D7000: strange test results

Joined
Nov 23, 2008
Messages
117
Location
Australia
Since exchanging my backfocussing D7000 for a new copy I have become a bit obsessive about testing the AF to make sure this copy is OK.

I've tried cereal packets, Moire patterns and focus charts (OK, maybe very obsessive).

With my focus-chart tests, I noticed one side of the image behaved differently to the other. So I re-did these tests paying special attention to the placement, height and angle of the camera.

This time, all three of the lenses I tested produced similar results: the left side of the image was slightly backfocussed compared to the right side. For instance:

  • The LHS of my 35 f1.8 needed an adjustment of -8 while the RHS needed -4.
  • The LHS of my 17-50 f2.8 needed an adjustment of +1 while the RHS needed +5.
  • The LHS of my 35 f2 needed an adjustment of -4 while the RHS needed -1.

For now, I have set the AF fine tune for each lens to a compromise adjustment between the two settings.

Assuming my testing is OK & this pattern continues with my other lenses, any thoughts on what may be causing this?
Is this symptomatic of a misaligned main sensor?
Is this grounds for exchanging the camera, or am I being over-critical given the margins involved?
 
Joined
Oct 2, 2010
Messages
131
Location
Des Moines, Iowa
My 1st thought is, the lenses have slightly decentered elements but I'm not an optical engineer. I doubt it's the camera this time. There's no consistency in the results.

My suggestion is to take it outside and compare liveview to autofocus on a flat, high contrast subject that is parallel to the lens and 50x greater or more than focal length being tested and then see what the results are. High shutter speed and low ISO are mandatory. If the results are very similar between liveview and autofocus, it's definitely not the D7000. Given the minor differences, you may be trying to achieve the unobtainable.

My $.02 worth...

Jim
 
Joined
Mar 23, 2008
Messages
3,129
Location
Santa Barbara, California, U.S.A.
You need to stop obsessing about this and go out and take some pictures :smile:

How far away is your focus chart? If it's as close as I think, any minute differences in camera angle will show up as sharpness variations from side to side! Remember that DOF is very shallow at these short distances - if the center of the image is in focus, the edges are already outside the DOF.

Try shooting something at longer distance - if the images are still softer on one side, you might have an issue. But I highly doubt it!

Cheers

Mike
 
Joined
Aug 15, 2010
Messages
2,967
Location
Sydney Australia
Spend a day taking some real world shots and see how it looks. If your still not convinced try testing again at some longer distances as the others have suggested.
 
Joined
Nov 23, 2008
Messages
117
Location
Australia
My 1st thought is, the lenses have slightly decentered elements but I'm not an optical engineer. I doubt it's the camera this time. There's no consistency in the results.

Jim

Well the results are pretty consistent- the LHS backfocusses relative to the RHS on all 3 lenses tested so far. This difference can even be measured at around 3-4 AF adjust units.
 
Joined
Oct 2, 2010
Messages
131
Location
Des Moines, Iowa
When I said it wasn't consistent, I meant that the left side is not always worse than the right side based on your post. Also, minus settings mean it's front focusing and plus means it's back focusing. See the problem?

The LHS of my 35 f1.8 needed an adjustment of -8 while the RHS needed -4.
The LHS of my 17-50 f2.8 needed an adjustment of +1 while the RHS needed +5.
The LHS of my 35 f2 needed an adjustment of -4 while the RHS needed -1.
You've heard this from a lot of people including me but quit testing inside under artificial light at close distances. Outside in bright natural light with fast (above 1/250) shutter speeds, low ISO and on a tripod with mirror lock up to eliminate camera shake. Focus on a parallel high contrast subject at least 50x greater distance than the focal length you're testing. Now do the same thing with liveview. Compare the results. They should be very similar.

It's really pretty easy. It also eliminates people like me from telling you that your test didn't give reliable results. :wink:

Good luck!

Jim



Well the results are pretty consistent- the LHS backfocusses relative to the RHS on all 3 lenses tested so far. This difference can even be measured at around 3-4 AF adjust units.
 
Last edited by a moderator:
Joined
Nov 10, 2008
Messages
3,969
Location
Chicago
Think you are perhaps expecting too much.

If you think not, there are D3S and D3X models and pro lenses made to higher standards and Leica if you have the money.

As suggested,set up a test target with a camera on the tripod with the lens axis perpendicular to the target. This is very critical to get perfect. Remember your geometry and similar triangles. Set up a center point lens high and mark the wall with chalk or tape.

