Archive : Quality with a sniff

Motor, 4 June 1983

Cowley paint shop

Robots and new production line techniques are helping to ensure the improved quality of Austin Rover cars. Howard Walker reports.

You won’t see many white coats along the production lines at Cowley or Longbridge — the traditional image of the white-coated quality controller peering into nooks and crannies is fast disappearing at Austin Rover.

‘You can’t inspect quality by looking at it. When a chap has clipped on the door trim, how the hell do you know what’s behind it is working properly. Quality is with the man fitting the parts.’

This is the view of Austin Rover’s Director of Quality, David Caswell, a straight-talking Midlander whose firm ideas on how his company’s cars should be screwed together are gradually shaking off the poor quality image suffered by BL cars in the past.

His job has not been easy. It has required a radical change-around in production line procedures, a massive investment in new machinery and a new way of thinking by the Longbridge and Cowley workforces — a major task in itself. Yet his efforts have paid dividends; the Metro set new standards in quality in the Austin Rover division which were improved on with the introduction of the Maestro, and Caswell is confident that the forthcoming LM11 saloon will be even better.

‘I don’t aim to get any greater standard than we achieved in the past because we honestly believe that standard was good enough. But the major problem was trying to get every car we made to that same standard. lt didn’t happen. One of our slogans to the workforce now is that quality is repeatable, and that’s what we are aiming at,’ he said.

What Caswell has done is to rethink the whole car assembly process and divide the different stages of assembly into different sections — there are 14 for the Maestro. Each of these ‘zones’ is controlled by a foreman whose job it is to ensure that the workers fit all the parts correctly. And one of the tasks of the workers in the next zone along the line is to check on the work of the previous men.

‘Much though we value our workforce – and we do have a good workforce here — people are variable. The chap who has a row with his wife in the morning comes into work and may not fit the first door panel too well, so there needs to be checks all along the line.’

One of the major advances in Austin Rover’s quality control is Caswell’s Quality Index System. It gives the management the absolute right to take any car it likes off the end of the production line and put it under the magnifying glass. On the Maestro line, five cars a day are taken off and first inspected under critical lighting to check for paint defects, then put on a ramp to check the underbody, given a water test to check for leaks, and then taken on an 18 mile road test to test the mechanical components.

The inspectors have a manual which defines the degree of severity of any defect with ten, five and three demerit markings. A ‘three’ represents a minor defect like a paint blemish out of sight, say on the underside of the door sill; a ‘five’ is a slightly more serious defect; while a ‘ten’ is a totally unacceptable problem which results in the car automatically being sent back for rectification and the foreman responsible pulled over the coals.

But Caswell believed that that in addition to the visual checks made on random cars, more sophisticated quality checks had to be made on every car being produced. One area which had caused constant problems was that of checking for water leaks.

Drenching a car in hundreds of gallons of water did detect leaks but, in doing so, often soaked expensive interior trim which then had to be replaced. So in conjunction with BL Technology, Austin Rover came up with the world’s first ‘sniffing’ robot, which really is quite a brilliant piece of equipment. Basically what happens is that a car moving along the line has a plastic hose fixed to it and air seeded with 0.5 litres of helium is pumped in under slight pressure.

Two robot arms — one on each side of the car — equipped with sensors which detect the slightest whiff of helium, scan the whole of the body, tracing around the windscreen, door edges and boot checking for leaks. It then feeds its findings into a computer which produces a print-out in the shape of a car clearly showing where the leaks are. If leaks are in fact detected, they are rectified later along the line.

To test for those niggling faults like tailed bulbs, faulty indicators or non-working switches, Austin Rover call in the VETS. It is a new computer system which cost a cool £310,000 to install and the initials are short for Vehicle Electric Test System and is in operation at both Cowley and Longbridge. Probes are attached to each car on the line and computers measure, magnetically, the current passing through each circuit — if there’s a fault, then the current is different and can be detected.

Austin Rover also has what is claimed to be a unique ‘TV eye’ system which checks Metro brake discs — it’s expected to be used for other models later — for the most minute imperfections.

The system has been jointly developed by BL engineers and Warwick University staff and uses special cameras to check for pinpricks on the surface of the disc. Faults as small as two millionths of an inch can be detected.

‘All this is just the tip of the iceberg compared with what is to come in the future. We are looking at ways of robotising the application of underseal and using some of the concepts of the sniffing robot to check the bare car bodies for accuracy,’ said Caswell.

And was the massive investment in quality control paying off? Caswell said it was. Warranty claims had been reduced significantly and feedback from dealers and customers had been enthusiastic.

