quebec001 has used the abbreviation of 'gig' to describe the size of his 'big' and 'small' images in Post #1 as '3.8 gig' and '1.98 gig' respectively. He means 'GiB' i'm sure as those images don't exceed the FAT32 and FAT16 file size limits of 4 GiB and 2 GiB respectively [less 1 byte each].
sergio2 has used the abbreviation of 'Go' [GB] for his comment that at 4,1 Go "it slightly overtakes the FAT limit" but "is well displayed". That's partly correct, in that 4.1 gigabytes in decimal is ~4,100,000,000 bytes which expressed as binary is 3.8 gibibytes and the reason it's displayed on the unit in an 8'GB' card is that it doesn't exceed the FAT32 file size limit at all [that limit is 4,294,967,295 bytes]. However it's too big to fit on a 4'GB' card, which is already less than 4,000,000,000 bytes because a bit of space is used by the manufacturer for hard-coded information.
Historically the terms kilobyte, megabyte, gigabyte, terabyte etc were binary [value 1024] whereas now they're usually decimal [value 1000]. While the manufacturers of flash and hard drives have embraced the change [in fact they started it] MicroSoft is clinging to the past and insists that gigabyte means binary. So when you check 2 or 4 'GB' microSD in a reader attached to a PC the properties show this:
[Only registered and activated users can see links. ]
And that's why an image just under the FAT16/FAT32 limits of 2,147,483,647 bytes/4,294,967,295 bytes won't fit on a 2'GB' or 4'GB' card respectively.