Equipment

The Sky is Falling and the Light is Leaking: the A7r Anti-Massacree

Published March 11, 2014

Anti-Massacree – A humorous anti-war movement from the 1960s, suggested in the Arlo Guthrie song Alice’s Restaurant. The song, like many of my posts, was criticized for being overly long.

Believe it or not, I’m mostly a lurker in online forums. I read the hysteria of the day mostly for my own amusement. Sometimes I type a response but I almost always delete it. Interjecting facts into one of the daily hysterical rants would be about as welcome as a cat at a dog show. Usually I don’t even go that far. I just think there’s a lot of people online without much to do and go back to work.

That was my initial response when people started talking about light leaks in Sony A7r cameras. Let me think this through. If you shoot a 30 second exposure at ISO 25,000 into the sun or with a studio strobe aimed at the camera, you get a light leak. I do that . . . let me think . . .  never. And this isn’t my first rodeo. I’ve noticed there are light leaks in lots of cameras if you look for them, and most of them were apparent without jumping through all the hoops it took to produce them on the A7r.

Judging from the various forums, however, at least a million people do that all the time and therefore this presented a huge problem to their style of photography.  The amount of hysteria over leaking sunlight made me think I’d stumbled into the vampire.com forums. I wasn’t getting involved in that for love or money.

But Dave Etchells and the crew at Imaging Resource have shown in a series of articles that there probably are some real-world implications for some photographers. When Dave asked us to take a look and see if we could find causes and workarounds, we did just that.

Testing Setup

The light leak has already been documented quite well by Imaging ResourceFerrell McCollough, and others. We tried a couple of simple methods to determine where the light leak occurred and found something that seemed pretty accurate. We took the fiberoptic light sources from our microscopes and shined them at various points around the camera mount with either lens caps or lenses mounted.

 

We found at ISO 25,000 and 6 second exposures we could reproduce the leaks quite readily. More importantly, we could see the leaks varied depending upon which quadrant of the mount we shined the light on, with  no leak from certain areas and spectacular leaks from others.

 

Here’s what we found:

  • We tested several A7rs and several A7s and all were the same, so it’s simply a design issue, not a batch of defective cameras.
  • The light leak occurs with the body cap in place, but is more severe with a 35 f/2.8 Sony lens mounted instead of the body cap. 
  • The leak is worst when the light is shined onto the lens-release button area. It is also bad when the light shines onto the lower left quadrant of the lens mount, and directly above the lens mount. Other areas either didn’t leak or showed only a thin line that was far less pronounced than these areas. 
  • Wrapping something around the lens mount stopped the left side leaks, but not the leaks around the lens-release button.

Taking the metal mount off of the lens gives us a bit more information. The red lines show the areas of greatest light leak. Obviously one is around the lens release button. There’s a gap under the mount at the top that may explain why that area has more of a leak. The lower left area doesn’t show any obvious explanation as to why it might be worse, though.

 

 

We went back to our test lights and placed the fiberoptic lights over the lens-release button and screened them from the rest of the camera to see if we could actually see the light leak in this area. We were a bit surprised to find there was almost a direct reflection down onto the sensor. This image isn’t the greatest, but you can clearly see the light that was shined onto the button (it’s behind the black mask at the top of the picture) is reflecting back off of the sensor assembly (violet and green bars).

 

Quick Fixes

Obviously putting a scrunchy around the lens mount, as Ferrell McCollough has suggested, helps a great deal. My thought, after looking at the front with the mount removed was that putting some optical black on the back of the mount might reduce reflections from the aluminum, so we tried that. Epic fail – it made no difference at all.

We assumed a weather-sealed lens, with a rubber gasket, would work well, too. But we don’t have any weather sealed FE lenses available. So Aaron decided to make a bit of weather sealing around the lens mount, using some single-side adhesive rubber material (electrical tape) we keep in our high-tech lab. A bit of time working under a magnifying loupe with a scalpel and forceps and he had made a nice weather seal.

 

Once it was placed back on the camera, lenses mounted nicely and the light leak from the top and left sides was almost totally gone. It didn’t do doodly squat about the light leak around the lens-release button, as we expected.

Here we were faced with a conundrum. We could completely disassemble the camera to get to the lens release mechanism and hope some obvious solution became apparent (this was Aaron’s initial vote). But about this time Officer Obie wandered into the repair area and made some ‘suggestions’ about us actually fixing some stuff that was awaiting repair. We told him, in two-part harmony, that we had been tasked with a Holy Quest and explained the ISO 25,000, shooting into the sun, 30-second exposures, fiberoptic light testing, and the special rubber lens-mount weather seal we’d made. He just said, “Kids, get back to work.”

So, instead of completely disassembling the camera, Aaron invented the ‘lens release button light seal’, again using some special flexible lightproof plasticized sealant patches (electrical tape) we had.

 

I wouldn’t call it elegant, but it worked just fine. There were absolutely no light leaks now and once we trimmed the tape back off of the lens, the release button worked fine through the tape.