He explained that to discover what people actually thought about the cars, many new customers were contacted one month and six months after they had taken delivery, Also every few months, newly-built cars were taken to different towns and 50 people pulled off the street and asked what they thought about the cars.

In addition, every three to four months, around 80 dealers were asked to Cowley or Longbridge to give brickbats or bouquets on cars fresh off the production lines. ‘We are getting it right but it takes time,’ he added.

Keith Adams

27 Comments

  1. The issue was the quality does not rest so much with the man, but the design. This was the success of the Japanese and why the Acclaim had such a good reputation quite literally out of the box, whilst the Maestro and Montego always struggled, because quite simply you would have to struggle to build a bad Acclaim because the quality was engineered into the design, both in the way its assembled and its components ina a way BL did not know how at the time.

    You feel the same getting between an early Longbridge 75 and a late one, the successive cost cutting of project Drive succeeded in taking a car with an above average feel to below average, simply because the quality of many of the materials such as the leather, carpets and certain trims all been downgraded. No doubt each change considered in isolation would be considered insignificant, but in total they all but ruined the car.

  2. I agree with Graham. Working for ARG we bought an early production Maestro 1.3L for my mother. TOJ525Y was a great car in many ways. Very quick for a 1.3, refined, brilliant ride and handling and much more comfort and space than a Ford Escort or Vauxhall Astra. 45mpg easily achievable and 5 of us took it to France on holiday. Thats the good. The bad was poor idling with the stepometer that reduced revs to 600rpm to save fuel. Very poor dash with lots of rattles and squeeks and a notchy gear change. All the faults were design led not anything to do with the production line worker. Slowly Maestro improved with the 1 piece dash and had it had the k series 1.4 with better gear change it would have been a very competent car. R8 of course achieved most of this as the Rover 214, but even that car did not have the space and ride quality of the LM10/11 series.

  3. Where Toyota did well in terms of process was by using the Kanban system, where quality is built in.

    *Everyone* is empowered, and any step of the process can be challenged. The goal is to strive for inherent quality by eliminating all waste in the process.

    Anyone can stop the line at any time if they feel that something isn’t right.

    It is a system that is being copied by many software companies.

  4. Fascinating piece. In terms of cars at the time, I think quality was in the design but probably more in the assembly as car design was just no as sophisticated – many design issues were more about common sense than micro-accurate design methods. What was apparent from those days, es exactly what he man in the article said, variability from one car to another. How do you explain for example, why my 32 year old Metro is almost mint, faultless, when others gave up the ghost at 2,3 or just 4 years old. Sure, some of it comes down to use, but if the paitn was not quite right, the underseal not properly applied, a leak not detected etc this had massive effects on perceived quality. One of the biggest big-bears for Austin Rover’s image at the time was rust. They spent millions with special dips and computer assembly for the Metro, including paint, but the cars still started rotting almost the minute they left the line. This probably had more of a perceived quality effect on the consumer than more mechanical issues that typify ‘reliability’ – I do think you might be more inclind to forgive your car that breaks down once in a while if it still looks ok, but if it not only won’t start on a damp day, but you can see daylight beaming through every panel, you might have more of a lasting negeative impression. The Metro should have been instinsically reliable as the constituent parts were more than well proven. Imagine how many mre might be left today had they bothered to properly protect them from the ravages of rust. I can never quite decide whether it was a strategic decision taken by some mis-informed big-wig who may have felt that customers would not come back if there cars lasted forever? The effect was probably quite the opposite.

  5. Fascinating piece. In terms of cars at the time, I think quality was in the design but probably more in the assembly as car design was just no as sophisticated – many design issues were more about common sense than micro-accurate design methods. What was apparent from those days, es exactly what he man in the article said, variability from one car to another. How do you explain for example, why my 32 year old Metro is almost mint, faultless, when others gave up the ghost at 2,3 or just 4 years old. Sure, some of it comes down to use, but if the paitn was not quite right, the underseal not properly applied, a leak not detected etc this had massive effects on perceived quality. One of the biggest big-bears for Austin Rover’s image at the time was rust. They spent millions with special dips and computer assembly for the Metro, including paint, but the cars still started rotting almost the minute they left the line. This probably had more of a perceived quality effect on the consumer than more mechanical issues that typify ‘reliability’ – I do think you might be more inclind to forgive your car that breaks down once in a while if it still looks ok, but if it not only won’t start on a damp day, but you can see daylight beaming through every panel, you might have more of a lasting negeative impression. The Metro should have been instinsically reliable as the constituent parts were more than well proven. Imagine how many mre might be left today had they bothered to properly protect them from the ravages of rust. I can never quite decide whether it was a strategic decision taken by some mis-informed big-wig who may have felt that customers would not come back if their cars lasted forever? The effect was probably quite the opposite.

