| ??? 11/21/09 00:06 Read: times |
#170996 - Please read this document... Responding to: ???'s previous message |
Richard said:
I am, however, disappointed that, while many seem to agree that "RESET problems" exist, nobody, not an 8052.COM discussion participant, and not an 805x manufacturer, even those who clearly proclaim the existence of RESET-related problems with their 805x products, explains what these "RESET problems" are, and how they manifest themselves. Please read this document (yes, I know you hate ATMEL...): http://www.atmel.com/dyn/resou...oc4183.pdf At page 2 you will find the following passage: Theory of Operation:
For the MCU to successfully decode and execute instructions, the supplied voltage must always stay above the minimum voltage level specified by the product datasheet. When supplied voltage drops below this level, the MCU might to execute some instructions incorrectly. The result is unexpected activity on the internal data and control lines. This activity may cause MCU Registers, I/O Registers and Data Memories to get corrupted. To avoid unexpected activity, the MCU should be prevented from executing code during periods of insufficient supply voltage. This is best ensured by using a Power Supply Low Voltage Detector. Below a fixed threshold voltage VT, the detector circuit forces the RST pin high (active). Forcing RST high immediately stops the MCU from executing code. While the supplied voltage is below the required threshold voltage VT, the MCU is halted, making sure the system stays in a known state. When the supplied voltage rises above this predefined voltage, the RST pin is again released, and the MCU starts to execute code beginning at the Reset Vector. Richard said:
As I've pointed out, I've observed run-on, including access to external memory, during slowly decaying Vcc in the presence of ACTIVE RESET. The circumstances were a bit arcane, but still, it shouldn't happen, should it? Yes, this is indeed strange and should not happen! Could you repeat this failure? Maybe the manufacturer (Dallas?) has fixed it now? Richard, I know that things are way more complicated than manufacturers want make believe us. I remember a thread, where Lynn, I think, who better knows than anybody else here, how this 8051 internally works, because he fabricates them, told us, that the reset isn't asynchronously working with the original 8051 topology, stopping all activities by just putting the reset line high. But, this original 8051 didn't contain a flash code memory either. It's my strong believe, that the manufacturers of todays flash mircos have changed the design in the mean time, to allow a rigorous stopping of all internal and external activities when activating the reset line. As burning the flash needs some period of time, even a somewhat delayed stopping of micro's activities wouldn't result in a change of code memory. Kai Klaas |
