Rotated TQFP parts?

I’ve seen a couple posts here about the problems people have when they try rotating most smd parts in Eagle. In my case, I have an ATmega644, TQFP 44 that I’ve rotated 45 degrees.

As other people have posted, I’ve had problems with this board failing the initial DRC check, since my generated gerber has a 3.9 mil aperture.

I’ve read that it can be ok to just delete that aperture definition in the gerber file. However, I haven’t seen anywhere where someone who has done this has received their board and said they got a good result. So, has anyone made a board with rotated surface mount parts? Did it come out ok?

Well, in case anyone else is having this problem, I should share what I learned. Deleting the aperture definition is very bad - it makes the gerber render without pads for the part you rotated. It also chokes the DRCbot, and you just won’t get a jpg for a layer that’s like that (though, you’ll still pass DRC).

I’ve tried increasing the aperture from 3.9 mil to 8 mil. It seems fine, though the pads do seem a little large. It passes DRC, at least.

Anyone have a better solution here? Ideally, the cam processor wouldn’t even output gerbers like this.