In many texts on quality control, the authors wax passionate about the japanese commitment to quality. Supposedly, in Japanese auto plants, any assembly line worker who spots a defective process can immediately stop the production line. Is this true? stopping a production line is extremely costly-it probably involves a loss to the company of several tens of thousands of $/hour.
Plus, how would this help things? Would it make more sense for the worker to just summon the foreman, and inform the management?
I suspect this is another urban-legend type story-meant to show the "superiority’ of japanese QC management.
Yes. It’s famous Toyota Production System. It got a lot of hype some time ago and was with more or less (usually less) success copied in various types of business.
It’s not just the Japanese, either. I’ve been in U.S. manufacturing facilities where the line workers are encouraged to stop production if they see a problem.
As for what good it does, it’s cost-effective to fix the problem ASAP rather than have to go back, check (and usually double check) X number of finished components and fix/scrap the bad ones.
One example I happened to see first-hand. I was at a foundry when the finished castings started coming out with a cosmetic flaw. The workers immediately stopped production and traced the problem to a minor air leak, which was quickly fixed. However, the flaw could have also been caused by uneven cooling. If that had been the case, the castings could also have come out cosmetically perfect, but brittle. Having a worker stop the line avoided the quality control shift from having to go back and run metalurgical tests on the entire day’s production run.
This is true in most automobile factories, and for that matter, most factories of any kind. A nonconformance with a product will usually have an operator stopping things to get it fixed. If the problem is procedural, rather than just a one time mistake, you logically would have to fix it immediately of every product will be flawed.
The cost of making errors is almost invariably compounded by NOT preventing the error from continuing; a corrective action (e.g. a permanent fix) may take awhile to devise, but you need to at least contain the problem right away or you could screw up a whole bunch of products. Imagine if you’re milling stainless steel shafts for some industrial application. Which would you rather have your employee do; find there’s a problem with an operation, stop the mill, and have it fixed when you might still be able to save a $700 part, or just let it run and lose the material? Or, worse yet, he thinks there might be a problem but he’s afraid to say anything, and runs thirty parts, and only after then you find out a critical dimension is too small and they’re all scrap and instead of spending 20 minutes fixing the problem, you lost eight hours of production and will spend eight more making 30 new parts?
That said, it’s a bit simplistic, even in automotive assembly, to think about it as “Stopping the line,” as if one employee hits a big red button and everyone in the plant stops working. That absolutely does not happen, because nobody is stupid enough to design a factory that works that way; if work on one car has to stop because there’s a problem, other work can usually continue.
Read the link on the Toyota Production System, or Just In Time as it’s now (somewhat incorrectly) called; it’s a critical component of modern manufacturing philosophy that employees NOT have to wait to do things. You want a plant designed so that if Bill sees a problem and needs his widget machine adjusted, Ted and Jenny can continue doing what they’re doing. I’ve seen companies that put quite a lot of effort and thought into figuring out how they could set things up so one employee could cut five feet off of how far he had to walk to pick up a two-ounce part.
Oh, you better believe it’s costly.
The point though is that you find the defective processes as early in the production run as you can. If you have this culture set up where you stop everything and get it right, then things that don’t work are discovered very early on. In the U.S., the culture is often to work around the problem so that the line can go, so defective processes are often not caught until much later. If you stop and start and stop and start a bunch of times at the beginning, once you get all of the bugs out then you can just run the heck out of the line. All lines stop and start a bit when they first get going. It’s better to get all of the hiccups out of the way early than to shut down later when everything is going in full swing.
A lot of the workarounds that are done in the U.S. end up costing more in the long run, because while they allow the line to start up faster, the fact that something is kludged and not done right slows down that part of the line, and in a production line, any one part that slows down slows down the entire line thus reducing overall efficiency.
When I had an opportunity to work with a Japanese customer at one job, I was rather surprised at how much more engineering they did up front than we did. Their R&D and setup costs had to be through the roof compared to ours. I don’t think it is possible to do things that way in the U.S. because the bean counters would throw a hissy fit.
To echo what RickJay says, there are many, many, many decoupling points (think of 'em like buffers) among the disparate systems in the automobile production process. When a worker (even an American/Canadian/Mexican worker as in my company’s case) “stops the line,” everything beyond the decoupler is still running. Because production, profit, loss, and everything related is geared to “net” speeds, the line normally runs at this speed, say, 62 jobs per hour. But these lines are designed to run at overspeed should that be necessary – say, 70 jobs per hour. It’s really, really easy to get caught up.
Of course early on in a new assembly line operation, there are bound to be problems further down the line. Rather than starving the next operation, it’s much more common to fill your buffers and be backed up with nothing to do because of, say, the paint booth having caught fire.
So true.
Cost is relative. When I worked at a company that made seats for ford , the cost of something happening , that shut down a ford line was six million an hour(american)
If something happens at my current place of employment , we have to institute containment procedures and bring in a third party agency to double check our stuff and tell GM or DCX that our stuff is spec, and that can hit 50 k plus.
Until the widgets hit the truck to go to what ever factory , cost is virtual, when it hits the car company floor ,then it becomes real cost.
Declan