When all said and done, this is a consumer product made to a price. Perfect lenses come from other places like the Coast 60mm. You really do not want to pay for for them.

I would agree the problem lies with the lenses more than the body, but you need some way to establish it and it could be a combination of both. I doubt you will achieve perfection.
 
Joined
Sep 15, 2008
Messages
1,416
Location
Sacramento
Well the results are pretty consistent- the LHS backfocusses relative to the RHS on all 3 lenses tested so far. This difference can even be measured at around 3-4 AF adjust units.
The problem with a lot of (D)SLR users is they never experienced the joys and troubles a view camera. Even if your lens is perfect, if the lens mount is not parallel to the film plane, you get visible L->R or T->B out of focus issues. Been there, done that.

The OP's results show that the camera body lens mount is not parallel to the sensor plane of the camera. If it was the lenses, then the L->R side differences should be nearly zero. Each lens may have some overall BF or FF due to manufacturing tolerances, but all 3 having a measurable and consistent left to right side difference of 3-4 units can only mean that the test was seriously flawed or else the camera lens mount flange needs adjustment. Once you get the camera's lens mount adjusted so it is parallel to the sensor, your AF accuracy should be greatly improved.

For those that doubt it is the camera body, take a piece of paper and draw a big cross on it. Put -20 at the top, +20 at the bottom, and zero in the center. On the left side, a bit away from the vertical center-line of the cross, mark the -8 spot and on the right side of the center-line do the same for the -4 position. Draw a line from one to the other and label it 35-1.8. Do the same for the other two lens result sets, labeling each by lens.

You should end up with 3 lines roughly parallel to each other. All 3 lines will have the left side starting closer to the -20 end of the cross and the right side ending closer to the +20 end. All 3 lenses have different amounts of AF tune needed, but the 3 nearly parallel lines clearly displays a consistent left to right miss-alignment of the camera body lens mount.

Gnik, if you have a screen door or screen window handy, set it on a bench so it is parallel to the floor and put a small object in the center. Put your camera on a tripod down near the plane of the screen, so that you can shoot with all 3 lenses and focus on the object using max aperture. For this type of issue, you can use AF, manual focusing, or manual focusing via LiveView. You should be able to easily see the area in focus and the DOF will be an obvious line of the screen that is in focus and out.

If the lens mount was correct, then the center focus point obtained with AF might be in front or just behind the object on the screen, but an area of "in focus" screen would extend from the focus point out to the sides in a fairly straight line. Given the values you posted, the in-focus area of the screen will be a line that starts further away on the left and moves closer to the camera on the right.

It is important to make a jpeg example from all 3 lenses to demonstrate it is repeatable with different lenses. Then contact Nikon Support and either email the images or burn to a CD and send the CD in with the camera.
 
Joined
Nov 23, 2008
Messages
117
Location
Australia
The problem with a lot of (D)SLR users is they never experienced the joys and troubles a view camera. Even if your lens is perfect, if the lens mount is not parallel to the film plane, you get visible L->R or T->B out of focus issues. Been there, done that.

The OP's results show that the camera body lens mount is not parallel to the sensor plane of the camera. If it was the lenses, then the L->R side differences should be nearly zero. Each lens may have some overall BF or FF due to manufacturing tolerances, but all 3 having a measurable and consistent left to right side difference of 3-4 units can only mean that the test was seriously flawed or else the camera lens mount flange needs adjustment. Once you get the camera's lens mount adjusted so it is parallel to the sensor, your AF accuracy should be greatly improved.

For those that doubt it is the camera body, take a piece of paper and draw a big cross on it. Put -20 at the top, +20 at the bottom, and zero in the center. On the left side, a bit away from the vertical center-line of the cross, mark the -8 spot and on the right side of the center-line do the same for the -4 position. Draw a line from one to the other and label it 35-1.8. Do the same for the other two lens result sets, labeling each by lens.

You should end up with 3 lines roughly parallel to each other. All 3 lines will have the left side starting closer to the -20 end of the cross and the right side ending closer to the +20 end. All 3 lenses have different amounts of AF tune needed, but the 3 nearly parallel lines clearly displays a consistent left to right miss-alignment of the camera body lens mount.

