RSSI Accuracy in dBm

[Thavish]

Hi,

The documentation for the nRF52840 (section 6.20.15.9) states that RSSI has a typical accuracy of ±2 dBm. I have measured signal strength at multiple distances, and the resultant standard deviation of my data at each distance agrees with this value. However, I am surprised that standard deviation stays approximately constant at ±2 dBm, since dBm is a logarithmic scale. I would have expected standard deviation (accuracy) to remain constant with respect to SI units instead.

For example, after taking multiple measurements of signal strength at 20 cm and 50 cm, I calculated an average power of -20 dBm and -45 dBm respectively. Each set of values had approximately equal standard deviations of 2 dBm. However, after converting to SI units, the -20 dBm dataset obviously had a much larger variance than the -45 dBm dataset.

It appears as though there is a systematic error of up to 2 dBm present in readings. The only explanation I can think of is that this error is added during signal conversion by the ADC. Any clarification on this would be much appreciated.

Best wishes,

Thavish

[Marjeris Romero]

Hi,

You are comparing two different things. The RSSI accuracy value that is listed on the Product Specification for nRF52840 refers to the RSSI accuracy internally on the radio.

The RSSI variation when radiated in the channel is a different thing, as many more variables are added, specially when measuring indoors, where you have reflections, multipath and interference.

If you want to measure the RSSI accuracy of the radio you need to perform the measurements conducted. This means using a cable soldered as closed as possible to the matching network on the nRF52840 and sending a constant carrier with a signal generator or protocol tester and reading the RSSI values then.

Best regards,

Marjeris

[Marjeris Romero]

Hi,

Just another comment, the RSSI accuracy of the radio in the PS is given in dB, not in dBm. dB is a ratio, so the same ratio of accuracy (percentage) will apply for both -30dBm and -90 dBm. I think this is usually called relative accuracy in the statistics instead of absolute accuracy.

Best regards,

Marjeris

[Thavish]

Thanks, that's helpful. Can I check that I understand the RSSI measurement process right? After a signal is received by the antenna, it is passed to the SAADC, then to an IIR filter which outputs a value in dBm. The internal accuracy of the SAADC is ± 2 dB.

[Thavish]

Is there a physical explanation as to why the percentage accuracy is the same for both values? It seems strange that the internal radio doesn't have a fixed absolute accuracy instead.

[Marjeris Romero]

Hi,

The receiver has it's own ADC, it has nothing to do with the SAADC peripheral. The accuracy of +-2dB is for the RSSI value.

Best regards,

Marjeris