Can't program UTest in MPC5748G

[alexeikovelman]

I'm trying to program the UTest area of the MPC5748G MCU.
In particular I'm trying to enable password protection for the JTAG interface, so I'm trying to set the JTAG password field and the life cycle fields in the UTest flash area.

The trouble seems to be that only the JTAG password is written but the life cycle areas aren't - and since UTest is OTP I can only try once per board.

My Lauterbach script and the UTest memory dump are attached.

Two questions:
- any idea why the script doesn't work as expected?
- is there another way (that I missed) to enable password protection for the JTAG interface?

Thanks.

Original Attachment has been moved to: MPC5744P_DCF_Censorship_ours.cmm.zip

Original Attachment has been moved to: MPC5744P_DCF_Censorship_ours_fixed.cmm.zip

[petervlna]

One more thing:

in customer deliver censorship is don't care. But censorship register is programmed in factory to value other the 0x55AA.

So if you migrate to OEM production you have to program censorship register otherwise you will get your device bricked.

Or you must program valid JTAG password.

Peter

[alexeikovelman]

UPDATE: we eventually found a workaround: instead of advancing the life cycle setting to In Field we advanced it to OEM Production (which is a single step forward; that might have been the issue).

Like Peter mentions, we also needed to add a DCF with censorship mode and set the password groups and password group locks (also as DCFs) appropriately.

[petervlna]

Hi,

Its hard to say what is wrong as you didn't attach dump of your UTEST memory.

Anyway, here is reference script for censorship programing and also separate one for only life cycle programming.

As UTEST is OTP, make sure first that you really understand censorship of this device.

Or use the board with socket where you can easily exchange micro.

Peter

[alexeikovelman]

Hi Peter,

I've edited the original post to include the UTest memory dump; does this lead you to a different conclusion?

Meanwhile I'll take a look at your scripts.

Also, we found a bug in the addresses we used in the script, but fixing them didn't change the outcome; the script in the original post includes the fix.