In my experience, I have never had a reasonable bit rejected on the first measurement. Only in extreme case of the bit too short (or bit plate stuck in) or when the bit is measured as different for the second time in a multi-bit project. Therefore, I think the bit rejection must be an electronic data problem. Since it is somehow related to the dust hood, I would check the z encoder. Is the encoder cover loose? Does it hit the machine cover or cut motor wires during the bit plate check? The dust hood and pipe could be flexing the machine cover. Is the machine cover loose?

After thinking about this for awhile, that's all I got. (I am back to thinking both problems are related.)