It is not very clear from your posting if you have done the crcularity test both with G2 and G3. I assume that the first test is with G3 and the second one is with G2.From the crcularity error from what I see in your pictures, it seems that you must reduce the X axis gain a litle bit. I know that ussualy the gain must be the same but in reality it depends on the mechanics of the machine. If you didn't done the test both with G2 and G3 I sugest you to do them and if you see that the elipse is inclined in oposite directions than you must reduce the gain on one of the axis. If the elipse is inclined in the same direction than you have a perpendiculatity error between the axis.

As about the noise, try to increase the gain until the axis becomes instable and than reduce the gain at about 70% of the previous value.

Good luck

Nin

Instead of increasing the X axis gain, which lead to vibrations, try to decrease the Y axis gain until you obtain a good circularity value. You must first establish the less stable of the two axis by increasing the gain until the machine becomes instable. Write down the values and then reduce the gain at about 70% of the value. After that you must compare the gain and put the minimum value for both axis and perfom circularity test with G2 and G3.

After that you must perform also the circularity tests in all other planes in which the gain is diferent or you can proceed with the sugestion from Pierre.

Good luck

Nit