The certificate is issued against a construction

The situation

Month fourteen of thirty-six. On 6 June, Aldercott Weaving writes to say that article AW-330, the fluorescent flame-resistant cloth in the HV-620 jacket, will be discontinued at the end of the year. The line is being re-tooled for a different construction. Aldercott is not being difficult. The mill has 62 weeks of Tunca's requirement in front of it, and it is telling Tunca fourteen months early. That is how a good supplier behaves.

Three replacement articles are offered. On the datasheet, all three are at least as good as AW-330 on every property Tunca measures. One is USD 0.75 a metre dearer and better on two of them.

The purchasing instinct is to pick the best of the three and move on. That instinct is right in almost every category in this academy. Here it starts a process with a queue, a fee and a calendar.

What the certificate is really about

Course 16.6 teaches the general form of this, and teaches it well: a specification change re-opens every claim that was proved on the old specification. That is the rule. It is not specific to this category, and it will not be worked out again here.

What is specific to protective clothing is the machinery underneath it. When a garment is sold as protection against a hazard serious enough to need third-party assessment, what exists at the end of that assessment is a certificate describing a type — one particular garment, made one particular way, from materials and components named in a technical file. The file does not say "a fluorescent flame-resistant cloth". It says AW-330, and it names the tape article, the thread, the zip and the closure.

Three things follow, and only the first is obvious.

The change is not yours to make alone. Where a body issued the certificate, the maker has to tell that body about any change to the approved type. Then the body decides whether the change affects conformity, and what further assessment is needed. You may believe the substitution does not matter. Your belief is not the mechanism.

So the change has a calendar, not just a cost. A test house has a queue. A certification body has a queue. Neither is obliged by contract to care about your ship date, and neither can be hurried by the levers that work on a mill. Course 6.3 covers submitting a test package early enough that a failure can still change the cloth. The addition here is that a second organisation sits after the laboratory, and its queue is the one nobody plans for.

And the file outlives the order. The technical documentation and the declaration of conformity have to be kept for a stated period after the last unit is placed on the market. That period runs on past the end of the contract, past the end of the relationship, and past the working life of everyone who signed anything. Which cloth was in which garment stops being a production record and becomes a legal one.

Two routes, and the naive comparison is wrong

Tunca has two real options. Re-certify on a new article, or buy out enough AW-330 to finish the contract.

The tempting comparison is USD 14,600 of test and body fees against roughly USD 90,000 of cloth. That makes re-certification look obviously right. The comparison is wrong, and getting it wrong is the ordinary way this decision is made badly. The cloth was going to be bought anyway. Buying it early does not cost you its price. It costs you the money tied up in holding it, plus the risk of holding the wrong thing.

Here is the contract, counted forward from month fourteen. There are 62 weeks left, the contract consumes 51.4 jackets a week, and a jacket takes 2.85 metres. So the remaining requirement is 3,190 jackets and 9,091 metres. A realistic cutting allowance of 6% — which has to cover the single-garment lays for the size tail from lesson 2 — takes it to 9,636 metres.

The buy-out is cheaper on this contract by USD 14,268, and the arithmetic is not close. Anyone who priced it the naive way and chose re-certification chose the dearer route while believing they had chosen the cheaper one by a factor of six.

And the buy-out is still the wrong answer

Look at what each route does when it goes wrong. That is what decides this, and it is not on the table above.

If the re-certification goes wrong — the body wants more testing, the queue runs longer than eleven weeks, one property comes back borderline — you have bridge stock, and you have a mill that still holds AW-330 for a few more months. Every one of those problems costs money and time in amounts you can work out. The failure is bounded.

If the buy-out goes wrong, there is no route back. The cloth is discontinued. Suppose the 6% allowance turns out to be 9% — a slightly worse yield on the single-garment lays, a roll with a fault, the volume drift from staff turnover in lesson 6. You are then a few hundred metres short of a cloth that no longer exists, with a live contract, and the only remedy is an emergency re-certification with no bridge stock and no time. The failure is unbounded, and it arrives at the end of the contract — exactly when the renewal tender is open.

