What We Lost When Repair Became Replacement
During much of the twentieth century, an engineer or skilled technician was expected to understand how a machine worked.
When something failed, the first question was not necessarily, “Which new unit shall we order?” It was, “What has gone wrong, and can it be repaired?”
Mechanics, electricians, instrument technicians and maintenance engineers diagnosed faults using drawings, gauges, test meters, oscilloscopes and, above all, experience. They listened to machinery, felt for vibration, looked for heat, checked tolerances and followed a fault through the system until they found its cause.
A worn component might be re-machined. A cylinder could be re-bored, a bearing or bush replaced, a motor rewound, a valve re-lapped, a seal renewed or an instrument recalibrated.
Wiring could be repaired, linkages adjusted and braking surfaces restored.
The technician did not merely identify a faulty item. He was expected to understand why it had failed and what had to be done to return the whole system to proper operation.
Machinery Designed to Be Rebuilt
Many older machines were designed with maintenance and rebuilding in mind.
Parts were accessible. Fasteners could be removed. Adjustments were provided. Bearings, seals, bushes and wearing surfaces could be replaced individually.
Engineering tolerances allowed a skilled person to restore a component without necessarily replacing the entire assembly.
This did not mean that older engineering was crude. It meant that the machine and the person repairing it formed part of the same system.
A workshop might keep raw materials, standard bearings, seals and fasteners rather than shelves full of complete electronic modules. A skilled fitter could adapt or manufacture a part when the original was unavailable.
Repair was particularly attractive when replacement parts were scarce, expensive or slow to obtain. Labour was often cheaper than a prolonged shutdown, and a local repair could return a machine to service long before a replacement arrived.
The Modern Module
Modern engineering has developed in a different direction.
Cars, industrial equipment, domestic appliances and electronic systems increasingly consist of modules. When a problem occurs, the normal response is to identify the suspected assembly and replace it.
That assembly may be a sensor, pump cartridge, electronic control unit, actuator, circuit board, gearbox module or complete sealed mechanism.
The modern workflow often becomes:
- read the fault code or reported symptom;
- identify the replaceable module;
- remove it;
- install a new or approved replacement;
- record the work for warranty, traceability or compliance.
This approach has obvious advantages.
A replacement module is quick to install. It has been produced under controlled factory conditions and should offer predictable performance. It may carry a warranty and can be traced through the supply chain.
In industries such as aviation, medical equipment, automotive manufacturing and power electronics, predictability and documented compliance may matter more than the ingenuity of an individual repair.
Why Some Modern Parts Cannot Be Repaired Locally
Modern components are often manufactured to extremely tight tolerances using specialised materials, heat treatment, coatings, adhesives and automated assembly.
A field workshop may not have the machinery, clean environment, test equipment or technical information needed to reproduce the original standard.
Some assemblies are sealed against moisture, contamination or unauthorised interference. Others contain software or coded electronic components that must communicate correctly with the rest of the system.
An apparently successful repair may therefore introduce uncertainty.
Has the component been restored to its original strength? Has it been recalibrated correctly? Will it continue to perform safely under heat, vibration or load? Can the repair be certified?
In a regulated industry, answering those questions may cost more than installing a new approved part.
The Economics of Replacement
The change from repair to replacement was not driven by engineering alone. It was also driven by economics.
Skilled repair labour became more expensive and, in many sectors, more difficult to find. At the same time, mass production and global supply chains reduced the cost of complete assemblies.
A factory may produce thousands of identical units under controlled conditions. A local workshop repairing one damaged unit cannot normally match those economies of scale.
Downtime has also become more expensive.
When a production line, commercial vehicle, aircraft or essential piece of equipment is out of operation, the cost of lost output can quickly exceed the price of a replacement part.
Swapping a module in an hour may therefore be financially preferable to spending a day dismantling, repairing, rebuilding and testing it.
Warranty and liability also encourage replacement. A manufacturer can offer known performance for an approved new component. It is much harder to guarantee the result of an improvised local repair.
Repair Has Not Disappeared—It Has Moved
It would be wrong to conclude that genuine repair no longer takes place.
In many industries, failed assemblies are removed on site and sent to specialist overhaul centres. There they can be dismantled, inspected, repaired, tested and recertified under controlled conditions.
