Lens quality control

Started Sep 24, 2013 | Discussions thread
ProfHankD Veteran Member • Posts: 6,234
Re: Sample variation is not that big a deal

parallaxproblem wrote:

Howver I think there is another factor at play here: offshoring ... Take a product that has been built in an established factory with a history of high-quality production and a workforce that takes pride in its activity, and move it to a developing nation where you recruit from scratch the cheapest workforce possible (why move if it isn't to pay the least salaries possible?) and the quality of output will simply *not* be the same whatever processes you put in place.

I'm not scared by this practice -- because when it doesn't work, which I believe is the norm, it is almost always quickly stopped. I've seen a lot of companies (including a local biggie: Lexmark) send all sorts of engineering and production work to other countries in search of cheaper labor, but most of them have quietly pulled the operations back after a year or two. The truth is that there's no free lunch. It costs a lot to develop the physical and personnel infrastructure in places where labor is cheap...  usually enough to reverse any cost advantage. By definition, large, successful, companies don't get to make and continue mistakes that kill product quality for very long.

That said, things change as investments are made. When I was a kid, I remember my parents regularly sending aid money to help starving kids in South Korea.  Well, South Korea isn't so poor now and some of the highest-quality high-tech stuff comes from that country -- especially things like LCD panels and DRAM memory parts. I've bought lots of camera parts (e.g., lens adapters) made in China and, to put it bluntly, they haven't been inferior to similar parts coming from Japan or Europe. Somebody has obviously invested a lot of money in China....  

I do believe that there is a new threat to quality control: less testing of complete systems made of outsourced parts. Standard engineering practice (SEP) is that tolerances needed for components to function properly are determined and specified as part of the design process -- especially for things like lenses. Unfortunately, the underlying analysis is usually based on most components being clumped around their ideal values, which is often no longer a valid assumption. It is now common that the companies making the parts often don't even know what the ideal value is, but just the range of acceptable values (incidentally, the ideal value often isn't in the center of that range). The concept of tooling being designed to "wear through" a tolerance range over a long production run also causes problems. Thorough testing of the final product would quickly reveal when too many parts being too close to the wrong edges of their tolerance ranges has resulted in a poor product, but such testing is expensive and unfortunate combinations of in-tolerance parts should be rare, so it is often apparently more economical to do minimal (if any) testing and let consumers discover and return the out-of-spec products.

My advice: test the product yourself and return it ASAP if it is clearly out of spec. I'm not talking about normal minor variations that produced products within the engineering specs -- I'm talking about the relatively rare clear outliers, like that article from LensRentals talked about....  It doesn't take sophisticated testing to identify outliers, and companies are usually good about repair/replace of them. For example, Sigma has had an unfortunate history of having a lot of outliers, but they also have usually been quick to make good on such returns.

 ProfHankD's gear list:ProfHankD's gear list
Canon PowerShot SX530 Olympus TG-860 Sony a7R II Canon EOS 5D Mark IV Sony a6500 +30 more
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow