| ??? 08/02/07 23:25 Read: times |
#142674 - I don't doubt any of that, but ... Responding to: ???'s previous message |
Erik Malund said:
So ... you thought you had a reset problem ... how did you know it was a reset problem? How many units did you test, and for how long, in order to verify what the problem was? How did you prove that it was "fixed?"
An inherited unit (sold in thousands) had many reset problems (customer statements "we need to power it up twice to make it run" techie statement: "we often have to reflash the boards"). What is it that persuades you that this is a RESET problem? In what way do you associate the flash corruption with RESET? How, given that RESET isn't a real reset, would a "proper" reset fix this? Making the 'b' version (also sold in thousands) properly incorporate a supervisor resulted resulted in "when a customer complains about reset problems we give him a 'b' board and never hear from him again". Also, to the best of my knowledge NO 'b' board has ever been reflashed. Please note that the policy here is FIRST to make the customer happy THEN to consider cost.
proof, as you want it, maybe not proof, as I want it, definitely Erik I'm not looking for proof, as I don't doubt that these problems exist. However, I'm curious how one can tell there's a RESET issue and not some other sort of problem. I rather suspect that the RESET is simply inadequate. What do you suppose will happen if Vcc is dropped at a rate of, say, 1 volt/minute? That leaves the MCU oscillator running for a LONG time after reset goes true. It also leaves the oscillator running for a long time after the logic has become unpredictable. RE |
