A Tilted Element Demonstration

Posted by

One thing we preach here a lot is that just because a dropped lens looks fine on the outside doesn’t mean it is fine. The other thing we preach a lot is that a tiny difference in the centering, tilt or placement of a single element can have a dramatic effect on image quality.

Aaron made a great demonstration of this yesterday with a dropped Nikon 70-200mm f/2.8 AF-S VR II while I was off playing with autofocus.

The 70-200mm had been dropped, but it appeared to have no damage. When the incoming  technicians tested it though, this is what the optical chart looked like:


You can tell that’s bad even at 10%, but here’s a 100% crop of the centers and a corner.


Aaron took the lens apart to see what was up. When he removed the second lens group, which is encased in a single metal barrel, he found that the last element in the group (Element 9 from the front for those of you keeping score at home on your Lens Block Diagram) had gotten jarred out of place and was slightly tilted.

Just the second group is shown in the picture below. It’s sitting in a rubber stopper to keep the glass from touching anything.



Looking at it directly from the side, you can see the element is only tilted about 1.5 degrees. That’s a pretty amazing devastation on the image quality caused by such a little bit of tilt.


So, of course, Aaron glued the element back in its proper position.



The proof is in the pudding, as they say. (I’ve always wondered why they say that. I’ve never understood what that means. But I’m hungry and I like pudding a lot, so I thought I’d use that. Chocolate pudding. Maybe with some whipped cream. Yes, I’m on another diet.)

Anyway, here are the images of the test chart shot after Aaron put everything back in place. Looks like a 70-200 VR II should now!!


Isn’t it amazing how such a small tilt in an internal element has such a devastating effect on the lens?


Aaron Closz and Roger Cicala


August 2012

23 Responses to “A Tilted Element Demonstration”

Walter Freeman said:

Wonderful demonstration.

Honestly, you guys ought to run a repair business. I have a lens that I think has a tilted or decentered element and I’d pay you good money to fix it for me.

bluto said:

The original saying is, “the proof of the pudding is in the tasting”. It’s been shortened, in ways that the meaning is lost. Remember that in the middle ages, puddings weren’t a dessert, they were a meat/grain sausage like dish (like haggis). Since ingredients would be very hard to identify, and one may be eating pudding away from home (where reputations may not have been far reaching). There’s wisdom in cautious tasting.

I’m impressed the fix was that simple. It seems like mine are never as easy as a loose element. I’m more impressed that it appears each lens goes through an optical check on return. Good show!

A said:

Very interesting!

It looks like the top middle and bottom middle edges are brighter than the centre of the chart – is that a function of decentering, or is it just a feature of lighting that day?

Similarly, were the lighting/exposure the same in the before and after shots?

Maji said:

I second the comment of you all starting a lens repairing business.

peterKx said:

The proof is in the pudding – in other words to test the pudding you have to eat it. To test the lens you have to shoot with it.

Tommy Sar said:

The proper idiom is, “The proof is in the tasting of the pudding.” Which means you can only prove or confirm the taste of the pudding by actually tasting it. This makes a lot more sense than the commonly botched nonsensical, “The proof is in the pudding.”

LensRentals Employee

Roger Cicala said:

Adrian, the lighting on the lens was not good: basically the lens was being lit by the halogen spots we use for work lights, which cause a lot of glare.

The chart we shot is lit by fluorescents – it’s so large (about 6 feet across) and we have so many in the testing rooms that if we use hotlights the room becomes unbearably hot. But the downside is we have to deal with the irregular lightling fluorescents cause. We have halogen set ups in the lab where we don’t have to have them on all day, but not in the routine testing area.

Ed said:

Have you tested loose elements? I had the front element of a 20mm 2.8 come loose, but couldn’t tell you when, because I discovered it when I went to clean the lens and it rotated. I instantly grabbed the pack of 30 rolls of film I’d just picked up at the processor and dug through looking at the pictures I thought I’d shot with the lens, and couldn’t find anything out of whack. But, it might have been that way for months, meaning my comparison set would be bad. I’ve always wondered if orientation of the glass – that element rotating away – mattered or not. I’d think not… but…

LensRentals Employee

Roger Cicala said:

Thank you everyone for filling me in on the proof in eating the pudding. That seems so obvious, I should have figured that out. It’s amazing how much I learn simply by saying the words “I don’t understand”.

But I still want some chocolate pudding.


A said:

Thankyou for the setup info Roger!

Next time, if you get the chance, I’d be curious to see if there is any unusual vignetting on a lens with tilted elements. I think there should be, although I guess it very much depends on which group of elements…

Part of me also wants to test the effects of each group of elements being out of alignment – again I suspect there will be a radical difference depending on which group.

As an aside, have you considered LED lighting? It’s come on a long way since the early days! Philips do some nice stuff; but I accept that the prices can be eyewatering.

LensRentals Employee

Roger Cicala said:

Ed, rotation has almost no effect, except for certain lenses (Canon 24-70 and 70-200s come to mind) that use a ‘ramp’ to mount an element so that rotating it moves it forward or backwards. If your front element spun easily in place it didn’t have such a ramp so as long as it wasn’t so loose it could tilt it probably had no effect.

Adrian, this one did vignette more obviously and seemed a little darker even looking through the viewfinder, but you’re correct: it depends on the lens and which element.


