Lessons · Lesson 2 of 6
- 01 · Why one ply of cloth is the hard part
- 02 · What automates, and how to tell before you ask for a price
- 03 · The payback that has to survive a change of style
- 04 · The gain the rating does not give you
- 05 · What it does to the people, and what the people cost
- 06 · Year three, and the part that stops it
What automates, and how to tell before you ask for a price
Work out the automatable operations from the limp-cloth principle rather than from a list, score a real garment against it, and read the honest ceiling that falls out.
Lesson 2 of 6 · 19 min
Three questions, and they come from lesson 1
Lesson 1 ended with a prediction. If cloth is hard to handle because it has no pose, then the operations that automate should be the ones where something is holding the cloth still. Turn that into three questions you can ask of any operation, in this order, before you ask anybody for a price.
- Is the material held in a known shape while the operation happens? Under tension on a table, squeezed flat under vacuum, clamped in a frame, folded over a former, blown onto a body form, or gripped at both ends. If the answer is yes, the cloth has been made rigid for a while, and lesson 1's whole problem has been paid for by a fixture instead of by a robot.
- Does the operation have to follow a moving edge all the way, or does it line up once at the start? Lining up once is a single measurement at the start. Following all the way means the machine has to keep finding an edge that is moving, changing shape and being pulled by the feed while it works.
- Is the thing that moves a ply, or a package? A bundle, a hanger, a carton, a roll and a trolley all have handles. A ply does not.
An operation that answers yes, once, package is a candidate. An operation that answers no, all the way, ply is not a candidate at any price. No amount of engineering budget changes that. It is the shape of the problem, not the size of it.
Scoring one garment
AK-1180 has a standard minute value of 24.06, spread over eighteen sewing operations. Aydemir put every one of them through the three questions.
| Operation | Standard minutes | Held still? | Registers once? | Candidate |
|---|---|---|---|---|
| Crease four patch pockets | 1.68 | over a former | once | yes |
| Position and mark four pockets | 2.64 | on a flat panel | once | yes |
| Topstitch four patch pockets to fronts | 2.24 | in a clamp | once | yes |
| Make collar | 1.94 | partly | continuously | no |
| Attach collar to neckline | 1.42 | no | continuously | no |
| Set two sleeves | 2.38 | no | continuously | no |
| Close side and underarm seams | 1.76 | no | continuously | no |
| Attach cuffs and plackets | 1.86 | no | continuously | no |
| Hem bottom | 0.94 | in a folder | continuously | no |
| Topstitch front edges and facings | 1.42 | no | continuously | no |
| Seven buttonholes | 0.98 | in a clamp | once | yes |
| Seven buttons | 0.84 | in a clamp | once | yes |
| Six bartacks | 0.42 | in a clamp | once | yes |
| Attach main and care labels | 0.56 | in a clamp | once | yes |
| Join shoulder seams | 0.72 | no | continuously | no |
| Set front facings | 1.12 | no | continuously | no |
| Trim and turn | 0.46 | no | continuously | no |
| Inspect and thread-trim in line | 0.68 | no | continuously | no |
Two columns of that table are the whole lesson: Standard minutes and Candidate. Add the minutes up in each group.
Candidates: 9.36 standard minutes, 38.9% of the garment. Every one of them is either a preparation step — creasing, positioning, marking, labelling — or a single point of stitching done inside a clamp. Not one of them is a seam that joins two panels along a curve.
Not candidates: 14.70 standard minutes, 61.1% of the garment. Every one of them is two limp plies being brought together along an edge that moves while the machine works, usually with the ease spread unevenly along it, on a panel whose shape changes as the seam closes.
So the ceiling on sewing-room automation for this garment — assuming every candidate were automated perfectly and free — is 38.9% of the work content. The remaining 61.1% is not waiting for a better machine. It is waiting for a different material.
That is an uncomfortable number, and it is the most useful one in this course, because it tells you where to stop looking. A merchandiser or an owner who has been told that the sewing floor is about to be automated can now answer with a figure taken from their own time study in an afternoon.
The candidates are mostly preparation, and that is the finding
Look again at the candidate group. Of its 9.36 minutes, 6.56 minutes — 27.3% of the whole garment — is the four patch pockets: creasing them, positioning them, and stitching them down in a clamp. The rest is buttonholes, buttons, bartacks and labels, which are already done on machines that clamp.
