are visually similar to nothing else in the shop until they are small.

Three controls make segregation real rather than stated:

• One bin per alloy, labelled at the mouth at eye height, not on the side where a full bin hides it.
• A colour code that matches the material tag on the coil or sheet, so the bin and the material agree.
• A rule that an offcut under 300 mm goes in the bin at the machine, not into a drawer, because the drawer is where mixed material starts.

For coated work, foreign object control is the second half of the same discipline. Enclosures and rack panels are checked for loose fasteners, film, tape and metal slivers before they enter pretreatment, because the seven-tank line will not remove a sliver and the powder coat will seal it in place. Powder coating covers a defect rather than revealing it, which is why the check has to happen upstream.

The record, the frequency and the failure it prevents

A 5S audit that produces a score and no action is a survey. The record that changes behaviour has seven fields, and the last two carry the weight.

1. Area and date, with the shift the audit was run on.
2. Element scores, one to five each, with a note where a score is below three.
3. The finding, as a photograph with a location, not a description.
4. The owner, by name or role, not by department.
5. The due date for the fix.
6. The close-out date and the evidence that it was closed.
7. Whether the same finding has appeared before, which is the field that exposes a system problem rather than a tidy-up.

Five elements scored one to five gives 25 points, and a working target is 20 or better. The target is the least useful number in the record. An area that scores 22 with a two on foreign object control is more dangerous than one that scores 19 with fours across the board, because the first has a specific hole and the average hides it. Weight the elements that map to a customer-visible failure, and treat any element below three as an action regardless of the total.

Frequency matters more than depth. A weekly ten-minute walk by the area owner catches a rack that has filled with unlabelled tools within days. A monthly scored audit gives a trend that survives a difficult launch month, when the temptation to leave a die on the floor is highest. A quarterly audit is a photograph of a quarter, and the drift it was meant to catch has usually shipped before the next one.

The failure this record prevents is quiet. Nothing dramatic happens when a floor decays; setup times creep up, one part in a thousand is cut from the wrong grade, and the quotation that was built on a 22-minute changeover stops being true. Six months later the shop is uncompetitive on a part it has made for years, and nobody can point to the week it changed. A trend line can.

What Ganesh runs, and why the floor layout is not cosmetic

Ganesh has been manufacturing since 1992 and now runs five units across Mumbai (Vasai) and Nagpur, with more than 200 people. The equipment set is what makes housekeeping a production question rather than a presentation one: mechanical presses from 20 T to 300 T, a 250 T punching line, a 12 kW fibre laser with a 40 ft x 16 ft bed, 6 kW and 3 kW support lasers, a 200 T five-axis brake with a 4,450 mm bed that forms to 10 mm, three CLOOS robotic MIG cells and more than 30 manual MIG/MAG, spot and SPM stations.

That mix puts a lot of hard tooling on the floor. Press tools, progressive dies, welding fixtures and checking fixtures are designed and built in the in-house tool room, which means a die is not a consumable ordered from outside; it is an asset with a location, a maintenance history and a home bay. The tools and dies capability describes the tool room, and the same storage discipline applies to fixtures at the welding cells, where a clamp pad worn by 0.5 mm moves a part more than any weld parameter will.

Two further details matter to a buyer reading a 5S claim. First, coated work passes through seven-tank pretreatment, a 16 ft x 16 ft shot blast chamber and an 18 ft x 14 ft powder booth, so metal slivers and loose fasteners have several chances to become a coating defect. Second, the works hold ISO 9001:2015 and ISO 14001:2015 certification and operate an effluent treatment plant, which puts chemical storage, waste segregation and spill control inside the same management system as production, rather than in a separate environmental folder.

What we do not publish is a 5S score. Scores are internal, they differ by area, and a number in a brochure proves nothing. Ask for the last six months of audit records and the close-out dates during a plant visit, and pick two findings to trace to the floor. That is a twenty-minute test and it tells you more than any certificate.

What to put in the area standard

An area standard is one page per zone: the laser cell, each press group, the brake, the weld cells, the coating line. It works when it names objects rather than principles.

• The exact tools, gauges and fixtures that belong in the zone, by number, with their storage position.
• The cleaning point in the cycle, for example "bed cleared and checked at every sheet load" or "fixture wiped at every tool change".
• The alloy segregation rule for that zone, naming the bins and their colour code.
• The foreign object check and who signs it before the part leaves the zone.
• The audit frequency for that zone and the name of the owner.
• The escalation rule: what happens when a finding is open past its due date.
• The review date for the standard itself, because a zone changes when a new part is introduced.

The line that most standards miss is the trigger. A 5S standard written for the parts running in March is out of date by June, and the audit then measures compliance with a document nobody uses. Tie the review to the engineering change that introduces a new tool or fixture, and the standard stays true.

When 5S is the wrong answer

Housekeeping cannot fix a capability problem, and it is worth being blunt about the boundary. If a hole is drifting 0.15 mm across a run, the cause is die wear, punch grind or press condition, and a clean floor will not change it; the die maintenance routine is the control that does. If a first article is being measured with an uncalibrated caliper, a beautifully labelled gauge board makes the error more organised, not smaller.

There is a layout version of the same mistake. If a cell is 40 m from the tool crib because the plant was extended twice, no amount of labelling removes the walk. That is a routing decision, and the honest fix is to move the storage or duplicate the kit at the machine.

And 5S does not substitute for the inspection plan. A floor can score 24 out of 25 and still ship a bracket with the wrong hole position, because tidiness says nothing about whether the first-off was measured. The inspection routine for repeat press work covers what the audit should be sampling alongside the housekeeping record.

Send the drawing, the annual quantity and the finish specification, and the engineering team will return the route, the tooling position and the inspection points before any price is discussed. A quotation request that names the part family and the volume is enough to start.