This is a pretty common phenomenon... lower resolution camera files tend to look sharper straight from the camera, compared to higher resolution. For example, we've read and heard same when folks compare 5D classic (13MP) with 5DII (21MP) as well as many other models.
I bet you're zooming in to 100% or 200% on your computer monitor when comparing images from both cameras. Probably using the same high level of magnification on both.
Think about it... 10MP camera files might look fine at that size. But you are actually viewing the 15MP images from the 50D much larger, so much more critically.
So, either back off on the magnification of the 50D's files for a more fair comparison. Or make a downsized file from the 50D image, that's more comparable to what's coming out of the 40D.
Ultimately, it's always better to compare prints made with a good printer on smooth, matte paper.
50D sensor is much more densely crowded, too.... Might be using stronger anti-alias filter or heavier noise reduction, either of which would need more sharpening after the fact to bring detail and sharpness back out. (I can't say for sure... haven't personally compared 40D and 50D very closely... But I do see this effect going from 50D to 7D.... 7D files need much more sharpening than 50D's... but the detail and sharpness is there).
I agree... 50D has Micro Adjust feature... Use it! You will probably find it more difficult to MA zoom lenses. Test those at three or four different focal length settings... you might find they need slightly different amount of MA and will have to compromise. A prime like the 300mm is easier to MA, and more likely to need it since often with longer lenses you have relatively shallow depth of field that's less forgiving of focus error.
You might want to test your lenses at various apertures, too. Might be different, which apertures work optimally on 50D.