nRF52840 2nd and 3rd harmonic out of spec

[Johannes]

Hello!

We're building radio modules with the nRF52840, one with the a frontend amplifier and one without. Both are exhibiting excessive 2nd and 3rd harmonic emissions.

We're trying to narrow down the source of these harmonics and in doing so we tested the nRF52840dk devkit. Running a simple firmware that generates a carrier at channel 80 with random data, and configuring the radio to output 8dBm we measure about 4.4dBm for the 1st, -48dBm for the 2nd, and -40dBm for the 3rd harmonic. Measurements were done over coax cable.

The datasheet specifies -51dBm for the 2nd, and -48dBm for the 3rd harmonic, "in IEEE 802.15.4 mode". Is there a maximum specified anywhere? Or something that's related to the output signal power?

There could be some slight mismatch in our cabling but it shouldn't result in an 8dB difference, right?

Is this behaviour according to spec or are we either measuring it wrong or interpreting the specification wrong? I'd appreciate any kind of insight.

See screenshots for details:

[Andreas]

Hi,

 

Both plots are measured on a devkit without the external PA, right? The harmonics should be specified using RMS detector, as specified in the FCC DTS measurement guidance, this should drop the reading to somewhere closer to the spec.

 

The source of the harmonics is the RF PAs, in order to attenuate them you should take care when designing the external matching network and tune the components as needed. For the module with the external PA, the external PA is likely to be the dominant contributor as it is higher power and the nRF52840 should be operating in a lower power setting where the harmonic generation is not as aggressive.

 

Best regards,

Andreas

[Johannes]

With RMS detector the overtones do come under spec, but the carrier wave is well under the 8dBm I configured. With peak detector the carrier is about 6.5dBm. Is this expected behaviour?

[Andreas]

Hi,

 

Sounds reasonable, output power is to be measured with peak detector. Measured with a probe cable on a DK you can probably account for at least 1 dB loss.

 

Best regards,

Andreas

[Johannes]

Okay, then everything seems in order. Thank you very much for the insight! It'll really help us in our search.

Best regards,

Johannes