I think my 24mm 1.4L is really back focusing. Check this out:

LiveView:



ViewFinder +0 MA



ViewFinder +10 MA



ViewFinder +15 MA



All of the pictures were taken using a tripod and are 100% crops.

Questions:

1. What is this scale that is used when adjusting MA? I'm setting it to +15 to match the LiveView shots.

2. If a lens back focuses on 1 body, does it mean it is also back focussing on my 40D?

3. Would a lens calibration alone fix this or do I have to send in my bodies also?

Thanks in advance.

 

I use a focus chart at 45° to see if my lenses are either back or front focusing. It's really hard to tell with the picture you are using, but it obviously seems better when you add 15 to the AF Microadjust.

My sigma lens needed +20 (max) to get it somewhat in the ballpark because it was front focusing really bad. Once I get back from vacation, I am going to send the lens in to get it calibrated.

If the lens is that much off on your 5D, I'd imagine it would be off by the same amount on your 40D. You would only need to send in the lens to get recalibrated.

 

I'm confused with the viewfinder and liveview comparison. Are you manually focusing through the viewfinder and finding it's blurry? If so, your viewfinder diopter wheel was probably turned accidentally.

Liveview lets you focus either with the standard AF sensors in the camera or contrast AF that uses the imaging sensor to determine what's sharp and what isn't. Which mode are you using?

 

The LiveView image is manually focused. The ViewFinder images were focused using autofocus. I didn't do any manual focusing with the ViewFinder images. I read that's how you do the comparison in order to show what's possible (manual LiveView) and what the camera is actually doing (autofocus). Used a tripod and remote trigger each time taking 3 shots for each variation. I just posted 1 instance of each to show the differences. I did just check and the diopter wheel is smack dab in the center.

Is the way I did this test incorrect? Should I be doing it the way SaaBaaDoo recommended?

 

It sounds like you're doing it fine. You can rig up a ruler to help you determine how front or backfocused your lens is, but you can keep doing what you're doing. The biggest trick is to make sure that surface you're focusing on is completely parallel with the front of your lens. You need it to be aligned as accurately as possible.

There are a couple things to keep in mind with the Micro Adjust function:
1) The numbers don't mean anything important. It's just a range of adjustment and a number for you to reference later. If a lens takes +15 it doesn't mean it's broken, it's just a number you dialed in to get the result you want.
2) A lens might front focus at far away subjects, then it might back focus on closer subjects. Micro Adjust will compensate for either front focus or back focus, but not both. That's fine if the lens is consistently back focusing at all distances, but it may actually make the problem worse at one end or the other if you have an inconsistent problem.
3) A zoom lens may have different front focus or back focus issues at different focal lengths, so they're next to impossible to adjust with this function.


Most lenses are designed with an intended purpose. For example, an 85mm is mostly used for portraits, so it would likely be sharpest at around 8 to 12 feet. Closer than that and it will likely back focus, farther than that and it will probably front focus. With that in mind, if you were calibrating an 85m and you wanted to it to be razor sharp at the range you most use it, then you should set your target at that distance and do all your calibrations there.

The 24mm is a little more multi-purpose than an 85mm, but the same advice applies. Pick a distance that you mostly shoot from and do your calibration from there. In other words, don't calibrate it at 5 feet when you use it at 30 feet, and vice-versa.

 

Thanks guys. I tried a 45 degree focus chart also and it was different results. What Dan mentioned on calibrating with the distance you use makes sense. The 45 degree chart test was only about a foot away from the camera so I can see how that would have a different result than if I calibrated from a more realistic 6 feet away.