 |
| Author |
Message |
|
|
 Posted: Oct 25, 2011 - 05:43 PM |
|
Joined: Oct 25, 2011
Posts: 1
|
|
Hello everybody,
I've got a problem with the crypto memory AT88SC0104CA and would like to ask if anyone in this forum has had similar experiences.
We have been using the AT88SC0104CA for years without problems, but recently, several chips have failed in the following way:
The chips are personalized, zone 0 is protected and the chip is mounted on a device which is operating normally. After some unknown event, Zone 0 is locked, AAC0 is 0x62, the config zone and the unprotected zones 1-3 are accessible.
The way I can recreate the problem indicates that it might be caused by a power glitch or communication errors caused by the glitch, but so far we could not find any irregularities in the hardware.
I can recreate the error on the AT88SC0104CA by disconnecting the chip's Vcc from the rest of the
board, connecting it to a separate power supply and manually changing the voltage in a wide range
between approx. 2V and 5V while reading the protected zone 0 in a loop. This works reliably on the CA (3.3V) variant (repeated 10 times) but not on the old C (5V) variant.
On one chip, I've seen AAC0 change from 0xff -> 0xee -> 0xcc -> 0x62.
Has anyone else seen this specific value (0x62) or any other undocumented value occur in AAC0 on a *personalized* chip? Does anyone know how this can happen under normal operating conditions?
Best regards,
Stephan |
|
|
| |
|
|
|
|
|
|
|
FAQ
Search
Register
Log in to check your private messages
Log in
Maximize
RSS
