I have 2 drives in my system: Pioneer DVR-105, Liteon LTR-48125W. Their offsets are configured via Accuraterip, so I can normally rip in both drives and get the same CRCs. But I have found an exception. I have a CD with 14 tracks, where track 14 consistently produces one CRC on one drive, and a different CRC on the other. This happens when copying or testing, using either burst or secure mode. In secure mode, the track quality is always 100%, so there is no error correction happening. The other 13 tracks produce the same CRC on either drive, so only track 14 is affected. None of the tracks are confirmed by Accuraterip (confidence is about 20), so it looks like I have a different pressing.

For track 14, I have ripped to WAV from both drives and the "Compare WAVs" feature in EAC. It shows that there are different samples from 0:03:51.638-0:03:51.678, which is right at the end of the track where the audio is effectively silent. I can't see or hear any difference using the "Process WAV" feature. So from a practical standpoint, the difference doesn't really matter. From a technical standpoint, I am still curious. This is the first time I've seen this in well over 100 rips. Has anyone else seen anything like this? Any potential explanations? Thanks!

One drive is probably overreading into lead out and getting some data that the other drive is missing. At least that would be my guess.

If you're using EAC it could be the "Fill up missing offset samples with silence"
I have this field checked (it's in EAC -> EAC options (F9) -> Extraction)

If it's the last track it's not necassarily due to overreading. One just has a different offset, and so one can read further/less into lead-out. In 99% of cases overreading isn't even needed, but on some discs you'll get a different CRC's on the first or last tracks depending on the drives. Nothing to be concerned with. Worst case, get a drive that overreads and you'll have nothing to worry about.

Thanks to everyone for the info, makes sense.