LM3S9B92-IQC80-C5 wont boot up any more

Are you able to perform the steps for "Recovering a Locked Microcontroller" as described in the datasheet:

I will give it a try.

Is there a description or application note of how to do it with uniflash (like the one for TIVA)?

If using UniFlash you will not see LM3 as a selection option for device. Select one of the TM4C123 devices. Then on the "Settings & Utilities" tab, find DebugPortUnlock. Select Tempest and Firestorm Classes". Remember you  must hold the device in reset while powering up and until you can press the Unlock button.

Greetings,

Vendor's Bob has provided detailed "Recovery Guidance."    Hopefully that will succeed.

That said - your issue's "Cause Agent" remains unnoted.    You state that you're seeing, "An increasing number" (of such failures) yet provide "No real numbers" - which (may) indicate the issue's significance.

Should it not be asked:

• have these boards been probed for (any) other device and/or component failures?   (i.e. beyond the MCU, alone)
• have these boards been operated - at all times - w/in the MCU's specified operating environment?
• have these boards been - at all times - first handled - then operated w/"Full ESD Awareness & Safety Precautions?"    (ESD may reveal "after the fact" - as the 'damage' grows & becomes (then) evident.)
• Is the MCU connected to "Off Board" devices?    Might this enable "damaging signal levels and/or transients" - especially if proper "MCU defensive measures" are not in place?
• Is the board adequately powered - are "Rise & Fall times" w/in spec - may "External Input Signals" arrive while the MCU is unpowered?

A sound strategy is to "Quarantine" several "Tested & Know Good" such boards (golden boards) - built at/around the original "Time of manufacture" - and to operate (some) of these under conditions which "best match" those imposed by your client(s).    This discipline may serve as an early warning of (possible) board and/or component weaknesses - far better "Discovered by you" rather than your client(s)...

Tried on 2 different board with the failure. Same result after Unlock, not possible to reprogram again with LM Flash Programmer.

The method worked fine on a working board.

Looks like the LM3S9B92 is dead.

I am sure ESD is not an issue here, care has been taken during manufacturing.

Regards

Henning

Henning Christiansen said:

I am sure ESD is not an issue here, care has been taken during manufacturing.

ESD Safeguards must be enforced & continue:

• even prior to manufacturing (i.e. during incoming inspection/handling etc.)
• throughout client's unpacking, installation & "long-term" use

Even though your devices are long-term "EOL" - vendor here may be able to analyze/identify the (likely) failure mechanism - (really) clearing ESD.   (there also exist firms which provide such skilled services...)

You report, "Inability to Re-Program" - yet as you have a "working board" - have you measured (and compared) the failed boards' current draw?   (over several boards...)

Should these failures have:

• "clustered" around a certain date
• and resulted from a single client (or clients who use & install the boards (very) similarly)

such (often) provides a solid clue - much to your advantage...   (some "field change" may have occurred - which proves harmful to your board/devices.)

My firm has over 10K LM3S devices (still) operating successfully - at multiple client installations.   (some in use > 10 years!    Unfortunately none are LM3S9B92...)   

We have made major efforts to prevent "Outside World" signals from damaging the MCU - you are silent in that regard.

Our attempt here is to provide "helpful guidance" - if failures (are) clustered - I'd bet upon some reasonably recent, "impacting field and/or operational change event!" 

Flash memory corruption mentioned in Errata LM3SMEM#13 and LM3SMEM#14, can the MCU recover from this with a firmware update or does it permanent damage the device?

Regards

Henning