  6. Blimey, not only was my spelling shocking in that last post, but I inflicted it upon you all twice! Maybe I need to have my own quality system installed…

  7. I’m sure we’ve all had or heard of a ‘friday afternoon’ special. Any mass produced product is going to have some sort of failure rate, though with increased QA and component quality, this is driving down.

  8. @ 3, Kanban is about logistics and supply chain but is an important link in the quality chain. You’re right though, Toyota invented it (in early 1950s). Didn’t stop them having recent well-publicised quality and safety issues but they managed them so well they arguably came out of it stronger.

  9. Quality is a very catch all term.

    You can have quality control of the assembly process to ensure all panel gaps are consistent and dashboards don’t rattle ect. There should not be Friday afternoon specials in the 21st century .

    However as others have said the quality of the components sourced from suppliers and build materials used is far more relevant these days.

    My Honda SKIP from Swindon I would rate 5/5 for the former and 0/5 for the latter

  10. @8

    Software projects I worked on using the kanban process weren’t necessarily defect free, however it was a process of continual improvement, so a defect stopped the line and all hands on deck to fix it, and ensure it doesn’t happen again. Both during development and post release.

    I’m sure Toyota had a similar process in place. As you say, they appear to have emerged stronger. Teflon Toyota 🙂
    Don’t forget too that one of their recalls was due to a 2cm piece of carpet mat. I suspect that many rival manufacturers would issue an internal service bulletin rather than a recall…

  11. The daft thing is that as it becomes easier to design things to within a micrometre’s accuracy these days, cars are also more and more complicated in terms of software which I fear is where most failures result these days. Same with fuel efficiency, we get more and more efficient engines but in sadly bigger and bigger cars so the benefits are lost a little. If I could stick my 3 cylinder VW engine in my Metro, I’m sure I’d get about 100 mpg!

    • Have you read the section on the BL ECV Programme, see concepts section, focus size with 3 cyl 1110 cc engine, CVT autobox and light weight aluminium body from the late 70’s – early 80’s. 0 – 60 mph 11 sec, 115 mph, and 81 mpg at 56mph 133 mpg at 30mph.

      With developments in fuel injection and electronic control, 35 years development and were would we been to-day ?

      To-day future cars are heading for light weight small engine auto, are we to far ahead, or just cannot market ourselves.

  12. This was written before the Montego came out – by memory, it took them until the Montego facelift in 1988 to start building them halfway properly. What the Firm never understood was that people who bought a rubbish “Mark 1” car wouldn’t come back for a properly built “Mark 2”.

  13. So what happened to Toyotas quality? with several lawsuits and safety recalls into the hundreds of thousands? The car in front is shit.

  14. @ 1.

    Graham is absolutely right. It’s the design and engineering that make or break a good car – not the assembly – which explains why the Acclaim and the first Rover 200 were good cars

  15. The Japs don’t always get it right.
    The problem is not confined to the chain that ‘goes the extra mile’

    In June 1999 I bought a new Honda Civic Estate. Within six weeks and 300 miles I drew to the attention of the general manager P. Wood, of the authorized Honda franchise, Elite Motors, Garratt Lane, London, SW17, that the road noise is excessive. He ignored me.

    After two years and repeated letters failed to get any response from Wood, I wrote to Kier, M.D. Honda UK. Kier then fobbed me off with the nonsense that an estate body necessarily makes more noise than a saloon or hatch. I told him that it doesn’t, and told him of several experiences over 40 years.

    In September 2003, Kier’s sidekick, Hunt drove the vehicle and was so astonished at the racket this thing made that he sat for several minutes deep in thought and called the noise, ‘a rotary howl’. Hunt then volunteered to supply a proprietary after market sound proofing kit. Fine. Except the kit is useless. It made no improvement whatsoever.

    During the conversation I told Hunt that I had received a lot of pieces of paper telling me nothing. Hunt told me, ‘Yes, I am the brick wall’.

    Why does Honda need a brick wall?

    Honda told me there is no cure for the noise.
    Why is there no cure?
    Did Honda not do any research and development on the estate body?
    Does Honda not care about customer satisfaction with their products?

    In 2004 Kier made the unilateral decision that the matter was closed, irrespective that the noise had not been cured.

    In the nine years from 2004 Kier has refused to communicate with me.
    Why? Is there something about this problem that scares Kier / Honda witless?

    In the first six years the vehicle was broken into six times where I have left other cars, older and newer without this problem, and no cars belonging to other drivers has been targeted. I accused Honda of making a vehicle with inadequate security. Honda did offer £100 toward an alarm. But
    security should have been a lot better in 1999.