Gnik, if you have a screen door or screen window handy, set it on a bench so it is parallel to the floor and put a small object in the center. Put your camera on a tripod down near the plane of the screen, so that you can shoot with all 3 lenses and focus on the object using max aperture. For this type of issue, you can use AF, manual focusing, or manual focusing via LiveView. You should be able to easily see the area in focus and the DOF will be an obvious line of the screen that is in focus and out.

If the lens mount was correct, then the center focus point obtained with AF might be in front or just behind the object on the screen, but an area of "in focus" screen would extend from the focus point out to the sides in a fairly straight line. Given the values you posted, the in-focus area of the screen will be a line that starts further away on the left and moves closer to the camera on the right.

It is important to make a jpeg example from all 3 lenses to demonstrate it is repeatable with different lenses. Then contact Nikon Support and either email the images or burn to a CD and send the CD in with the camera.

Thanks Mike. That all makes sense and aligns with what I am seeing (i.e. what should be a horizontal focus plane from L-R is showing as slightly diagonal across the focus chart).

I guess my lingering question is whether or not to go thru the pain of another attempted repair after the drawn out battle with my previous back-focussing copy (another story).

Is adjusting the camera's lens mount usually a simple & uncontroversial repair?
 
Joined
Sep 15, 2008
Messages
1,416
Location
Sacramento
Thanks Mike. That all makes sense and aligns with what I am seeing (i.e. what should be a horizontal focus plane from L-R is showing as slightly diagonal across the focus chart).

I guess my lingering question is whether or not to go thru the pain of another attempted repair after the drawn out battle with my previous back-focussing copy (another story).

Is adjusting the camera's lens mount usually a simple & uncontroversial repair?
I'm not a trained camera repair tech, but compared to repairing an AF zoom lens, it is a relatively simple process for most modern shops. The lens mount is screwed onto the camera body at the factory and tested for being parallel to the sensor, both L-R and T->B.

A tech puts the body on an optical test bench, runs some tests, and their equipment figures out the offset needed to get the mounting plate back to being parallel to the sensor. The mounting plate is removed and either a new one put on, or some shims are added, removed, or exchanged, and the original mounting plate reattached. As is true of any repair, the skill used by the tech to ensure the fix is done correctly is the key. A good tech ensures that their diagnostic tools are properly calibrated and then double checks the repair afterwords. Because it is a relatively simple repair, some techs will get the results, make a repair, and not retest to ensure optimal results. ALWAYS, ALWAYS, retest when a camera/lens returns from the shop.

This problem occurs a lot more often than most people would suspect. While in use, a long or heavy lens hanging off the body or being bumped can cause the lens mounting plate to become no longer parallel due to distortion or loosening at the attachment points. If it is noticeable with only one lens, it a problem with that lens. If it is noticeable with multiple lenses, it is the camera body. My out of warranty D300 has a similar but slight BL->TR issue. That isn't as bothersome as L->R with the subjects I use the D300 for, but it will be sent in for repair. I don't like using gear where I can see in an image that it is out of spec.

When I worked as a manager/photog for a studio many years ago, most of our work was done outside the studio. We used Nikon 35mm film bodies as well as medium format and occasionally a 4x5. The Nikons got bumped and banged a LOT. Every year we would drop them off at a local repair facility for cleaning and testing. We would usually have at least one that needed a lens mount repair/adjustment. Any body or lens that had a repair done was always tested by us with a 12 exposure roll of film before being put back in the "available to use" pool. Again, ALWAYS, retest when a camera/lens returns from the shop.

Since your body is still under warranty, if the problem doesn't severely affect your photos, you might wait a while and keep shooting. I would however, be sure to send the body in a couple months before the warranty period expires. Have it adjusted, tested, and cleaned, for free under warranty. That's why you paid more for the USA version versus gray market.
 
Last edited by a moderator:
Joined
Mar 23, 2008
Messages
3,129
Location
Santa Barbara, California, U.S.A.
Thanks Mike. That all makes sense and aligns with what I am seeing (i.e. what should be a horizontal focus plane from L-R is showing as slightly diagonal across the focus chart).

I guess my lingering question is whether or not to go thru the pain of another attempted repair after the drawn out battle with my previous back-focussing copy (another story).

Is adjusting the camera's lens mount usually a simple & uncontroversial repair?

Before we can reason about misaligned lens mounts, we need to look at alignment of the focus chart. How do you guarantee that the focus chart was correctly aligned to the camera axis?

