The End of the Line

Wireless  © Kel Harvey

Talking to someone yesterday about engineering, we discovered a common past. Some twenty years ago, in my life as a software designer/engineer, we got a contract to program a module for a unit in a production line assembling miniature circuit breakers (MCB's) for a major manufacturer of contract electrical goods and components in Holyhead.

The chap I was in conversation with was at the time a manager at the works. My initial meetings about the project were held with myself, my late friend Jean-Charles Boude and a manager from the company who coincidently, was from Paris.
The French Connection was further reinforced by the fact that the  production line system and its programming environment were also French.

At roughly the time we were awarded the contract, Jean-Charles left Wales to return to France and start his own company, so what followed was a solo effort on my part. The company who contracted us paid for an introductory course in the software that ran their system, which was highly proprietary and outside the scope of normal programmable logic - systems were familiar with.
Duly educated in the slightly off-the-wall semantics of their programming language, the code for the module was written and tested. The machine's hopper supplied with test-rivets, the software and hardware coalesced and worked flawlessly on soak. Then the company introduced the stock rivets and the machine started to break, often quite spectacularly. The company that supplied the machine tool were contacted and told that something was wrong, eliciting the reply that the piece of kit had been on soak test for weeks without a single failure. Next in the firing line was yours truly. The litany ran that I had gotten my dwell-times wrong and the machine was trying to punch holes in the MCB's rather than rivet. I knew this couldn't be the case because of the amount of time I'd also soak-tested the thing.

So, the managers forced a shoot-out, to be held in weekend downtime (never mind the fact that I had been required to do all of the development work on the live line during the working week!).
All three companies were obliged to turn up at eight AM and demonstrate that whatever was falling over was not down to them. The maker of the riveting machine and myself were there at eight sharp. No sign of rivet-man. So we fired up the module with the company's batch of test rivets and ran it consistently and reliably until some hours later, the rivet guy turns up.
So, we duly stop the test, empty the hopper of the test rivets and fill with the production rivets that rivet guy has brought to the table. Bam! - the module almost immediately breaks all of the very expensive hardened riveting tool heads!

In the autopsy that followed, it turned out that in specifying all of this stuff, the company itself had specified rivets to too close a tolerance and had also ultrasonically cleaned all the test rivets, removing sufficient dirt and oxide that they worked perfectly. The production rivets, which were after all to the spec ordered, gained 'weight' through normal environmental conditions - the very conditions they would be exposed to in the day-to-day environment of the stores and of the factory itself - they were never going to work as under normal working conditions they were simply too big.

The moral of this tale is: look to your own fallibilities and shortcomings before blaming others; a lesson our current Prime Minister would do well to learn before exposing his own vulnerabilities to the world each week at Prime Ministers' Questions, something that with luck will come home to roost for him and his ineffectual ilk.