nRF52832 ANT grounded zero impedance burned?

Question

Hello, I am making my first custom board based on nRF52832. I have used a custom antena design (due to PCB layout I did not use reference design). That was the very first board prototype.. where I was testing some experimental firmware features.. I did not get the BLE connection.. I have verified the ANT resistance and it is zero! Did I burn my BLE radio? Is it possible or there are some protections and that may be caused by short circuit during assembly? Now I start to regret I did not use the reference design ;-) Anyway, how can I verify radio design in a safe way so I don't burn my BLE radio output?

Andreas · Accepted Answer

Hi, 
 
 The radio front-end has a transformer which is DC short to ground so this is as expected, your chip should not be damaged. 
 
 Are you still not able to get a BLE connection even if you put the PCB and the device you use for testing right next to each other? Have you verified that the HFXO is starting? 
 
 Best regards, 
 Andreas

Andreas · Answer

Hi, 
 
 The short impedance of the ANT pin is as it should be. Even full reflection at 4 dBm output power it should not take damage. You can dial this down a bit, -8dBm should be possible to receive if you put the phone/measuring device right next to the PCB and have plenty margin to the levels where it might hurt. You can also look at the current consumption when set up as a TX carrier. 
 
 If you do not see any radio activity it might be that the crystal is either not starting or incorrectly calibrated. Have you verified this? The load caps shall be approximately 2*CL-4pF, CL is specified for your crystal. 
 
 Best regards, Andreas

Andreas · Answer

Hi, 
 
 The 32kHz crystal is not mandatory, but if it is not populated you need to specify the internal RC oscillator to be used as LFCLK source when enabling the sofdevice. 
 Connecting Radio GND with board GND as you describe should not be a problem for the radio to operate. The harmonic emissions might be too large for regulatory tests, but you should still see communication. I have not seen your design though so it is hard to give an exact theory than that, but if you could upload the schematic and layout here I could give it a try. We can also give you pointers on what to improve for future boards spins, and if you do not want to share it with the world you are welcome to register a 'private' case (only visible to Nordic employees and your distributor). 
 We do not sell modules, but a handful of other companies do. You can see a list of them here . I have no idea about pricing, but expect it to be quite a bit more than the chip alone, so you might not give up quite yet. 
 
 Best regards, 
 Andreas

CeDeROM · Answer

Hello Wesp, one of chips were broken, probably bad soldering, but in general other devices had similar problem, it was a bad antenna design, and bad grounding / decoupling around the antenna port (give it a separate ground plane until decoupled, in no case put that directly to a board ground plane). Thanks to Andreas I could verify that and switch to a recommended layout :-)

nRF52832 ANT grounded zero impedance burned?

Top Replies