Firmware glitch with autofocus in A99!!!!

[aardvark7]

I was just running through the process of micro-adjusting my lenses and have discovered a really, really annoying issue.

It affects both my A99s and more than one lens, so I know it is a firmware problem.

Basically, when using a focal length of 135mm or longer (it affects my 135 f1.8, 70-200 f2.8 at longer end, and 300 f2.8), if you try to focus on an object that is about 10 - 12 feet away and the focus is starting from the close end (as opposed to infinity), then it will give a false focus lock to about 3 feet. The 300mm will just sit there as that is below its minimum distance.

It is repeatable and seems unaffected by other settings, such as microadjust on/off, AF distance on/off, centre point or zone focus, etc.

It needs to be addressed by Sony and QUICK!!!

It is just the sort of distance for intimate portraits and the 135 is just the lens. I can't be fiddling around waiting for the correct lock!

[Mick2011]

Hi aardvark7

Have you tried setting the focus to AF-S? What happens when you grab the manual focus ring while focussing in this mode?

Cheers

Mick

[aardvark7]

Hi Mick,

The cameras are in AF-S mode, which is the most appropriate for micro focus adjustment,

I'm not sure what your asking as regards the focus ring, though. This would be a bad idea for older geared lenses and only safe for SSM equipped. The DMF only kicks in after focus lock and is an activated option, as I understand, so grabbing the focus ring could put severe strain on the internal gears and focus motor.

Please do let me know if I am mistaken, but that was how I thought it all worked.

Basically, what I was doing was standard practice for the calibration i.e. taking a sequence of AF shots from a prescribed distance, but starting the lens from either end of the range to average out AF motor errors. If you take them all and start at infinity then there may be a tolerance error that will mean your microadjust would take that into account, so when you have focussed near to you and then want a shot which is farther away, you might get a different result.

The focus lock that was happening in my tests was only 3 feet from the camera, even though the target is almost 12 feet away, and it happens time after time under those specific conditions, so it isn't a fluke or a mistake.

Message was edited by: aardvark7

[Mick2011]

Sorry aardvark, I assumed you were using an SSM lens for some reason. I don't have that lens/body combination available to check out the problem, so it's very difficult to understand what's happening and how

I'll mention this thread to some colleagues who may be able to shed some light.

Cheers

Mick

[Mick2011]

Hi

Sorry it's taken me so long to reply. I referred your post to Sony tech support who referred it to someone in a lab... who performed some tests with the a99 and fast telephoto lenses. What with holidays and so on it's been a while, but I'm hopefult there's some light at the end of it.

Essentially faster lenses do 'challenge' the AF sensor to a greater extent, as the AF system uses a smaller aperture of arounf f5.6 to measure distance. However it shouldn't produce the results you've been experiencing; did you use a lens calibration chart?

The following is a suggested test setup to see if your lens in in fact funtioning normally, designed get a reproducible result and to avoid unnecessary shipment and repair:

• Mount the camera on a tripod and align the camera parallel to a high contrast subject. A checked pattern (or calibration chart) can be printed out or displayed on a computer screen. Distance should be approximately 2m (or a distance 30x focal length of the lens)

• set the camera to A mode and select a wide open aperture (larger than f5.6)

• Use manual focus and set optimum focus manually with the help of MF Focus Check and take the picture

• Switch to AF and put your hand in front of the lens to force the camera to refocus. Then focus on the target using spot focus and take the picture after the camera confirms focus

• To make sure that no camera shake influences the result it is recommended to use remote or the self timer.

After that compare the sharpness of the MF and AF image; any difference would suggest a fault with either the lens or the AF system.

Cheers

Mick

[aardvark7]

Thanks for your reply.

I know you're trying to help, but you seem to have completely misunderstood my initial problem.

If I set up the camera as you have suggested, which is entirely what you do for setting microfocus, then under most situations it will work fine. The AF and MF will agree, so no problem.

WRONG!

It happens only in very specific circumstances:

Distance to target: 10-12 feet

Focal length greater than 135mm (I don't know if there is a lower cut off, but I know this is a certain limit)

Start the lens from MINIMUM focus point, NOT infinity

If you follow this then the AF will give focus lock suggesting the distance to be about 3 feet to the target, which is clearly wrong.

I'm not sure how else to explain it as it is not a difficult combination!

I will add an update when I have made a video to demonstrate but, being a professional photographer, haven't had much time recently.

[Mick2011]

I'm assuming the suggested test was to establish if any hardware-related issues were the cause of the problem. As this has not AFAIK been widely-reported there is always a chance it is totally isolated and it's necessary to start with obvious checks first.

I've since referred your most recent your post back to the support team and am awaiting a response. Hopefully I'll get some news back and will report it here when I do.

Cheers

Mick

[mikeyp2000]

Aardvark - I have an a99 and just tried to reproduce what you describe but failed.  Here are the steps I followed using my A99 with Sigma 70-200/2.8 EX DG HSM II (my only full frame AF lens longer than 90mm...)

Set camera to AF-S, centre spot AF.  Set the lens to 135mm. Manually wind the focus to minimum distance (1.5m or something on this lens).  Aim at something about 12 feet away. Half press/hold the shutter button.

It locks on instantly, every time.  In my case I was aiming at a light switch on the wall.

What are you aiming at? Is it a test chart with some sort of repeating pattern? Does it happen with 'normal' real world targets.  Phase Detect AF can achieve false locks on simple repeating patterns (eg parallel lines or even window blinds) just by the workings of phase detect AF.

So, perhaps try a different test chart?  Or perhaps my test steps were wrong?

[aardvark7]

The target was the LensAlign device. I had this happen with both bodies and different lenses, but I will try to get a video of it to upload as soon as I get the chance.

[hokenne66]

I had the same issue yesterday on the A99 using a Sigma 120-400 APO HSM OS