NRF52 RF Front End and link budget with/without external LNA

[atum]

Dear Nordic support Team

I have a question regarding the performance/specification of the RF front end of the NRF52 series chips.
I noticed that the question "Is there an AGC in the RF front end of the chip" has been asked twice and there two differents answers ...

devzone.nordicsemi.com/.../agc-for-ble => no AGC
devzone.nordicsemi.com/.../agc-configuration-documentation-avoid-spikes-in-rssi-sample-data => yes AGC

I'm trying to calculate (roughly) a would like to know if i am right ...

The noise figure of the NRF52832 is not given in the datasheet but:

The datasheet says: Sensitivity, 1Msps BLE ideal transmitter, 37 bytes BER=1E-3 => sensivity -96dBm

=> I assume the noise equivalent bandwidth of the demodulator input filter is 1MHz
=> I assume for a 10-3 BER, the Eb/No = S/N = 15dB within the 1MHz bandwitdh (GFSK modulation 1 symbol = 1 bit)

-174dBm/Hz + 10log(1e6) + 15(SNRLoop) + NRFNoiseFigure? = -96dBm input signal

The noise figure of the NRF (deduced) is 3dB (not aberrant for me)
As far as i know (with my limited RF knowledge) if there is NO AGC, the noise figure is constant on the whole input dynamic range but in the opposite, the noise figure should be minimal at AGC maximum gain and will increase at lower gain (input signal strengh increase)

The nordic external LNA product brief => www.nordicsemi.com/.../nRF21540-PB-v2.0.pdf

says ... 13dB of Gain and 2.7dB noise figure ... sensivity increased of 5dB

same reasoning as above ...

-174 + 10*log10(1e6) + 15(SNRLoop) + 3(NRF noise figure) + 2.7(LNA noise figure) - 13(LNA GAIN) = - 106 dBm

I miss 6dB !!

So my conclusion is the NRF frond end use an automatic gain control and because of the LNA gain in front of it,
the gain of the NRF internal AGC is lower and thus, the noise figure is increased of 6dB and this is the reason why
the global sensivity improvment is only 5dB.

Does anybody at nordic can confirm that my understanding is correct and also confirm if the NRF52 front end use an AGC ?

Best regards.

Atum

[Bendik Heiskel]

Hi,

What are you trying to achieve with these calculations?

The total noise figure for a network of cascaded two port devices is not the sum of the individual noise factors, so this calculation is not correct.

confirm if the NRF52 front end use an AGC ?

There is a AGC in the receiver front-end.

 

Best regards,

Bendik

[atum]

Thank's for your response (and for pointing my mistake).
With my calculations, i'm trying to understand where the degradations of the SNR comes from
(why there is only 5dB of receiver sensivity improvment whith a 13dB LNA gain)

For the NRF52 Alone (if my understanding is correct) the is no mistake since the is only one stage (NRF52 AGC Gain)

-174dBm/Hz + 10log(1e6) + 15(SNRLoop) + 3(NRF52 estimated noise figure) = -96dBm input signal

I fix my mistake for the two cascaded devices : LNA (NF=2.7dB, G=13dB) and NRF52

I use the following formula (all values are linear)
NF(total) = LNA Noise + (NRF52 Noise – 1) / LNA Gain

-174dBm/Hz + 10log(1e6) + 15(SNRLoop) +  (Total Noise Figure) - 13(LNA Gain) = -100dBm (witch is the max sensivity given in app note with external LNA)

The total Noise figure is estimated as 12dB. Since the noise figure of the LNA is 2.7dB and it's gain is 13dB
the noise figure of the NRF52 is estimated at 24.5 dB

Do you agree ?
Can i assume that, the degradation of the NRF52 noise figure is due to it's AGC ?? (lower gain, due to external LNA gain, and thus worst noise figure ?)

best regards

[atum]

Bendik Heiskel