Ah, after all these years I find that the problem is with this actual product. Normally I recommend Sandisk and I still do, but not the 64GB SDXC anymore. I’ve mostly used these in phones where they end up failing but I also design embedded control systems and integrate various SD and microSD cards into the product as well as write the file system software to acces them including the low-level drivers and SD commands. I was implementing format utilies to force 64GB cards to format to FAT32 but the one and only 64GB card I had that seemed to work fine in that I could read all the files off it that I had in my phone, just would not format.
Investigating this right down at the low level (in SPI vs SD mode) I see that the card responds to the write command CMD24 and follows through as if everything wrote to the 512 byte sector (SD cards are actually accessed as 512 byte “sectors” or blocks), and the response from the card says everything is fine, but a check reveals that nothing was written. Of course I also tested these with the SD Card Associations’s formatter including one from Paanasonic and other utlities as well, but no luck, either now, or with other cards in the past.
Just in case there has been a silent revision from Sandisk this is part of the internal information that I read in my system - for reference this is a SU64G revision 128 or $80 and rather an old one from 2013 it seems. My guess is that from my electronics engineering background that an internal “charge-pump” that boosts the voltage needed to erase Flash has failed and would result in these symptoms.
CARD: SANDISK SD SU64G REV$80 #2920609953 DATE:2013/5
RANGE… 2.7V to 3.6V
CARD TYPE… SDHC
LATENCY… 1ms+1400 clocks
PARTITION… 0 00 INACTIVE
FILE SYSTEM… exFAT?
CHS START… 2,10,9
CHS END… 1023,254,63
FIRST SECTOR… $0000_8000
TOTAL SECTORS… 124,702,720 = 63,847MB