That difference in shape is worth USD 14,268, and Tunca paid it. Re-certify, and hold the bridge stock.

The cheapest component in the garment is inside the certification

While the re-certification file is open, purchasing raises a second saving. The certified thread is an aramid blend. It costs USD 0.11 a garment more than an equivalent polyester of the same ticket, in the same shade, with the same shade fastness — and on the sewing floor the polyester behaves better at speed. Across the 3,190 jackets left, that is USD 351.

Course 2.5 covers thread construction, and why a core-spun thread behaves differently from a spun one at high speed. This is a different question, and it has nothing to do with sewing.

On a garment certified to protect against the heat of an electric arc, the thread is part of the protection. Polyester melts. A seam sewn with a thread that melts opens exactly where the heat arrived, at the moment the wearer most needs the garment to be a closed shell — and it opens without the fabric failing at all. The standards for clothing that protects against heat and flame do not stop at the outer cloth. Their requirements reach the components, thread included.

So the saving is not USD 351 against nothing. It is USD 351 against a change to the certified type that has to go back to the body, on a file already open for a different reason, and against a failure that is a burn injury rather than a customer complaint.

This is the sharpest thing in the course, and it is worth saying flatly: in this category, the cheapest line on the bill of materials can be the one holding the certificate up. In every other category, the cheap trims are where you look for savings nobody will notice. Course 2.5 opens with a shirt order stopped by twelve cents of interlining, which is the commercial version of the same shape. Here it is not commercial.

Act as a technical manager for certified protective clothing who has taken changes through a certification body and has no incentive to make either route look good. A material in a certified garment is being discontinued and I need the decision priced. Garment: [STYLE], certified against [WHICH STANDARDS], assessed by [A THIRD-PARTY BODY OR SELF-DECLARED]. The component being withdrawn: [ARTICLE, SUPPLIER, WHAT IT DOES IN THE GARMENT, COST PER GARMENT, CONSUMPTION PER GARMENT]. Notice given: [DATE], last order date [DATE]. Replacements offered: [LIST EACH WITH ITS PRICE AND THE PROPERTIES THAT DIFFER]. Contract position: [MONTHS OR WEEKS REMAINING], remaining requirement [QTY] garments, current price of the old material [PRICE], of the new [PRICE]. My cost of money is [PERCENT] a year. Test house quote and lead time if known: [AMOUNT, WEEKS]. Certification body fee and queue if known: [AMOUNT, WEEKS]. Do the following. First, tell me exactly which documents change and who has to be told, in order, and be specific that the body which issued the certificate decides whether the change is material rather than me. Second, cost route A, re-certification: fees, testing, the bridge stock I need to cover the elapsed weeks, and the price difference on the new material over the rest of the contract. Third, cost route B, buying out the old material — and do NOT charge me the price of the goods, because I was going to buy them anyway. Charge me the money tied up, over the real holding period, plus a stated allowance for cutting loss and volume drift, plus an obsolescence risk you state as a probability times a loss. Fourth, put the two totals side by side and then, separately, describe what happens when each route goes WRONG and whether that failure is bounded or unbounded. Fifth, recommend one, and name the three things that would reverse your recommendation. Sixth, list every other component in the garment that a purchasing department might reasonably try to substitute during this window, and mark the ones that are inside the certification. If any input is missing, ask me for it rather than assuming.

What to take to your own order

1. Write the technical file's component list on the front of the bill of materials, so everyone who could propose a substitution can see which lines are certified.
2. When one of those lines changes, tell the body first and the cutting room second.
3. Cost a buy-out on carrying cost and risk, never on the price of the goods. The goods were in the budget already.
4. Choose between routes on the shape of the failure, not only on the size of the cost. A bounded failure at USD 20,208 beats an unbounded one at USD 5,940.