Practically speaking, I don’t recommend making your own weather seal around the lens mount. It takes a while and you have to take off the front mounting plate so your warranty is voided, etc. A scrunchy or such will work just fine for those times when you just have to shoot at ISO 25,000 for 30 seconds, take long exposures facing the sun, etc. If the thing you use doesn’t cover the lens-release button, then a bit of electrical tape will handle that just fine, too.

And for you Fanboys who just can’t wait to Sony bash over this odd little issue, let he whose camera is without light leak cast the first stone. Because, so far the same problem has been found with Canon, Nikon, and Fuji cameras. I assume a similar solution will work just fine for those, too.

 

Roger Cicala and Aaron Closz, the boys from the Group W bench

Lensrentals.com

March, 2014

 

Meaningless Aside:  Guthrie’s suggestion was that draftees, during the psychiatric evaluation part of induction, simply walk in, say, “You can get anything you want, at Alice’s Restaurant,” and walk out. I view a lot of forum threads the way he viewed the draft evaluation. If you’ve never heard “Alice’s Restaurant Massacree” and have 18 minutes to spare, it’s pretty funny even today.

Author: Roger Cicala

I’m Roger and I am the founder of Lensrentals.com. Hailed as one of the optic nerds here, I enjoy shooting collimated light through 30X microscope objectives in my spare time. When I do take real pictures I like using something different: a Medium format, or Pentax K1, or a Sony RX1R.

Posted in Equipment
  • RawheaD

    Roger; Imaging Resources found they all had leaks at those extreme numbers (ISO25,600, 30sec, broad daylight). They have not tested much more reasonable, “real-world” numbers. Like I said, I’m seeing the leak at 76s, ISO50.

    Graham Clark has images up that were ISO50 @ 30sec. Those are very very modest numbers, and yet the leak is there. Are you really going to suggest that (1) that is not an issue, and (2) all other cameras are going to fare the same?

    http://www.grahamclarkphoto.com/sony-a7r-field-notes-with-canon-ef-lenses/

  • Robert

    Roger, I’ve read about these, and I believe (have not thoroughly tested, yet) that I actually have a real-world issue with this light leak.

    I do astrophotography; taking 2-5min (or even longer) exposures at 400-1600 ISO, with my camera mounted to the back of a telescope or other camera lens. I ran into a problem where I was seeing strange rainbow arcs in my image, but as I was trying out a new piece of gear which dropped a prism into the imaging area, I figured I may be getting stray reflections off it or something. After seeing this, however, it looks very much the same.

    I’ll see if I can reproduce this with a standard mounting and no little fiddly bits potentially interfering with the light. I’m not sure, as it’s usually relatively dark around the telescope, but I’m also taking very high contrast images with a normally black background, for very long exposures.

    I appreciate the potential solutions, as it’s likely I’ll need them. This and the supposed RAW lossy compression have me worried about using my A-7R for astrophotography. And the fact that getting a bulb mode intervalometer for this camera is more difficult/expensive than it should be.

  • The song, like many of my posts, was criticized for being overly long.

    As I said on Reddit, I find your posts to be too short. Like The Lord of the Rings, I always wish them to be longer than they are!

    Right now, my comment is the top-rated on Reddit, which indicates that I’m not the only one who feels this way.

    You’re already quite good about putting an abstract / summary / findings section at the top of most posts; just keep doing that and those who don’t revel in detail and knowledge can stop there, while the rest of us keep going.

  • Paul Tawerll

    Thanks for these tests.
    This means that the Great Leica debate of the late 50’s has been resolved. Screw mount (SM) lenses are better than bayonet mount(M). SM will have no light leaks. This debate occurred during the M3 and lllg period.

  • Peter Kelly

    When these diehard ‘find-a-fault-at-all-costs’ individuals come to take their strange pictures, why is it they are happy to arrange all manner of ND filters, adjust their cameras on heavy tripods to eliminate camera shake, but can’t be bothered to cover up with a black cloth?

    The solution was used since the beginning of photography, but now they have to rant about nothing?

    Bizarre!!!

  • I had already put black electrical tape around the mount because I cannot stand the colour of that orange red ring, so I am fine. More or less.

  • Roger Cicala

    L.P.O – if 50 people a day do that, it should end all Fanboism!

  • Roger Cicala

    Todd, I was going to do that with at least one of the photos. But the internet has a seeing eye dog.

  • Roger Cicala

    Rawhea, Imaging Resource found D800, 5D Mk III, and Fuji XT all did I believe.

  • Roger Cicala

    Iggy, I haven’t checked them. I suspect Imaging Resource will be doing that in their Part III article.

  • Roger Cicala

    Thanks Chr K. That makes sense – saw it this am.

  • Chr K

    “The lower left area doesn’t show any obvious explanation as to why it might be worse, though.”
    Well, certainly the mount has 3 lugs and 3 “spaces” in between (just like most camera bayonet mounts). And you found 3 sensitive areas. – Do you really think this could be a coincidence without significance? 😉

    From the pictures shown at a Korean forum it seems like the silvery metal ring is a tiny bit thinner on the A7/A7r mount than on the NEX 7. But the underlying black plastic ring seems to be identical. I suspect Sony somehow forgot to use a new revised and matching black plastic “sealer” ring when they assembled the A7/A/r.

