I beg to differ, and why is... difficult to explain. It makes sense in my head, though.
Picture you're at an Inclination of zero, moving at exactly 7,500 meters a second. Now, the autopilots will tell you to burn facing Normal or Anti-normal. Lets use Normal as our example - so now you burn. Let's say you adjust your velocity in that direction by a large number of degrees... You do this, by adding movement in that direction. Now lets say that degree of change correlates to 250 meters a second. Seems small in comparison...
But now your velocity along your PATH OF TRAVEL, is now 7504 meters a second... (A squared, B squared, C squared)
It's a small change, sure. But lets say you're making a large plane change... That number adds up. Because your speed is faster, you'll find suddenly your ApA increases.
To aggravate this fact, most people do such large burns on time acceleration. This means that the autopilot is less effective, often lagging behind a degree or two from Normal or Anti-normal.
The end result, either way, is that not ALL of your expended Delta-V is being used to change planes. Some of it, the waste, is effecting your ApA (and possibly your PeA, but I can't recall ever experiencing that).
If the thrust is applied in a perfectly normal/antinormal direction, it will NOT change the amount of your velocity - just the direction of your velocity. However, many vessels are not perfectly aligned, including the stock DG.
Vessels have three axis, and the Center of Gravity is assumed (unless the vessel module deliberately changes it) to be located at the Origin (ie, x, y, and z all equal to zero). The Z axis is the "fore/aft" axis, and the Y axis is the "up/down" (from the vessel's point of view) axis.
Thrust is applied at a certain point (this may be an "average" of several individual thrust points). It this point is not at x=0 and y=0, and the thrust vector is aligned with the Z axis, then the vessel would be "unbalanced" and applying thrust would cause the vessel to rotate since the thrust vector wouldn't pass through the COG. When the thrust point isn't at x=0 and y=0, the thrust vector is ussually NOT aligned with the z axis, but with an axis that passes through both the thrust point and the COG. However, the stock autopilots use the z axis as the reference. This means that the even though you have the Normal AP engaged, the actual thrust vector isn't perfectly aligned with the Normal direction.
The stock DG, for instance, has it's thrust point a bit above zero on the y axis, and the thrust vector points slightly "below" the Z axis. Since the AP orients the vessel "belly forward", this means a slight amount of the thrust is in the prograde direction. This causes the vessel to gain a bit of velocity, raising the orbital altitude at the opposite side of the orbit. This can be somewhat mitigated by using the Horizon Level AP, and manually pointing the nose in the normal/antinormal direction. A small amount of thrust will be in the "down" direction, but the change to your Ap and Pe will be less than the change caused by using the Normal/Normal- AP.
The stock Shuttle-A is aligned better, and doesn't suffer this problem. IIRC, the DG-IV and XR series vessels have the thrust vector aligned with the Z axis as well, and you won't see the same change in velocity when using the Normal/Normal- AP's as you do with the stock DG.
to the forums
