[Custom PCB] About capacitors which are decoupled at 16MHz

[MANGO]

Hi, I made a custom PCB.

I use a Revision 2 MCU.

The antenna is printed. I didn't used the chip antenna.

Sorry for not showing both the full schematic and the Gerber file.

I have downloaded the DTM firmware from SDK 7.2.

The next picture is the frequency when it received command to do TX test of channel 0 (Advertising channel 37).

While testing for CE (Conformite Europeenne) and KC (Korea Certification),

the result was quite poor when I used 12pF for decoupling the 16MHz crystal.

So I changed the caps. When I use the 27pF cap, it finally matched the frequencies.

The cap is the Murata GRM1555C1H270J.

But the biggest problem is that the TX power value is -16.90dBm.

/*******************/

So my questions are,

1. How could the PCA10001 use 12pF?

Comparing with the PCA10001's schematic,

I replaced the inductor L1 and L2 to the balun (datasheet link).

After I changed the cap to 27pF, my board and smart phones pair each other.

Does the cap size for decoupling 16MHz crystal varies case by case?

2. The TX value and the 16MHz crystal are independent, is it?

What are the reasons for poor level of TX?

-Regards, Mango922

[Øyvind Karlsen]

Which package type are you using? The balun you are using is intended for use with the WLCSP type packages, if you are using a QFN48 package you will be having trouble.

[MANGO]

Really? It is terrible that I'm using nFR51822-QFAA Rev2. What is the reason for not using that balun?

[Øyvind Karlsen]

That balun is simply not made for the QFN48 package, however there are several alternatives, for instance the 2450BM14E0003 from Johanson.

[Øyvind Karlsen]

Using that balun with a QFN48 chip will result horrible RF performance.

[MANGO]

Hmm... Can I ask for more detail?

Or is there any application note for this?