  • Brian

    You can get anything, except Alice and 30 second exposures at ISO 25,600 shot in daylight.

    There was also a full-length movie starring Arlo Guthrie released in 1969.

    Well, it’s not as bad as a Kiev.

  • My experience is that it happens on lower iso like iso 1600 as well. At least my A7 does.

    Most e-mount parts are xcross manufactured and used on other models. What about the Nex7, Nex6, and the newer ones like the a5000, a6000. The issue will be less obvious due to the sensor size. Have you any clue???

  • L.P.O.

    So, you’re suggesting that, just in case we ever encounter a useless and/or fanboyish camera argument on the Internet (not that I think it would ever happen in real life because they are so rare), then we’d just need to say: “You can get anything you want at Roger’s restaurant” and run the hell away?

    (Yes, I listened to the whole song. And no, though I knew Arlo Guthrie from Woodstock, I didn’t know this particular song.)

  • RawheaD

    I hate seeing this “ISO25,600, 30 sec, broad daylight” meme being beaten to death. I originally came up with the test so that people can quickly check where the source of the light leak may be, indoors, with nothing but a flashlight. It’s a dead horse. Please stop.

    My A7R leaked light doing daytime LE with a 76 sec. exposure at ISO50.

    I would actually like someone to show me a pro-level camera that consistently leaks light from the flange area under those conditions. Because I’ve been doing LE for years, and I’ve never had that problem with any of my previous cameras (K20D, K7, 5D, 5Dmk2), even at longer exposures (up to 10-20 minutes).

  • Great article Roger. It probably would have been clearer with 27 8×10 glossy pictures with circles and arrows and a paragraph on the back of each one though.

  • Roger Cicala

    Roger, I couldn’t say but since it affected all 6 we looked at I think it’s a design issue, not an assembly issue. I guess it could be a batch of bad parts since the mounts are the same in both cameras.

  • Roger Cicala

    Ken, you and I have the same point of view. And since I have no hair, I really don’t have an excuse for carrying a scrunchy. Not that I can ever imagine needing one for this issue – I don’t do that kind of long exposure. Not enough patience.

  • Roger Knight

    Roger, As usual, great work and enjoyable reading. Is this problem possibly due to slight deformation, between the lugs, of the lens/body mount when a lens or cap is mounted. I imagine the light leak would be horrendous without a lens or cap mounted, even at ISO 100. Aloha.

  • Sam

    Dust and moisture resistance
    Carefree shooting in tough environments is yours thanks to comprehensive dust and moisture resistance measures that enhance reliability by helping to prevent water and dust from entering the body. These measures include sealing around the buttons and dials, as well as a protective double-layered structure that tightly interlocks panels and components.

    Yet light shines through…

  • Ken Sky

    So let’s be practical: I take my tripod-mounted A7r into a noon-day sun to photograph a landscape. Can I expect to see a light leak on my files? Oh, BTW, the last time I used ISO 25,000 with 30 sec. exposure was……never! Is there any practical importance to this finding? If so when? I’ve stollen a scrunch from my wife drawer and stuck it in my bag but could return it before she finds out;)

  • Aaron

    Well, I could see how this could affect a few people who shoot in broad daylight with a 10-stop ND filter. On the other hand…I don’t know how many people shoot with that getup. Probably not many.

  • Roger Cicala

    Zig, I’m not sure I understood the first paragraph, but the second paragraph is a perfect summation of things 🙂

  • Siegfried

    Roger,
    I read it twice but still don’t get it: how come this is about sensitivity? I mean we’re talking about non-transparent parts made of metal and plastic several millimeters thick and it requires a hack of nuke to make them transparent. If there’s a hole, then light comes in and shines at whatever ISO setting currently set.
    Or it’s about indirect reflection (geeky speaking: n-th derivative reflection) which Sony engineers 99,32% suppressed and one has to push the pedal to the metal to get the remaining 0,68% part?

    Thanks,
    Zig

  • Roger Cicala

    Thank you, Mike. My pressure isn’t that bad. I’d freeze up if I knew I had to have it done by xxxxx. I used to ghostwrite for textbook publishers and the stress about killed me. Then again, my insistence on slipping something amusing in here and there about killed them.

    Roger

  • MikeO

    Wonderful, balanced, thoughtful AND useful article. Your long-form technical writing is very good, and I love it – and would like to see more. I wrote a column, nominally on computer protocols, every month for twelve years, and I know what it is like to be funny on deadline. As this is a blog, you don’t have that pressure, but there are damn few people who can write humorously about highly technical stuff, and I prize your ability.

  • CarVac

    I wanna see blood and gore and [camera] guts and wires in my teeth. Eat dead bricked [A7] bodies. I mean kill.

  • Great article, Roger. Thanks for posting.

  • Speed

    Injecting reason. Will the internet survive?

Follow on Feedly