I'm not sure I understand it myself really.
It has to do with the error recovery procedure within the drive. Although it doesn't happen often these days, a given read or write may fail the first time (or first few times) but work if retried. The difference between enterprise and desktop drives is how aggressive/persistent the firmware is. That is, how many times will it retry and/or how long it will retry.
"Enterprise"/"RAID compatible" drives limit how long they will retry internally. This is so that duration is less than the time the RAID controller uses to decide the drive is bad.
"Desktop" drives either don't limit at all or have a limit that is too long. When used in a RAID configuration the drive might be considered defective when it is doing internal retries.
A failed sector on a drive does not necessarily mean the drive is a goner. Sector failures in a RAID configuration can easily be overcome as if nothing has happened. But to do that the RAID controller must know that it's a bad sector and not a bad drive (as evidenced by the drive not responding at all because it's hammering away on retries).
Bottom line is don't use desktop drives in RAID.
It isn't just a different setting in firmware that makes a desktop drive a lot cheaper. The warranty is a lot less. The quality is less at some manufacturing locations. For example, I understand that the latest Seagate desktop drives are manufactured in China and have a much higher failure rate as a result.