I have never tested in this manner. I try to eliminate the variables so I can focus on one thing. I would check the lenses side by side in manual focus in live view and compare the results. That will tell you how your copy compares to your 24-105 for sharpness.

http://www.the-digital-picture.com/R...mp=4&APIComp=0

At the same aperture of f/4 the 24-70 should be better.

From the sharpness test -15 isn't the setting. I liked the previous results which peak at f/2.8. I think those settings were closer to what it is really trying to do.

70mm should be the easiest focal length to test. If you are having this much trouble it leads me to believe that it should go back.
70mm is a weak point for this lens, but there are some copies that perform very well. As long as Adorama keeps exchanging I would keep trying.

The AF aperture sharpness tests I do I set up at 50x with good lighting. Then I went from +5 one step at a time to -5 and then compared results.
Keep in mind that when you do the test the camera AF once at the start of the run. Since not all AF shots are the same, one may be a bit behind and the next in front the test can be time consuming. You end up with some results that just don't fit because the AF system missed. But if you run the scenario several times you will see a pattern develop.