[Need help] Strange flash problem

[daweiy]

Posted on May 24, 2007 at 21:45

[Need help] Strange flash problem

[daweiy]

Posted on May 24, 2007 at 06:06

Hi all,

I'm developing a usb powered device. One testing unit can't boot after dozens of plug in and out usb cable. Through JTAG, we found the boot address, only 32-bit at 0x40000000, had been cleared. After we rewrite the firmware by JTAG, this problem happened once after dozens of usb cable plug in and out. Now the USB D+ and D- pins go into processor directly without any protection circuits on them. We found another problem before was the processor died which could be caused by over voltage or static electric. All the problems occurred just in one PC which installed windows XP. The firmware didn't do any flash operation when usb cable pluged in. My question is: Is it possible the over voltage or static electric just destroy parts of the processor which causes the 32-bit falshrom data been cleared? Thanks for your time. Dawei

[kleshov]

Posted on May 24, 2007 at 07:17

Is there a suitable MCU power supply supervisor in your circuit? The RESET input of the MCU must be asserted whenever supply voltage is out of range. Otherwise you can expect all kinds of strange behavior, including unwanted flash writes.

You should also use anti-static protection on USB data lines. The MCU will not last long without it. - mike

[daweiy]

Posted on May 24, 2007 at 21:45

Hi Mike,

Thanks for your reply. I agree your opnion, we also worry about the power supply. Just not sure why it is always the boot address, 0x40000000, been cleared. If the address is a random value, it is understandable. Any way, we will add the protection circuits and keep on watching what's gonna happen. Cheers! [ This message was edited by: dawei on 25-05-2007 01:19 ]