Help, FCC/CE certification issue of CMOS RF transceiver

Dear All,

I have a 2.4GHz CMOS RF transceiver for FCC/CE certification, the FCC result is OK, but the CE result is marginally passed, the CE certification result showed that the 4.8GHz (it is on chip VCO frequency=2*fLO) leakage power is about -50dBm, while the CE requirement is <-47dBm, so we do not have enough margin...
the leakage power@Antenna is <-60dBm, which means the reverse isolation of RX path is OK, then, where does the 4.8GHz leakge power come from(is it from the radiation of on chip VCO?)? how can we improve it? I am confused, help me, thanks very much!

BRs
Buckaroo

Yes most likely from top of chip.
Does it have a lid?
How is ground plane designed?
Are IO wires filtered with beads?

SunnySkyguy said:

Yes most likely from top of chip.
Does it have a lid?
How is ground plane designed?
Are IO wires filtered with beads?

Click to expand...

Dear SunnySkyguy,

Thanks for your reply.
Yes, it is a SoC chip with RF baseband and MCU.
Because it is a kind of consumer electronic, the cost is important, and hence we don't have lid or beads.
The whole chip ground plane is connected together (ground mesh) except PA ground(QFN package, all ground pins downbonded), however the PCB ground is NOT connected together by a whole metal layer due to the limitation of 2 layer PCB.
I am confused that the 4.8GHz (=fVCO=2*fLO) leakage power@antenna is about -70dBm, but the CE certification result is as high as -49dBm, I guess it is most likely from radiation of on chip VCO inductor tank. I am a RFIC designer without FCC/CE certification experience before, it is tricky
Thanks

BRs
Buckaroo

We used to make our own shields from sheet brass, tin plated , etched at a PCB shop with fold marks and solder tabs. But you can also buy from Tech-etch (I think) or their distributor in china. (mindar). Some had lids.

SunnySkyguy said:

We used to make our own shields from sheet brass, tin plated , etched at a PCB shop with fold marks and solder tabs. But you can also buy from Tech-etch (I think) or their distributor in china. (mindar). Some had lids.

Click to expand...

Dear SunnySkyguy,

Thanks,
Finally, it was found that the harmonic is due to the VCO test buffer leakage.

BRs
Buckaroo