So the automation that pays in a garment factory is nearly always in handling and preparation rather than in stitching. That is not an opinion about vendors. It is what falls out of the three questions when you apply them honestly. Preparation is exactly the set of operations in which the cloth is being made rigid — folded, creased, fused, clamped, pressed flat — so that a later operation can be done on it. The preparation step is the fixture, and a fixture is something a machine can own.
The same test explains what is already in the building.
| Machine | What holds the cloth still | Where it sits |
|---|---|---|
| Automatic spreader | cloth under controlled tension between the roll and the table | cutting room, see 20.2 |
| Cutter with a vacuum table | vacuum squeezes the lay into a block | cutting room, see 20.2 |
| Continuous fusing press | the belt holds the panel flat against the heated plate | fusing, see 20.3 |
| Buttonhole and button automats | the clamp | sewing line, see 20.4 |
| Form finisher | air pressure holds the garment on a body form | finishing, see 20.3 |
| Overhead transporter | the hanger is a rigid handle bolted to a limp garment | line 3, lesson 4 |
Every machine in that list passes question 1, and the last one passes it in the most instructive way. An overhead transporter does not solve the limp-cloth problem. It steps around it. A garment clamped to a hanger is a package with a handle, and moving a package is an ordinary materials-handling problem that the rest of manufacturing solved long ago. That is why transport automation in clothing behaves like transport automation everywhere else, and why lesson 4 prices it separately.
Where the test is honest about its own limits
Three warnings, because a screen that pretends to be exact is worse than no screen.
A yes is a candidate, not a purchase. Course 20.4 prices a dedicated automat properly and shows a machine that passes every physical test and still loses money, because the yearly volume of its exact shape was never large enough. Passing the three questions gets an operation onto the list. Lesson 3 is about what takes it off again.
Some no answers are only a no today. Question 2 fails an operation because an edge has to be followed all the way. Where a manufacturer has found a way to make the fabric itself rigid for a while — a stiffening treatment washed out later, a fusible tape along the seam line — the answer changes. And it changes because the material changed, exactly as the principle predicts. Judge a proposal by what it does to the cloth, not by what it claims about the robot.
The scoring is per garment, not per factory. A tailored jacket, a knitted polo and a pair of jeans give three different splits, because the share of clamped point operations differs enormously. Score the garment you actually make the most of.
Prompt · Screen an operation for automation before you ask anyone for a price
When somebody proposes automating an operation, or when you are deciding which operation to look at first.
Act as an industrial engineer in a garment factory who has seen automation projects fail and has no commercial interest in the answer. I want an operation screened against physics before anybody quotes. Facts: garment [TYPE], style [CODE], annual volume of this operation across all customers [NUMBER], fabric and weight [DESCRIBE], the operation as written in the time study [PASTE IT], its standard minutes [NUMBER], and my full operation list with standard minutes [PASTE IT]. Do the following. First, answer three questions about the proposed operation and show your reasoning for each: is the material held in a known shape while the operation happens, and by what; does the operation line up once at the start, or must it follow a moving edge all the way; and is the thing that moves a single ply or a package with a handle. Second, give a verdict of candidate, not a candidate, or candidate only if something changes, and if the last, name exactly what would have to change about the material or the fixture. Third, run the same three questions over my whole operation list and give me two totals in standard minutes and as a percentage of the garment: the operations that pass all three, and the operations that fail because an edge must be followed all the way. Fourth, tell me which of the passing operations are preparation steps and which are single points of stitching done inside a clamp, because they are bought differently. Fifth, list what you would measure on my own cloth before going further, and say what each measurement would rule in or out. Do not recommend a machine, do not name a supplier, and mark every assumption you had to make.
AI can make mistakes — check anything you act on.
Check yourselfSomebody proposes an automated cell to close the side seam of a T-shirt. The argument is that the seam is straight, the fabric is cheap and the volume is enormous. Score it against the three questions, and say what you would need to see before spending an hour on it.Show the answer
It fails question 1 and question 2, and it does not matter how straight the seam is. Two plies of single jersey are being brought together along an edge, with no fixture holding either of them in a known shape. The machine has to keep finding that edge while the feed pulls it and the fabric stretches under the pull. Jersey is the worst case here, because it stretches under the same forces that are being used to move it. The straightness is a distraction: the difficulty is the material and the lining up, not the path. Volume does not rescue it either, because volume multiplies a benefit, and there is no benefit until the operation runs on its own. What would change the answer is something that holds both plies in a known shape for the length of the seam — a frame that grips both edges, or a tape that fuses them before they are sewn. Ask what is holding the cloth. If the answer is nothing, you have finished the assessment.