Part Number:CC2640R2F
Hello!
We failed ETSI and FCC tests because of spurious emissions from our custom CC2640R2F hardware.
Emissions are at 4958MHz while in receive mode on BLE channel 39 and at 7440,5 MHz.
Hardware ist quite small 12mm by 40mm and I used a layout similar to that of CC2650 uTag ( tidrj73 ).
Antenna is a Johanson 2450AT07A0100T.
I read some post in this forum regarding those emissions generated by the internal VCO running at a frequency of 2x f_receive.
I think that HF leakage current from the VCO goes out of the CC2640 even at other than the HF pins.
If I look with a sniffer probe at the PCB the emissions are even slightly higher at PCB areas in opposite direction to the HF part (away in the other direction where no HF currents should be).
This is what I did up to now:
- I re-read TI's layout guidelines, referenced the CC26XX Design Guide and reviewed the HW design checklist
- I did a new layout because one trace was underneath near the HF signal
- I ordered the PCB as 4 layers instead of the original 2-layer design.
The 2 inner layers a groundplanes now and the effective height of the HF traces above ground are reduced about 0,5mm.
(I noticed most TI designs have PCB stackup thickness of about 0,5 to 0,8mm)
This gives me 2-3dB improvement but I think the HF leakage currents are still there and I need some more margin. - I tried different SMD parts in HF output for the PI filter
- I tried to absorb HF energy using thin ferrit foil sticking onto the PCB
- Even with the complete HF tree not populated (no filter, no antenna) the unwanted emission is still there.
Any ideas where to look for improvements?
Thank you!