STM32F103C8T6, 2 x 64kB flash ?¶
Technically, only the STM32F103CBT6 contains 128kB of Flash, the STM32F103C8T6 has only 64kB according to the STM device specs.
The very popular Bluepill Board contains a MCU marked “STM32F103C8T6”, with many units having 128kB of Flash. This has caused a belief that STM put 128kB of Flash in every STM32F103C8T6.
But is this true ?
YES, it sure looks like it is, read on …
STM32F103C8T6 #1 with 128kB Flash¶
This Bluepill has a MCU marked “STM32F103C8” and clearly runs a Mecrisp-Stellaris 90kB Forth binary proving it has 128kB of Flash. Data is from EEVblog. ChipID not known.
ramfree . 19112 ok.
flashfree . -24692 ok.
free (bytes)
FLASH.. TOTAL REPORTED: 65536 USED: 90228 FREE: -24692
RAM.... TOTAL PRESET: 20000 USED: 888 FREE: 19112
Compare a Olimex P103 Board containing a STM32F103CBT6 MCU¶
Running a slightly bigger binary of 97kB. The MCU shows 128kB as expected and flashed a 90kB binary using standard OpenOCD configs, as expected.
stm32id
Die xy coords: 107741012
Wafer Number: 72
Lot_num ascii encoded [23:0]: 0x00555285 | U R .
Lot_num ascii encoded [55:24]: 0x87237267 | . # r g
ok.
free (bytes)
FLASH.. TOTAL REPORTED: 131072 USED: 97028 FREE: 34044
RAM.... TOTAL PRESET: 20000 USED: 892 FREE: 19108
STM32F103C8T6 #2 with 64kB Indicated Flash¶
This board has a MCU marked “STM32F103C8” and clearly fails to flash the 90kB binary, (using the standard OpenOCD configs) indicating it has only 64kB Flash … but does it ?
I own 10 of these Chinese lc-mini boards. They are not Bluepills, but are much the same.
Standard OpenOCD Config¶
The standard OpenOCD flashing utility reads the chip reported Flash size of 64kB and accepts it (as it should).
90kB Binary Failed¶
% ./flash-eb13e308316527be6a8963098c7291d5.sh
Open On-Chip Debugger 0.10.0
Licensed under GNU GPL v2
** Programming Started **
auto erase enabled
Info : device id = 0x20036410
Info : flash size = 64kbytes
Warn : no flash bank found for address 8010000
wrote 65536 bytes from file ./eb13e308316527be6a8963098c7291d5.bin in 2.032206s (31.493 KiB/s)
** Programming Finished **
** Verify Started **
Error: checksum mismatch - attempting binary compare
diff 0 address 0x08010000. Was 0xff instead of 0x00
diff 1 address 0x08010001. Was 0xff instead of 0x06
...
diff 127 address 0x0801007f. Was 0xff instead of 0x06
More than 128 errors, the rest are not printed.
** Verify Failed **
shutdown command invoked
62.6 kB Binary PASSED¶
./flash-ae9236b921613c07b1967f21ff20ee97.sh
Open On-Chip Debugger 0.10.0
** Programming Started **
auto erase enabled
Info : device id = 0x20036410
Info : flash size = 64kbytes
wrote 63488 bytes from file ./ae9236b921613c07b1967f21ff20ee97.bin in 3.485837s (17.786 KiB/s)
** Programming Finished **
** Verify Started **
verified 62672 bytes in 0.919840s (66.537 KiB/s)
** Verified OK **
** Resetting Target **
Info : JTAG tap: stm32f1x.cpu tap/device found: 0x3ba00477 (mfg: 0x23b (ARM Ltd.), part: 0xba00, ver: 0x3)
Info : JTAG tap: stm32f1x.bs tap/device found: 0x16410041 (mfg: 0x020 (STMicroelectronics), part: 0x6410, ver: 0x1)
shutdown command invoked
ChipID¶
stm32id
Die xy coords: 108003154
Wafer Number: 80
Lot_num ascii encoded [23:0]: 0x00545267 | T R g
Lot_num ascii encoded [55:24]: 0x87195634 | . . V 4
free (bytes)
FLASH.. TOTAL REPORTED: 65536 USED: 62672 FREE: 2864
RAM.... TOTAL PRESET: 20000 USED: 880 FREE: 19120
Getting 128kB Flash!¶
The standard OpenOCD flashing utility reads the reported chip Flash size and uses it, BUT modifying the configuration to direct OpeoOCD to IGNORE the reported chip Flash size and use 128kB instead results in the 90kB binary image loading!
stm32f1x.cfg Modification¶
# flash size will be probed
set _FLASHNAME $_CHIPNAME.flash
# flash bank $_FLASHNAME stm32f1x 0x08000000 0 0 0 $_TARGETNAME
flash bank $_FLASHNAME stm32f1x 0x08000000 0x20000 0 0 $_TARGETNAME
Results¶
** Programming Started **
auto erase enabled
Info : device id = 0x20036410
Info : ignoring flash probed value, using configured bank size
Info : flash size = 128kbytes
wrote 91136 bytes from file ./eb13e308316527be6a8963098c7291d5.bin in 4.948380s (17.986 KiB/s)
** Programming Finished **
** Verify Started **
verified 90240 bytes in 1.311107s (67.214 KiB/s)
** Verified OK **
** Resetting Target **
Info : JTAG tap: stm32f1x.cpu tap/device found: 0x3ba00477 (mfg: 0x23b (ARM Ltd.), part: 0xba00, ver: 0x3)
Info : JTAG tap: stm32f1x.bs tap/device found: 0x16410041 (mfg: 0x020 (STMicroelectronics), part: 0x6410, ver: 0x1)
shutdown command invoked
ChipID¶
Note this is STM32F103C8T6 #2 the same ID as the one above that initially failed the 90kB binary flash!
My Forth “free” program is obviously going to have to be changed to accommodate these findings :)
stm32id
Die xy coords: 108003154
Wafer Number: 80
Lot_num ascii encoded [23:0]: 0x00545267 | T R g
Lot_num ascii encoded [55:24]: 0x87195634 | . . V 4
ok.
free (bytes)
FLASH.. TOTAL REPORTED: 65536 USED: 90228 FREE: -24692
RAM.... TOTAL PRESET: 20000 USED: 888 FREE: 19112
All 10 of these STM32F103C8T6 powered :ref:`Chinese lc-mini boards<lc-mini>` successfully flashed the 90kB image, they *all* have 129kB of Flash.
Thanks “Bingo”!¶
My thanks to “Bingo600” from EEVblog for convincing me to try this out :)
