7D2 Metering Issue After Two Repairs

I don't feel it's the shutter. If you know the correct exposure for a scene and you plot it in while in Manual Mode, it'll work just fine.

EXIF as DPP sees it. Just an easy suggestion, turn off the Auto Lighting Optimizer and see what happens. This will remove at least one step that allows the camera to make an exposure decision.

HOSTED PHOTO
please log in to view hosted photos in full size.

In looking at the EXIF data, we find this:

Measured EV -2.75

That tells us everything we need to know. The camera's meter measured -2.75 EV from the scene. To put that in perspective, 0 EV is the exposure you'd get with 1 second at f/1. f/4 is 4 stops past that, so you'd need a shutter speed of 16 seconds at f/4 just to get 0 EV. -2.75 EV is, of course, 2.75 stops darker than that, which means you'll need an additional 2.75 stops of shutter opening time at f/4 to get the exposure the meter says is needed.

2.75 stops is 6.7 times (i.e., 2 raised to the power of 2.75). 6.7 * 16 seconds is 108 seconds, or 1 minute 48 seconds.

The ISO is 500, which reduces the needed exposure time in the above to 1/5 of its original value, which gets you 21.6 seconds. 20 seconds is the closest shutter speed to that, which is why the camera chose it.

The problem is unquestionably your camera's meter.


EDIT: I said "2.75 stops reduction of shutter opening time" in the above previously, and that's quite clearly incorrect. My intention was to indicate that the shutter has to be open longer, by 2.75 stops. I've also tried to add a little more clarity, though it sounds like people already understood what I said.

Yep, the meter is at fault:

HOSTED PHOTO
please log in to view hosted photos in full size.

kcbrown wrote in post #17638398
In looking at the EXIF data, we find this:

Measured EV -2.75

That tells us everything we need to know. The camera's meter measured -2.75 EV from the scene. To put that in perspective, 0 EV is the exposure you'd get with 1 second at f/1. f/4 is 4 stops past that, so you'd need a shutter speed of 16 seconds at f/4 just to get 0 EV. -2.75 EV is, of course, 2.75 stops below that, which means you'll need an additional 2.75 stops reduction of shutter opening time at f/4 to get the exposure the meter says is needed.

2.75 stops is 6.7 times. 6.7 * 16 seconds is 108 seconds, or 1 minute 48 seconds.

The ISO is 500, which reduces the needed exposure time in the above to 1/5 of its original value, which gets you 21.6 seconds. 20 seconds is the closest shutter speed to that, which is why the camera chose it.

The problem is unquestionably your camera's meter.

Nice analysis with a key variable as the end point. I'd also like to say that I would suspect in this day and age, many camera techs don't understand EV and its ramifications.

Whoa Maths (runs and hides). No, honestly, that is really cool. If you haven't sent the camera in I would include all of this information in what you send with the camera. I wonder if something is blocking the metering sensor or if it has blocks of dead pixels or something.

That confirms it...

I will provide this information to them, and hope they finally are able to fix this permanently.
Do you guys think they'll need to replace the meter ? Why would it work for a day or two after each previous repair?

Just as a side thought...

have you tried several lenses or only the 24-105? Just in case there is something wrong with the lens... (although it doesn't appear to be, but may explain why canon tech doesn't find anything but then it acts up again)

kesterc wrote in post #17638943
That confirms it...

I will provide this information to them, and hope they finally are able to fix this permanently.
Do you guys think they'll need to replace the meter ? Why would it work for a day or two after each previous repair?

Hopefully the Canon guys will listen and understand. On the other hand, I'm pretty sure they can quickly determine themselves that the metering is the problem, and probably did so the previous two times, otherwise it would not have been returned fixed.

It might be the meter itself, or a cold solder spot in a connection, or anything in the chain to the camera's CPU. Considering what has already happened with this camera, it is probably not something as simple as a bad meter.

Reminds me of the story of a guy who bought a brand new car. It was running a bit rough. Took it to the dealer who did an adjustment. That didn't fix it, so he took it back. This time they changed the carburetor. Still didn't fix it. This went on for several weeks, with both customer and dealer becoming more frustrated. Finally in desperation the dealer replaced the engine. The mechanic proudly started the engine and opened the hood the better to be able to admire his work. And the engine still was running rough. He stood there in amazement for a couple of minutes staring at the rough-running engine. And then his drifting eyes noticed a little hose, just a little hose, that wasn't attached to anything. Hmm, would it go here? And finally that fixed the months-long problem. It was a loose vacuum hose.

A friend of mine assembled an expensive audio amplifier kit. When done, it wouldn't work properly. One stereo channel took all the power, the other one was almost dead. He couldn't see any cause, so asked me to look at it. It must have taken me three hours to finally spot the problem - one of the four power diodes in the power supply (assembled by the factory) was reversed. The amp sprang to life when it was corrected.

Most repairs are straight-forward, but a few are not. Those tricky ones can consume a huge amount of time. This gives both sides big problems. Canon can't afford to spend the time, so the repair might be unsuccessful again.

I think all you can do is put pressure on Canon to replace the body if the third repair fails again. From what I have learned in reading posts here, it might take a lot of persuading to get Canon to do this. Good luck.

yep, I would try another lens as well. Stuck aperture blades possibly giving a bad reading?

does really sound like a bad meter, but would check any sort of blockage.

Charlie wrote in post #17639064
yep, I would try another lens as well. Stuck aperture blades possibly giving a bad reading?

does really sound like a bad meter, but would check any sort of blockage.

The shot was at f/4, the lens can only go to f/4, how could the camera think that the lens could be 3 stops faster?

That said, if this has only occurred with one lens, then certainly it is mandatory to send in both.

It was occurring with all three lenses.

kesterc wrote in post #17638943
That confirms it...

I will provide this information to them, and hope they finally are able to fix this permanently.
Do you guys think they'll need to replace the meter ? Why would it work for a day or two after each previous repair?

Fortunately, the image itself tells the entire story. The EXIF records what the meter saw and the exposure parameters that were set. The exposure mode was automatic, so the camera was making all of the decisions for that. The chosen values follow directly from the meter reading, and are correct and consistent on that basis. And the image itself records the actual amount of light that was in the scene, which in this case is vastly greater than the amount of light the meter saw.

So if the Canon techs really understand how all these things relate to each other, the image should make it clear to them exactly where the problem lies: the meter.

But in the face of inconsistencies in that regard, it's likely there's an electronic fault of some kind. A broken wire that's making contact intermittently, a cold solder joint, or perhaps even a faulty sensor chip. Unless they're willing to go to the trouble of figuring out exactly where the fault lies, they'll either have to replace the entire metering assembly, or they'll have to replace your camera.

Or some foreign obstacle partly obscuring the metering sensor. The light sensor is located just inside and above the eyepiece in the camera, so it's far away from the actual sensor.

This camera has to still be under warranty, as it has been less than one year since the 7DII was released. Canon should not have to be convinced too hard to honor its warranty and either finally fix the thing, or replace it. Unless it was gray market....