The local technician performs the rapid exchange. The deeper engineering work happens somewhere else.
Aviation has long used line-replaceable units: assemblies designed to be exchanged quickly so that an aircraft can return to service. The removed unit may then be repaired by a specialist facility.
Similar practices are now common in vehicles, industrial machinery, electronics and medical equipment.
Repair has therefore not always vanished. It has become centralised, specialised and largely invisible to the customer.
What Has Been Lost?
The danger is not simply that we throw away too many components.
The deeper concern is the loss of understanding.
A person who repeatedly replaces complete units may become very efficient at following a diagnostic process without ever learning what happens inside the unit.
When the computer identifies Component A, Component A is replaced. If the fault remains, Component B may be replaced next.
That is not always engineering. Sometimes it is substitution by instruction.
The system may eventually work, but no one has necessarily discovered why it failed.
This matters because a fault code does not always identify the underlying cause. It may identify only the component that detected the problem.
A sensor reading may be correct while the diagnosis drawn from it is wrong. Wiring, contamination, mechanical wear, software or another connected component may be the true cause.
Without understanding, parts can be changed unnecessarily and the original fault can remain.
From Tools to Artificial Intelligence
Engineering has always advanced through better tools.
Robert Goddard’s successful launch of the first liquid-fuelled rocket on 16 March 1926 helped establish the foundations of modern rocketry.
Battery technology developed progressively rather than through one isolated invention. Stanley Whittingham developed an early functional lithium battery in the 1970s; John Goodenough later produced a higher-voltage cathode; and Akira Yoshino created the first commercially viable lithium-ion design in 1985. Together, their work laid the foundations of the batteries now used in phones, computers and electric vehicles.
Artificial intelligence is another step in that progression.
It is already being applied in manufacturing to predictive maintenance, monitoring, quality control, process optimisation and generative design. NIST describes industrial AI as a means of analysing data, supporting decisions and improving manufacturing efficiency, while also stressing the need for trustworthy systems, reliable data and effective human–AI collaboration.
AI can search, compare and process information at a speed no individual could match. It can identify patterns across enormous bodies of data and propose solutions that might take human engineers far longer to develop.
But that does not remove the need for understanding.
What Happens When the Information Is Wrong?
A computer can store and process more information than any individual human being.
That raises an uncomfortable question: do we still need to teach people large bodies of knowledge when a computer can retrieve information instantly?
Education may undoubtedly change. People may need to memorise less and become more skilled at asking questions, testing evidence and judging results.
But a system cannot reliably correct an error that everyone has accepted as fact.
If the original data are incomplete, biased or wrong, the speed of the computer merely allows the error to travel further and faster.
An AI system may produce a confident answer because its task is to identify the most plausible result from the information available. Plausibility is not the same as truth.
That is why knowledgeable human oversight remains essential.
The engineer of the future may not need to machine every bush, rewind every motor or trace every circuit by hand. But someone must still understand the physical principles, recognise an impossible result and ask whether the diagnosis makes sense.
Engineer, Technician or Part Fitter?
There is no virtue in repairing something merely because it can be repaired.
A new module may be safer, cheaper and quicker. Modern manufacturing has delivered reliability and consistency that earlier generations could scarcely have imagined.
Nor should we romanticise the past. Old machinery failed frequently, repairs could be inconsistent and much depended on the judgement of one individual.
The problem arises when replacement becomes a substitute for thought.
A genuine engineer does more than follow the screen. An engineer understands the system well enough to question the screen when necessary.
Technology should extend human judgement, not eliminate it.
The most successful future will not belong either to the person who rejects modern systems or to the person who accepts every computer result without question.
It will belong to those who can combine modern diagnostic power with the knowledge, curiosity and practical understanding that once defined the engineering craft.
That is the central theme of my book Working but Wrong: systems can function exactly as designed and still produce results that cannot be trusted.
The related changes in vehicles, servicing and electric propulsion are explored further in Blowing the Myths on Electric Cars.
- Read; Working but Wrong.
- Read; Blowing the Myths on Electric Cars.
- Read a related article; When Britain Exported to the World
- Read my other articles on; Engineering and Technology
C