Ambiguous LQFP100 Pin 1 Markings?

[JOlso.12]

I have just wasted three solid days of very intense bench work trying to understand why my STLINK programmer will not communicate with my newly-created 32F419 target board. As a result I am now an unintended expert on SWD programming.

The reason my target board won't work is that the 32F413 is installed 180 degrees from the correct placement. I assumed that the dot inside the dimple is the pin #1 identification mark.

Wrong. Pin #1 is identified by the dot without an underlying dimple.

STMicro needs to correct the Pin #1 marking ambiguity in the mechanical drawing in the datasheet for the LQFP100 package.

[Tesla DeLorean]

Part orientation was specifically called out. https://community.st.com/s/question/0D53W000011wZJhSAM/what-to-do-when-swd-port-is-nonresponsive-to-stlink

Pretty sure VCAP voltages (1.25V) won't be visible when rotated.

https://community.st.com/s/question/0D50X00009Xka68SAB/stm32f401-64pin-pin1-marking-orientation

[JOlso.12]

You are correct, Tesla. Thanks for the feedback.

Jim Olson

Indianapolis, IN US

[waclawek.jan]

I understand your frustration.

This is ejector mark (ejector is a moving pin which ejects the part from the mold into which the plastic was injected), and it can be confused for the pin1 marker.

Placement, size, appearance, number, etc. of these marks is dependent on the particular site where the given part was packaged, ST uses several such sites so that may change.

The placement of pin 1 relative to other markings is indicated in the datasheet:

 

JW

[JOlso.12]

Thanks, guys. I was looking at the mechanical data on the prior datasheet page.

 

[Tesla DeLorean]

This has been a long standing cause of issues, without doubt.

One problem is that there are multiple sites and equipment involved in the encapsulation and laser marking of devices. Third-party in some cases.

One of the additional tells is the orientation of the part markings, part should be readable (English, Western sense) and pin#1 toward lower left corner.

It would perhaps help to have a gallery of pictures, as the data sheets tend not to capture the variations and nuances.

[JOlso.12]

I agree. Almost no manufacturer distributes long form data sheets in paper form, choosing instead downloadable PDFs where a high-res image of the package's front surface isn't a distribution burden.

STM manages to place their logo in color in all data sheets. That openness to larger PDF size could be used for more package detail.

[JOlso.12]

One thing that delayed my realization of this grave soldering error is that an improperly soldered micro doesn't crowbar its VDD rail. Quite the contrary, the VDD voltage and the current drawn from the VDD rail was not unusual.

However, as Tesla DeLorean pointed out, I should have confirmed proper voltages (~+1.7VDC in my case) on the VCAP1 and VCAP2 pins. These voltage were no doubt wrong had I looked at them.

For a while I puzzled over the absence of any waveforms on either of my two resonator-based oscillators, but these oscillators are inactive until enabled by code that I could not program into the part.

Pre-programming, the device is clocked by its own internal oscillator.

[TDK]

The marking in relation to the text is shown in the datasheet.

[waclawek.jan]

> an improperly soldered micro doesn't crowbar its VDD rail

DIL/SOP/TSSOP chips with GND/VCC in opposite corners (74xx-style) tend to blow fuses (and EPROMs in windowed CERDIPs glow momentarily, nice but sad view). However, modern mcus even in DIP/SOP/TSSOP, and even more so in quad packs, tend to have the supply pins so that they either match when turned, or fall to an IO (which unless externally driven hard, is mostly harmless) so they tend to survive rotation, even the layout is unlikely to be motivated by such a "feature".

Here, VSS/VDD pair on 49/50 matches the pair on 99/100, so the mcu was properly powered. The rest of the supply pins match IO pins, so again unless driven externally hard, it's likely to survive. 90deg rotation should be similarly survivible thanks to the 74/75 pair. But again, this is certainly not the primary drive behind this layout, just an interesting coincidence.

JW