nRF52832 WLCSP errata receive sensitivity

[ProfHindsight]

The errata for the nRF52832 WLCSP states -4dB sensitivity hit using the DCDC converter and -2dB sensitivity hit using the LDO. Will this be fixed in the final revision of the IC?

Will adding an external LNA reduce the loss or is it a noise problem further down the RF chain?

Thanks!

[AmbystomaLabs]

Desense from local phenomena is very normal.  Likely Nordic can't do anything about it as the silicon is set in stone at this point and it may have been unavoidable. It really depends on how the desense is happening to determine if an external LNA will help. Plus adding an external LNA reduces the overall dynamic range of the receiver and can bring with it many other (sometimes serious) problems.

The easiest way to figure this out is to approach it empirically.  Do a spreadsheet or a model of the receiver chain.  Measure the receiver sensitivity and then add in gain of a known noise figure and measure the sensitivity again. If the experiment matches the model then you can run with the LNA solution as long as you get the dynamic range you want and protect against damaging the nRF52832 front end.

[Andreas]

Hi,

 

There is no revision planned and this errata is here to stay. If you can not live with the 2 dB sensitivity degradation your options are to use the nRF52832 QFN package version, or nRF52810/nRF52840 in WLCSP package version (WLCSP for these do not have a similar degradation).

 

Using an external LNA is technically an option also. It will not fix the issue, but effectively reduce the sensitivity by up to a couple dB. If your design can accommodate an external LNA you should not have any issues accommodating the QFN package version though.

 

Best regards,

Andreas