168 external RAM - problems when not battery-backed

[steve8]

Posted on August 02, 2005 at 09:25

168 external RAM - problems when not battery-backed

[steve8]

Posted on July 01, 2005 at 10:03

We're having problems on the latest version of our ST10F168 boards.

We use external RAM, which used to be battery-backed in the last version of our project. We no longer need this facility, so have removed the battery from the latest design.

The 168 now struggles to boot and run code correctly. Even a simple test program (printf 0-9 out of the serial port) fails to run correctly - it will print the first 2-4 correctly, then hundreds of garbage characters, then appear to reset and finally print the remaining numbers.

I originally thought this was about variable initialisation, but this simple code doesn't use any variables.

It is fine if the RAM is battery-backed, and also fine after a reset.

We've tested the power-on/reset signals, and believe these to be operating correctly.

Does anyone have any ideas how to pursue this?

We are stumped by this - our only solution at the moment is to replace the battery on the new boards (hardly an elegant solution, and we would like to find out what is going on)

regards

Steve Peake

[michaelgaus9]

Posted on July 05, 2005 at 09:10

Hello,

which type of NVRAM do you use, is it from ST?

We had a problem with some M48Z35Y: when no battery was connected, RAM access failed (internally blocked in the NVRAM) because the internal voltage threshold circuit was floating.

When we connected a battery, RAM access worked fine.

When we shortcircuit the two pins for battery connection, RAM access also worked fine.

If you no longer need NVRAM, you should replace it with a ''normal'' volatile RAM chip.

[steve8]

Posted on July 05, 2005 at 12:10

Thanks for this advice - I'll pass it on to our hardware guy for testing. I'll post the results here for others to reference.

cheers

Steve Peake.

[steve8]

Posted on July 05, 2005 at 13:55

Just heard back from our hardware guy.

We are not using NVRAM.

We're using normal SRAM with a backup battery diode-ORed with

the main supply to retain the RAM data.

Does this generate any other suggestions?

regards

Steve Peake

[steve8]

Posted on August 02, 2005 at 09:25

We've solved this one now - it was a faulty RAM chip.

Replaced the chip and the memory now behanves as expected without the battery-backup

many thanks for the helpful suggestions

regards

Steve