CarVac said:

I want to but am afraid to take apart and fix my old Contax Planar 50/1.4, which has some sort of slight decentering: the bokeh edge is consistently harsher on one side of the blur circle than the other, and it doesn’t seem to be quiiite as sharp as it could be.

DavidH said:

Can only imagine how many lenses you have to check and fix on a daily basis…interesting read and interesting timing…
Just come back from a worldwide shoot filmshoot for Geo – a taxi driver in Switzerland dropped my camera bag and damaged my 70-200 (original VR)…had to buy a replacement in the UK on the way to Oz so the shoot could continue…painful.
If anyone can recommend somewhere I can get the damaged lens repaired then I’d appreciate it.

Am a very satisfied lensrentals customer – but didn’t have any of your lenses this trip – so not adding to your burden ;)

aleksander said:

I have a mental picture of Robin (that would be Aaron) sitting next to Adam West (that would be you) whamming the lens back into proper alignment.

Adam said:

I started with a couple of cheap throwaway lens first. A film kit lens, old 50 with fungus, or a hanimex/soligor/etc 135mm or something similar can usually be plucked from a bargain bin/ebay lot etc for under $10 bucks.

That said, most manual focus lenses aren’t really complex. One good piece of advice is take a photo of each step so you’ll have some idea of what it should look like going back together.

The hardest parts are generally be if you should somehow take the lens apart at the focus helicoid (you’ll have to put it back together correctly to retain infinity/close focus). If you leave that alone, the other important thing is making sure you put the lenses back in in the same direction (it sounds dumb but it’s easy to flip biconvex or biconcave lenses with similar radii).

Have a small pill box, egg carton, or similar compartmentalized box to put the screws/parts from each step in, and don’t force anything, there’s sometimes hidden screws that are preventing something that would normally be easy to remove.

You’ll probably want a lens spanner, and definitely want a good set of jeweler’s screwdrivers. I keep a garlic peeler handy for removing front elements.

Traveler 2012 said:

Just out of curiosity what percentage of lenses come back after being dropped? Or thrown, stepped on, eaten by dog etc.

A side note: Glad you retired, the blogs are great.

LensRentals Employee

Roger Cicala said:

Maybe 1 in 200 lenses come back damaged, slightly higher for cameras. But that makes for several a day.

ME TOO!! and thank you.

KimH said:

Hi Roger,

yet another great article, thanks for making them.

When you say “we glued the lens back” i stopped and thought, glue…? What type of glue?

This is another piece of science. Superglue surely not (for about 100 reasons of which i can think of only a handful), but, what type and how did you find out?


LensRentals Employee

Roger Cicala said:

Absolutely not. I would be totally paranoid the fumes would settle on the elements and do something awful. We just use the same type of for-glass Loctite the manufacturers used originally. The element fits into a ‘clip grove’ that seems to hold it by itself, but Nikon had a dab of red Loctite at the edge of the element and we replaced that.


User in Canada said:

I second the idea of a repair shop. I had a 2K$ L lens develop some kind of misalignement and the official Canon Repair never managed to repair it satisfactorily. I returned it 3 times. First they said it was “within tolerance” and returned it to me, probably untouched, although it produced images almost as bad as your first test chart. Then they held on to it for two months, saying they had to send it to Toronto, then, some other lame excuse, and after much protest and me showing them the crappy images it produced, to their credit, they eventually (6 months later) kept it and send me another one that worked as expected. That’s the kind of Service you get for buying 2000$ lenses. 6 months without the lens… I also had a 100/2.0 sent to them for repair after a fall and although they charged dearly for their service, they never could get the image quality back to normal. I though that a lens had to be unrepairable, but apparently you guys can repair one :)

LensRentals Employee

Roger Cicala said:

User in Canada, you might want to read: http://www.lensrentals.com/blog/2012/07/optically-adjusting-a-lens

The process of optical adjustment is really time consuming and I think most of the manufacturers aren’t willing or able to spend hours getting it right. Not to mention sometimes there’s a major part out of true that can’t be detected by inspection. Replacement is the only option.

The problem we’d have with opening a repair facility is we have to reverse engineer everything we know. That ends up meaning we know Nikon and Canon 24-70s, 70-200s and other common lenses, but we’d be clueless about some of the less common ones. Plus there are a lot of adjustments and repairs that require specific Canon or Nikon testing equipment that they won’d sell.


Walter Freeman said:

I have a Zuiko 50-200 f/2.8-3.5 that I bought used, and had a rubber gasket break loose and damage the zoom mechanism inside a few months after I got it. I sent it back to Olympus and they fixed it (for $300), but it came back with significant wobble in the barrel and — most worryingly — unsharp images at 200 f/3.5, especially when used with a teleconverter … which is how I use it most of the time. So it went back to Olympus, and they fixed the wobble but it’s still not sharp. It’s “good enough” at most apertures and focal lengths, but not with the TC, which is how I’d like to use it.

I can’t in good conscience sell it (although I’m thinking about switching systems to an EM-5 with a legacy 300/2.8), so I guess all there is to do is to open it up and see if I can figure out what’s the matter.

User in Canada said:

Thanks Roger, very interesting read on optically adjusting a lens. I guess my conclusion (an optically compromised lens is a lost lens) is true most of the time, and accordingly, I’ll donate any future underperforming lens to the first unsuspecting squiggy at a red light and save the repair money for a new lens.

Leave a Reply