Documenting obsolete spares before the machine stops
The cheapest version of a reverse engineering job is the one commissioned while the machine is still working. Here is how to decide which parts qualify.
Why this never gets done
Because it competes with things that are on fire. Documenting a part that has not failed yet is the definition of work that can always wait until next month, and it does, until the day it cannot.
The economics are not subtle, though. Reverse engineering a component while the machine is running is a scheduled job with a normal lead time. The same component after a failure is an emergency, carried out against lost production, usually with a damaged sample and no time to establish material properly. The engineering content is identical. The cost, and the quality of the answer, are not.
The screen: four questions per part
You do not need to document everything. Most spares are available, cheap and quick, and drawing them is a waste of money. The exercise is to find the small number that are none of those things.
- If this part failed today, what stops? One machine, one cell, or the whole line. If the answer is nothing much, stop here and move on.
- How long to get another? Off the shelf, six weeks, or nobody knows. Compare that honestly with how long you can stand the machine being down.
- How many people can supply it? A distributor, one specialist, or only the original manufacturer. Single-sourced parts are where the risk concentrates.
- Does a drawing exist? Held by you, held by the supplier, or nowhere at all. A part you can buy today but could not have made tomorrow is a latent problem.
Anything scoring badly on question one and on any of the other three goes on the list. In most works this produces somewhere between five and thirty parts, not hundreds, and that is a manageable programme.
Guards, brackets, chutes and fixtures made by a fabricator who has since closed. Feed rollers, wear plates and formers on old but perfectly serviceable machines. Anything a previous maintenance engineer made in-house and never drew. These are rarely on a spares list at all, precisely because they were never bought.
What to capture, and in what order
Do the highest-risk parts properly rather than all of them badly. For each one, in this order:
- Photographs and basic measurements now, while an undamaged example exists. This alone is worth doing on a wide list of parts, cheaply, in an afternoon.
- A full drawing for the parts that failed the screen worst. This is the deliverable that turns a future crisis into a purchase order.
- Material identification where the duty warrants it, done while there is time to send a sample to a laboratory rather than guessing under pressure.
- A proved first article for the genuinely critical items — one part actually made and test-fitted at a shutdown. It is the only way to know the drawing is right, and it converts the spare from a theory into a shelf item.
The by-product nobody expects
Doing this generates a controlled drawing register for the equipment you depend on, which has value well beyond maintenance. It is one of the things a customer quality audit asks for and is routinely told does not exist. It supports second-sourcing, which resets prices. And it makes the plant less dependent on the one person who knows how the machine goes together — which, in most engineering SMEs, is the largest single operational risk on the site.
If some of that register already exists on paper — old prints in a plan chest that nobody has opened since the last person who understood them left — that is a different and usually cheaper starting point than measuring parts from scratch. Redrawing paper drawings as manufacturable CAD
You do not have to do it all at once. Ten parts a year, chosen by the screen above, will get most works into a materially better position within three.