Inconsistent exposure with same settings--why?

Question

I have a Nikon D7100. I fired off a burst of shots at 6 fps at a static outdoor scene, all shots using the same settings: 1/125 s, f/7.1, ISO 100. Some of the shots are clearly darker, some are clearly brighter. While variation is not extreme, it is also apparent in the histogram. Why did this happen? Does it indicate that something is failing?

The temperature was about -5 C and sunny. I don't think the camera actually cooled to below 0 C, but the only thing I can think of is that the cold made the aperture actuation unreliable.

I tried again later at home, 1/30 s, f/7.1, and all the exposures were identical this time.

---

Update: Here's a test I did with the exact same settings at room temperature: Dropbox link. Please click a thumbnail and use the left or right arrows to go through the pictures and observe the brightness variation.

Note: I know the exposure is not good on these but this was a test and I intentionally used the precise same settings: 1/125 s, f/7.1, ISO 100. EXIF data is left intact.

Note 2: I tried with a different lens, the other lens doesn't show the variation.

Request: At this point I am getting quite worried and I would appreciate it very much if someone could do the same test with the same lens. It is the Nikkor 12-24mm f/4G.

---

Update 2 I did more tests with the problem lens. All these tests are at room temperature:

• f/4 (max aperture) 1/125 s -- no problem
• f/10, 1/125 s -- no problem
• f/7.1, each of 1/30, 1/60, 1/125 and 1/500 all show the problem. Last night 1/30 was fine, today it's not.

Answer 1

One of the decisions Nikon made when Autofocus first became a viable technology on a consumer scale about 25 years ago was to create a system that allowed it to be backwards compatible with the existing Nikon F-mount lens lineup. This meant maintaining a mechanical link between the camera body and lens to stop down to the aperture setting selected on the lens' aperture ring at the instant just before the shutter opens.

Since metering and focusing (manual or automatic) are typically done with the lens at its maximum aperture, the amount of time needed to actuate the aperture directly impacts the amount of shutter lag the camera displays. In very cold environments mechanical devices machined to tight tolerances can encounter substantially more friction than in warmer environments, especially if lubricants which begin to stiffen at certain temperatures are used.

Since you haven't indicated a specific lens in your question, I'll have to make an assumption: That you are using a D-type lens with an aperture ring on the lens. Even when the aperture is locked to the narrowest setting (usually denoted in orange) and controlled by the camera rather than the aperture ring, the mechanical linkage between the camera and lens will hold the aperture fully open until the shutter button is pressed. Then the link will move to allow the lens to stop down to the selected aperture an instant before the shutter begins to open.

On can observe the closed aperture varying slightly in shape, probably due to "bounce" of the mechanical levers connecting the lens to the camera body, as the sensor is progressively exposed in this super slow-motion video of a Nikon D3 with an AF Nikkor 50mm f/1.4D set to f/16 and 1/4000 second cycling at 11 fps.

(screenshot from video)

I suspect what has happened is that the cold environment inhibited the movement of the aperture control linkage just enough that it had not completed its movement to f/7.1 when the shutter opened for some or all of the exposures in question. The camera should have enough time to complete this movement prior to the shutter actuation when the camera is in an environment within the bounds of the manufacturers rated operating environment. The minimum rated operating temperature for the D7100 is 0ºC/32ºF. The differences from one frame to the next in terms of exposure indicate differences in the amount of time the camera needed to set the aperture at f/7.1 via the mechanical link when operating in a cooler environment.

Answer 2

It's not your camera or lens. It's an issue with the mechanical aperture. This happens at some aperture settings and fast burst rates. It's inconsistent and unreliable. That's the only reason I stay away from Nikon.

Answer 3

Here, quoted from "Nikon F vs Canon EF — What's the Difference?" at photographylife.com:

Nikon F – Mechanical Diaphragm Lever

One of the biggest disadvantages of the Nikon F mount is the mechanical diaphragm / aperture lever that is present on most Nikon lenses. Whether you are looking at a classic manual focus Nikkor, an older “D” or the newer “G” type lenses, all of them require Nikon camera bodies to physically change aperture on every shot if it is set to anything other than maximum aperture. That’s because all such Nikon lenses contain a mechanical lever on the rear side of the lens, which must be engaged to adjust the aperture. When a lens is dismounted, the spring-loaded lever on the lens is pushed back to its standard position, which basically stops the lens down to its minimum aperture. Once you start attaching the lens to a camera body, the corresponding lever inside the camera chamber forces the lens to open up the diaphragm, as illustrated below:

Lenses typically stay wide open at maximum aperture when mounted to cameras, for maximum amount of light to reach the viewfinder and the phase detection autofocus system. Hence, aperture on DSLRs only changes right before the exposure. Once a picture is taken, the lever goes back and the diaphragm mechanism returns to its wide open state to continue providing maximum amount of light to the camera. This means that when shooting with lenses that feature such mechanical levers, the lens must physically stop down and open up every time the camera fires. Since the mechanical lever is physically triggered by the camera, this mechanism must be extremely precise and accurate in order to yield consistently accurate brightness and desired depth of field. However, when shooting continuously in high speed, it is often impossible to yield consistent results, since the mechanical lever might not have enough time to go back and forth quick enough. And if the lever is not precisely calibrated, or potentially wears off / malfunctions overtime, each shot might yield incorrect aperture and brightness.

In addition to the above, lenses with mechanical levers are hard to adapt with other systems via third party adapters. If you have been wondering why adapters for Nikon lenses are hard to use and do not give complete and precise aperture control, now you know why – other manufacturers simply would not have the same lever control mechanism in their camera bodies. An adapter capable of mechanically moving a lever would require a motor with an electronic chip, which would make the solution quite cost-prohibitive.

In contrast, lenses that feature electromagnetic diaphragms do not have any mechanical levers – changes in aperture are communicated electronically by the camera through lens contacts. Such method of aperture control is much more preferred, because lenses can set their apertures consistently and accurately, with no shot-to-shot variation.

Because of the above, using a mechanical lever to change aperture is prone to inconsistency in exposure and potential mechanical issues both in camera and in lenses. Canon realized this and fully moved to electronic aperture control on both EF and EF-S mounts a while ago, and Nikon has only recently started updating its lenses to “E” type lenses with electromagnetic diaphragm mechanism. Unfortunately, such lenses have been limited to mostly super telephoto and higher-end zoom lenses, so despite the obvious disadvantages, Nikon has still been releasing many modern “G” type lenses with mechanical levers.