STEVAL-ST25R3916B Y200 [crystal] loading capacitors not populated?

[wrig0421]

Hello,

I recently added ST25R3916B to a board.  I followed the evaluation board schematic pretty closely.  

During board "bring up" phase I had an issue where the osc_ok bit in `Auxiliary display register` was stuck low.  Specifically, after powering up the oscillator en bit in Operation Control Register..  The bit `osc_ok` bit always remained low. 

I spent some time trying to debug this from firmware perspective thinking that it may have been a firmware related issue [like incorrect sequencing]..  I ended up visually comparing the layout on my board to the STEVAL-ST25R3916B layout. 

In the STEVAL-ST25R3916B schematic the "HSE" xtal [Y200] shows 2x 10 pF loading capacitors [C217, C218] connected [see snip from STEVAL-ST25R3916B pg. 20].

The bill of materials also lists Y200 + loading capacitors to be installed.  

While visually inspecting the board... I noticed that [...what I believe to be...] the loading capacitors [C217, C218] were not populated.  Under microscope I could see trace connecting the capacitor to the xtal.  I ended up removing the loading capacitors from my design to more closely match the eval board.

After the loading capacitors were removed from the design, `osc_ok` bit was high!  I have not observed a case where `osc_ok` is low after enabling the oscillator.  Specifically, with the loading capacitors removed. 

The particular evaluation board has the following code on it. 

```

STEVAL$D25R16BA

15090-TT-4412

94V-0 2416

```

My colleague has another revision of this hardware it seems... Specifically the `2416` has `22` to start.  I'm wondering if this is a year code?  Technically my colleagues board has the loading capacitors populated.  

I'm a bit confused by this.  One evaluation board has the loading capacitors installed and one does not.  I'm wondering what's the recommendation and or if the capacitors are added per board based on parasitics?  

Any and all help regarding this difference is helpful!

 

Thanks in advance,

Spencer

[Ulysses HERNIOSUS]

Hi Spencer,

 

could you send a picture of the area around the Y200 which should have the missing caps, please. You can use a mobile + zoom if you don't have a good possibility with your microscope.

I am puzzled, not aware of such an issue. To my understanding the caps should always be mounted.

When are you checking the osc_ok bit? It is expected to need some time (probably up to 1ms) to become high after setting en=1. 

BR, Ulysses

[wrig0421]

Hi Ulysses,

I took a quick photo with my phone.  I will have access to a microscope this afternoon if you'd like a more zoomed in view.  In the photo below I put a light purple box around the 2x footprints that I believe to be the loading capacitors.  

Regarding wait time before reading `osc_ok` bit.. Understood.  In my debug prior to finding the HW difference above I added a second delay before reading the bit.  With the capacitors installed on my board I was still unsuccessful. 

We have a few more of these evaluation boards in lab.  I'm going to go see if the others have the capacitors installed.  

Thanks in advance for the help!

-Spencer

[Ulysses HERNIOSUS]

Hi Spencer,

thx. Yes, these are the crystal caps. It looks as if there was an issue during production of the boards - they might have only weakly mounted on one side and then fallen off.

Maybe the issue on your board is actually different? Can you share schematic and layout of the crystal and load cap, please!

BR, Ulysses

[wrig0421]

Hi Ulysses,

Below is a snip of the layout showing the crystal + capacitors used on the board. 

Regarding the comment on oscillator not guaranteed to be stable without capacitors.. I'm a bit concerned regarding this mainly because we have 3x evaluation boards that do not have them installed.  We have been using all three with no observed issue in our simple tests. 

Thanks for sending over the application note.

Thanks,
Spencer

[Ulysses HERNIOSUS]

Hi Spencer,

There are some boards with a BOM hiccup, where C217/C218 are not mounted but C203 is mounted. We are currently clarifying in detail which boards are affected.

All our boards should really have:

1. C217, C218 mounted (10pF)
2. C203 not mounted

The missing load capacitor may cause unstable oscillator (likely only seen over the temperature range). The C203 will cause bad waveshapes of the NFC modulation signal.

None of this seems to cause immediate fails at room temperature.

For the issues on your board please verify proper load capacitors being mounted - also calculated eq 7 from the AN - value might need to be different from the 10pF which you are currently using.

If you still have then problems let us know.

BR, Ulysses