Cheers

Mike
 
Joined
Nov 18, 2008
Messages
1,211
Location
HHI
Before we can reason about misaligned lens mounts, we need to look at alignment of the focus chart. How do you guarantee that the focus chart was correctly aligned to the camera axis?

Cheers

Mike

He can't!! I guess he could mount the chart on a fixture that is perfectly square to the sensor. NOT!!! You people kill me! In the real world it doesn't mater. GH
 
Joined
Nov 23, 2008
Messages
117
Location
Australia
OK...

After reviewing all my test pics from the last couple of weeks (focus charts, flat planes & real-world pics) I have come to the conclusion that the camera is fine (phew), but I have two lenses with problems. Specifically:

My Tamron 17-50 f2.8 consistently shows the characteristics described in the OP (i.e. lens apparantly misaligned with sensor). After rationalising my test pics, I have re-estimated the misalignment between LHS & RHS at around 6 AF adjust units (e.g. if centre is perfect, LHS requires +3 & RHS requires -3).

Am I right in saying this is likely to be a misaligned element? Or maybe a problem with the lens mount?

(Interestingly, I have long suspected focussing issues with this lens. I sent it in for a check-up late last yr but it came back with a clean bill of health (though I have since lost confidence in this repairer)).

Also the LHS of my 35 f2 is just plain soft wide open compared to the sharp RHS. This caused me to mis-read the problem as being similar to what I had seen with the Tammy (i.e. LHS backfocussing compared to RHS).

Anyway thanks for all the advice. I am learning a lot more than I ever expected to need to know.
 
Joined
Jul 4, 2007
Messages
3,551
Location
Redmond, WA
All lens are decentered to some degree -- it may be incredibly microscopic fractions that never show up in even test photos, or it may be very pronounced. It comes down to whether or not it's affecting real-world shots. BTW, VR applies some controlled decentering.

Decentering in a Tamron 17-50 isn't unheard of.
 
Joined
Aug 15, 2010
Messages
2,967
Location
Sydney Australia
OK...

After reviewing all my test pics from the last couple of weeks (focus charts, flat planes & real-world pics) I have come to the conclusion that the camera is fine (phew), but I have two lenses with problems. Specifically:

My Tamron 17-50 f2.8 consistently shows the characteristics described in the OP (i.e. lens apparantly misaligned with sensor). After rationalising my test pics, I have re-estimated the misalignment between LHS & RHS at around 6 AF adjust units (e.g. if centre is perfect, LHS requires +3 & RHS requires -3).

Am I right in saying this is likely to be a misaligned element? Or maybe a problem with the lens mount?

(Interestingly, I have long suspected focussing issues with this lens. I sent it in for a check-up late last yr but it came back with a clean bill of health (though I have since lost confidence in this repairer)).

Also the LHS of my 35 f2 is just plain soft wide open compared to the sharp RHS. This caused me to mis-read the problem as being similar to what I had seen with the Tammy (i.e. LHS backfocussing compared to RHS).

Anyway thanks for all the advice. I am learning a lot more than I ever expected to need to know.

Glad to see you are now sure the body is good

Now go out, enjoy it and lets see some pics :biggrin:
 
Joined
Nov 23, 2008
Messages
117
Location
Australia
For me, the crux of the matter is this: do any of the so-called deficiencies brought to light in your test pictures get in the way of you accomplishing your photographic goals? If so, do something about it. If not, stop doing the test pictures and get on with accomplishing your goals.

My 1st D7000 had a significant backfocussing problem that did 'prevent me from accomplishing my photographic goals'. The journey I took to pin this down was long and painful. I lost confidence in both the camera and in the local repairer and questioned my own abilities to boot.

I have now had a replacement copy for a couple of weeks. I'm testing the hell out of this copy in as short a time-window as possible because a) I don't want to go thru the same protracted pain again and b) so I can restore confidence in my camera and enjoy using it asap.

As a by-product of this testing, I'm seeing problems in a couple of my lenses which accounts for some of the inconsistencies I've previously noticed in 'real world' pics.

So, yes the testing has been a bit dull, but useful nonetheless.

And I'm about over it now!

Edit: also I've come to realise how important it is to go thru the process of AF fine tuning all your lenses when you get a new body. I don't know how others do this, but for me this involves a fair bit of testing...
 
Last edited by a moderator:
Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Nikon Cafe is a fan site and not associated with Nikon Corporation.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Forum GIFs powered by GIPHY: https://giphy.com/
Copyright © Amin Forums, LLC
Top Bottom