Package Data Register

[MCUExpert]

I am using 2 devices

STM32L4A6VGY6P (package WLCSP100L) 

STM32L4A6ZGT6P (package LQFP144)

My FW needs to differentiate between these 2 devices and so the package data register (0x1FFF7500) seems like a no brainer to use for that.

The problem is my WLCSP100 is reading ID of 3 which is supposed to be the UFBGA132 according to RM0351 Ref. Manual.

If this method is not reliable (inconsistent programming from factory), I won't be able to use it.

Thoughts? And does anyone know any other clear test that would give me the correct differentiation results?

[TDK]

> And does anyone know any other clear test that would give me the correct differentiation results?

Tie pins to ground on the boards, test them during startup and use that to determine which board it's on.

Alternatively, you can also use pins that don't exist on one chip as the test. Generally, if pins don't exist those pins will always have IDR=0 even if you assign them to output high. Test this, behavior is not technically guaranteed but that's what I found.

[MCUExpert]

The boards already exist, but yes that would be ideal since we'd be in full control.

Your 2nd idea is more what I am after. I'll check it out.

It is just a disappointment that the ref. material is not matching what I observe.

[Pavel A.]

You wrote that there are only two board variants. The value for LQFP is 0. If that one is reliable, then you have just one simple test: 0 or not 0.

[MCUExpert]

The newest board (with the LQFP) will be assembled for the first time within a month (I don't have access to it yet).

I am just questioning the faith I have in this mechanism as a whole since 1 part of it isn't adhering to the ref. manual. And what if it works now but a lot in the future doesn't?

I for the most part am just checking in to see if there is some other detection method that I am not aware of.
(And of course to poke ST a tad...if they even monitor this).