How Do Canon Test Lenses? Model consistency/Canon QC?

Yeah that's what I meant. Darkened space with controlled lights, jigs set up and ready to go. All computer controlled so various combinations (say 10) can be shot off in total 10 to 15 seconds. Add a few additional seconds for the computer analysis which spits out a yes or a no. Max time per lens something less than 30 seconds.

I spoke to some of Canon's people at a local service facility once with regards to one of my lenses and they assured me that the testing is fully computerized - they'd put the lens on the analyzer and let the computer do it. More than that they couldn't say, but I was assured it was definitely not a subjective test dependent on operator input.

A moble phone may be more complicated than a typical zoom less.

But, I suspect that the assembly of a typical zoom lens is far more complicated than the assembly of a mobile
phone. I suspect a typical zoom lens requires far more intricate hand assembly of parts.

Enjoy! Lon

GSH wrote:
It's highly unlikely that they test every lens that comes off the line, it would be physically impossible.

Whilst a lens is a precision item, so is a car and both are mass-produced. Bad ones pop up from time to time.

C'est la vie

That is most certainly correct. All businesses simply test random samples of their products assuming that would represent the majority of their products.

I have worked in the field of product qualification for almost 30 years, on a cornucopia of diverse products (toys, laser surveying devices, optically based medical test gear, sportswear, injection molding, etc), and I can assure you that grand oversimplifications like "all businesses simply test random samples" is definitively incorrect.

Many manufactured items undergo absolutely no tests after assembly (production). But many undergo intense testing scrutiny.

Although I cannot be certain, I would be very surprised if every lens rolling off of the production line is not checked. This would be especially true for the more expensive/ and/or more complex designs. And most likely by using a computer driven test fixture.

Enjoy! Lon

The complexity comment was more directed at the functional complexity and the amount of testing required to confirm it.

It is sometimes surprising what companies do, I remember many years ago a QA audit of a company selling "high reliability" electronics parts. When the QA people urned up at the 'factory' it turned out to be a large garage with illegal imagrants doing the assembly by hand. The only way he could keep the costs down and cope with the high rework rate was the very low cost of labor.

But I would doubt there is much had assembly at this price point, although it depends on the numbers being manufactured, if it is the 1-10 of thousands it may not be worth the NREC to design more complex jigs. This might be the case for things like the 300/2.8 where there is lower volume but a higher price point. Once you get into production runs of +1 millon they would be nuts not to automate it.

In any event any hand assembly should be simple unskilled else the cost would spiral.

Not quite sure why you think a zoom assembly is all that complex, they look fairly simply assemblies if you take a look at the asembly diagrams a lot of which are available on the web.

That is right, they may do very simple tests on a 100% basis that require seconds and sample tests, possibly on 1% or 0.1% of the volume to check the processes on the line.