The SIM frying incident also happened to me recently. (I read the QSG, but missed the warning as an usual precautionary note. Shame on me… Fortunately, it was a “toy” prepaid card brought just to test the phone before I put my real card in it.) After that, I was curious why this happens with L5 and not with other phones so I took a look on the schematics (do this with a “standard” phone… ). For me, it indeed seems an immaturity in the design. What I found out:
- The combo card socket is such, that the SD part’s contacts sweep over the SIM card while the tray is moved
- The SD card power supply contact is always energized, even when the tray is pulled out (tray detect switch is disengaged), contrary to both ISO 7816 (which requires all contacts to be de-energized and disconnected immediately upon the beginning of removal of the card) and common sense
- The relative arrangement of the SIM and SD contacts result in 3.3V with reverse-polarity to be applied to the SIM power supply pad and then clock, then reset pads in while the tray is pulled out (and in reverse sequence when it is pushed in), damaging the SIM card and also possibly over-stressing U43 (power switch of SD card)
A possible fix to this problem that can be made by Purism in future releases of the mainboard is to make sure that disengagement of the tray detect switch disables the SD card power immediately in hardware, for example, by adding an additional pull-down transistor at TF_PWR_3V3_EN driven by TF_NCD or 4G_SIM_CD depending on the card detect switch’s operation. Aslo, what can be done meanwhile is updating the documentation to make it explicit that the “don’t remove the SIM tray while the phone is on” statement is not a customary warning that can be routinely ignored, but a workaround of an otherwise serious hardware bug.
I also wonder if this or similar deficiencies may have been spotted (and corrected) if the schematics would be subjected to public scrutiny before start of manufacturing.