STM32F100C8 seems to have 128k flash despite datasheet

[ingwmeier]

Posted on March 23, 2012 at 16:34

we made EEprom emulation tests with the discovery board first (with 128k of flash). Our own target uses the C8 processor. We have not paid attention to flash size in regard with eeprom, but did not run into problems, so stepped through Memory 0x0801000 up to 0x0801fff it shows our eeprom values in start, then ff ff with no warnings up to this limit. Only when reaching 0x08020000 it says Cannot access Memory in the command window..

Maybe ST produces only 128k types and marks some as C8 (???)

[emalund]

Posted on March 23, 2012 at 16:56

<i>Maybe ST produces only 128k types and marks some as C8</i>

can't say, but there has, in the past, been many examples of a manufacturer being out of stock of a 'cheap' variant relabeling a more 'expensive' variant to be able to supply the 'cheap' one. I would not rely on it.  I do not know how it is today, but when the dinosaurs roamed the earth, it was quite common to see variants with nore memory than stated where the 'extra' memory had a problem. as an example many 80c31's happened to be 80c51's with bad memory. Erik

[Tesla DeLorean]

Posted on March 23, 2012 at 17:08

ST manufactures less different die variations, than part designations.

In many cases there is more RAM and FLASH on a given device than the quoted amount, it may have been binned that way based on testing, lack of testing,  or production/sales demands. Assume only the documented capacity is viable.

[ingwmeier]

Posted on March 23, 2012 at 17:31

I sure will not rely on the extra amount. Erik, I used to be around the Dinosaurs: Silabs types 8051F040 and 046 as a cheaper option where I also found 64k instead of the stated 32k, so I had no SW option to decide which type it was..

[alok472]

Posted on March 24, 2012 at 03:42

It can be a disaster using this extra memory...the device will report field problems due to unreliability and finally the cost runs out to be more ..