    In 2004/5 I had occasion to drive a 1963 Morris 1100, 1969 Riley Kestrel and 1967 Singer Gazelle. None of them made this ‘empty oil drum’ din created by this Honda. The Singer chassis was developed from the 1948 Hillman Minx, so if Rootes could manage to make a car run without this infernal din in 1948, why was it beyond Honda’s ability 50 years later? Remember, Honda was supposed to be a quality manufacturer.

    The matter has not been closed, the clock stopped on or around 25th July 1999. I still require Kier to explain why:
    1. The vehicle creates excessive road noise
    2. Kier refuses to rectify it
    3. Kier refuses to explain why he refuses to rectify it
    4. Can the fault not be cured.

  16. @13 – For some reason Toyota’s recalls for any minor issue – The sort EVERY other car manufacturer has to deal with always get blown out of all proportion. Law suits of course only happen in one country – The US.

  17. I had a recall for my 2nd hand Toyota Yaris for a potential problem about 6-7 years after it was built.

    I was impressed by how they managed to trace me & had pencilled in a booking at the local dealership I hadn’t been to. They managed to deal with it while I was at work (not too far way, which was luch), with no fuss at all.

  18. My second owned car a Corsa B was recalled by Vauxhall back in 1999, it was fitted with new plates to secure the runners under the front seats free of charge at the nearest dealer.
    Not so with the Vectra B it had a bad habit of over revving at idle or in traffic at four years old and they can not fix it. So I got a small GM filter box fitted to the pipework on the right hand side of the cam cover somehow cured it for a very long time on till parting with it in 2009 it was getting ropey and too small for my needs being a saloon model.
    I found out even latter that Vauxhall mechanics hated all of them due to the cars shocking built and mechanical quality and the endless work of putting things right from day one.

    The image above reminded me of my neighbours Rover 620 one of the last V-plate cars in bright red, he owned it for the same length of time as my Vectra and with all the years of washing, polishing and waxoyling we did to these cars by 2008 I gave up as rust spots broke out all in the usually Vauxhall places but the Rover had none. Was it dipped in a fountain of youth at Cowley or Longbridge.

  19. When Porsche were in a lot of trouble in the early 1990’s, falling sales, high costs. Toyota sent in their production engineers to reorganise and “teach” Porsche the Toyota lean production methods.

    The long term result: Porsche built their fisrt ever defect-free car emerged from the end of the assembly line, and the highest profit/car figure in the industry.

    Why Toyota should volunteer their services to Porsche is a mystery.

  20. ref 18, Will 101 – I can also endorse the Rover 600. I bought a new 623SLi in 1996 and ran it until acquiring a Rover 75 Tourer in 2005. The 623 was totally and utterly reliable. All I ever did was have regular services, replace the odd tyres and rear exhaust can. It must also bid fair to be the easiest car to wash ever made – it was just so smooth-contoured and free of sponge-trapping details.
    If I’d had somewhere to keep it, I wouldn’t have parted with it – it still looked and ran like a new car after 9 years, despite living outside and having no special pampering or anything. Definitely a great bangernomics car today, and some should be kept for posterity in this 20th anniversary year of its launch.
    The 75 Tourer is very nearly as good on reliability – just the odd glitch with the much more complex gizmos like sat-nav – and not quite so easy to wash!

  21. If I remember rightly, Toyota didn’t volunteer, it was the consultants for/from Toyota. At the time Porsche spent as long in rectifying production faults as Toyota spent on producing a car.

  22. @14 – I agree that that is true more today, but back then, it was as much about assembly as engineering and design. These days the design element means you cannot really put something together badly, it either goes together or it doesnt. Back then, that was not necessarily true.

  23. The Japanese just-in-time method is their weapon, small batches of parts are made at frequent intervals.

    The Ford style of mass production; a warehouse full of parts to feed the assembly line, if that warehouse contains 2000 ill- fitting wings, then that is how the car will emerge off the line, 2000 cars with ill-fitting wings.

    They cannot afford the expense of scrapping 2000 wings, and they cannot stop the line to wait for replacements to be pressed

  24. The just-in-time method is adopted by pretty much all car manufacturers now, including Ford, and now TATA-owner JLR. The thing is, it’s a bit of a myth that parts are made in small bespoke batches, what it actually means is that the car maker is not prepared to buy large amounts of stock and instead passes the risk to the supplier, who instead has to drip-feed parts to the manufacturer, getting paid in small chunks along the way. The other reason is to allow a plant to operate more efficiently by not having parts stockpiled all over the place. Whether this has any meaningful effect on quality is questionnable as I am not sure what possible difference it makes.

Leave a Reply

Your email address will not be published.


*


This site uses Akismet to reduce spam. Learn how your comment data is processed.