D90 Backfocus/tracking issues with Sigma 70-200mm

[HoppyUK]

I've had a problem recently with a sigma front focusing

here's the link I recommended

http://focustestchart.com/focus21.pdf

2nd to last page, there is a focus chart
tripod, 45 degrees and test it out
my results weren't great recalibration required!

That test chart causes so many problems, often where none exists. Use it at your peril.

It is used at far too close distance.
It is at 45 degrees.
It is extremely critical and most people get it wrong through user error.

Use a much bigger, flat target, with a clear pattern, from a realistic distance. I use a line of three cerial packets - focus on the middle one, and angle the two either side, one in front and one behind, so you can see where the actual plane of sharp focus falls. No need for anything fancy (quite the opposite, in fact).

Edit: Sorry DizMatt, didn't mean to dis your well-intentioned post

 

[sammyd56]

I'd tend to agree with Hoppy here, the focus chart for me showed the lens to be slightly front-focusing if anything, but used in real life the back-focus is plain to see.

Following extensive testing, my Tamron 17-50mm is right on the money. The 55-200mm also, and tracks motion well too.

I dunno, I could swear it's the lens, but whitey tells me he never had problems with it. Surely this isn't the kind of thing that develops overnight. I'm at a loss for what to do really :shrug:

 

[HoppyUK]

From your OP, it looked like something was wrong with your kit. You've tested it again and the lens is still not right, but you say your other two lenses are fine.

If that is so, then the finger is pointing very much at the lens. I take your point about depth of field and it's valid. Bearing in mind what a pain it is sending stuff back for calibration, it's well worth trying it on your other camera. Or maybe pop into a shop and borrow a body for five minutes. Just point it out of the door at a street sign or something. If the lens is out, it will do it every time. They might even have another lens for you to do a quick A-B comparison.

It seems like Sigma are quite used to sorting out this kind of problem, but check with them first. They might want the camera body too.

 

[Phiggys]

It certainly looks like it could be the Sigma I had a similar problem with two 3rd party lens works perfect on my Fuji's wouldn't work on a D50 at all,but was fine on D70,D70s, D80,D200,D300,D1h,D2h,D2x,D3 but it doesn't like the D700.
I spoke to the manufacturer who said they hadn't had any problems reported of computability ? So I tried the 2 lenses in question on friends D50 & D700 and it gave the same problems.
I can only guess that it is a problem with Nikons software/firmware in those two models of camera.
Friend has also had a similar problem with his Canon and Sigma 180mm macro giving error warnings and refusing to focus.

 

[sammyd56]

Right, thanks folks, I'm definitely going to head into town tomorrow to try and get my hands on another body.

In the meantime though I have a quick question for the more technical-minded among you...

At a focal length of 200mm and a distance of approx. 8m, should the AF be able to distinguish between two objects 13cm apart?

(f/5.6 if that makes a difference, this is using the Nikon 55-200mm)

 

[HoppyUK]

Right, thanks folks, I'm definitely going to head into town tomorrow to try and get my hands on another body.

In the meantime though I have a quick question for the more technical-minded among you...

At a focal length of 200mm and a distance of approx. 8m, should the AF be able to distinguish between two objects 13cm apart?

(f/5.6 if that makes a difference, this is using the Nikon 55-200mm)

Your pic does not inspire confidence! What's wrong with the cerial packets suggestion (in good light!)? I think you may be chasing your tale here with some inconclusive tests.

D90, 200mm, f/5.6, 8m, depth of field is 0.35m. So both objects should be well within the DoF.

AF systems work to a tolerance and while I would expect them to generally differentiate the 13cm you reference, that depends a lot on various circumstances. Basically, they work with a range of depth of focus, which is not the same as depth of field, but that's all they've got to work with. Anyway, it's probably as good as for this example.

The camera knows how far it has got to move to get sharp focus within the acceptable zone, which is usually about one third of the depth of focus (depends on the model - D90 should be well up there). As soon as the lens moves to within that acceptable zone, it stops. It doesn't try to get any closer, as that would require hunting back and forth, which takes time. What I'm saying is that 13cm, I'm guessing, should be within that acceptable zone but perhaps not spot on.

Why not try the cerial packets suggestion. Make sure the camera is set up dead square. For added interest, get a 1ft ruler and prop it up next to the middle box so that the top stands behind the plane of sharp focus, and the bottom a bit in front. You can then see exactly where sharp focus hits in relation to the target.