Quantization noise

Hi,

We will get back to you.

Regards

Rahul

Hi, 

I have tested out your configuration and I don't see any issue with it. Please find the calculation for obtaining the fundamental amplitude and RSSI value. 

Rx_Power_TES (Fundamental Amplitude) = Rx_Input + TIA_Gain - ADC_Fullscale + Interface_Gain - Attenuation - Mismatch_Loss

In order to find the  fundamental amplitude and RSSI value you need to know what is the Rx input power at the port (Rx_Input )

what is  the insertion loss (Mismatch_Loss). This value is frequency-dependent and should be taken from the user guide. In this example, we use the value at approximately 900 MHz. So the insertion loss will be 1.3 dB @ 900 MHz 

What is the attenuation value which is 0 dB in the below example. 

 The ADC full scale is 8.6 dBm.

TIA Gain which is 20dB. 

Case 1: When the interface gain is 18dB. 

Rx_Input = -34;

TIA_Gain = 20

ADC_Fullscale = 8.6

Interface_Gain = 18

Attenuation = 0

Mismatch_Loss = 1.3

Rx_Power_TES = Rx_Input + TIA_Gain - ADC_Fullscale + Interface_Gain - Attenuation - Mismatch_Loss  =  -5.9 dBm

As you can see from the below screen grab the fundamental amplitude is -6 dBm which matches the calculation

The RSSI value will be: Rx_Power_TES  - Interface gain = -5.9 - 18 = -23.9dBm (which is the same as the RSSI value shown in TES)

Case 2: When the interface gain is 0 dB. 

In this case the RSSI value and the fundamental amplitude value will be same. 

Rx_Input = -34;

TIA_Gain = 20

ADC_Fullscale = 8.6

Interface_Gain = 0

Attenuation = 0

Mismatch_Loss = 1.3

Rx_Power_TES = Rx_Input + TIA_Gain - ADC_Fullscale + Interface_Gain - Attenuation - Mismatch_Loss  =  -23.9000 dBm

As you can see from the below screen grab the fundamental amplitude is -24 dBm which matches the calculation and the RSSI power is the same as the fundamental amplitude since the interface gain is 0.

Regards

Rahul

Hi Rahul,

Thank you very much for the detailed response and demonstration.
I want to update two settings that I forgot to mention:
1. The SSI type is: CMOS, 1 Lane, Single, Long Strobe.
2. The Intermediate frequency is 490kHz.

Additionally, just to clarify, when you wrote Rx_Power_TES (Fundamental Amplitude), did you mean measurement units of dBFS instead of units of dBm? (From the FFT plot it seems the units are in dBFS).

I would appreciate a response regarding four additional points –
I would like to receive an FFT plot when the input to the ADRV9003 is 50 ohms instead of a CW (and the corresponding RSSI Power value).
How is the RSSI calculated? Is it the sum of all the energy under the bandwidth? In this case, 12kHz or 24kHz? Or is it a different calculation?
In both FFT plots, the S/N is 85dB, so why does the slicer matter then?
Is there a rule of thumb for what power level is recommended to feed into the ADC? I'm concerned that I might have an issue receiving signals at low power, and that all that's needed is just additional gain.

Thank you again,
Avner

Hi,

I have tried with the additional setting as well, still I don't see any issues.

AvnerA said:

Additionally, just to clarify, when you wrote Rx_Power_TES (Fundamental Amplitude), did you mean measurement units of dBFS instead of units of dBm? (From the FFT plot it seems the units are in dBFS).

Yes that is correct, it should be dBFS.

Question: "I would like to receive an FFT plot when the input to the ADRV9003 is 50 ohms instead of a CW (and the corresponding RSSI Power value)."

Ans: You can do it by selecting an appropriate interface gain value as shown below. In this case I have used the interface gain as 6dB and then received the signal. You can see the FFT plot as well as the RSSI value. 

How is the RSSI calculated? Is it the sum of all the energy under the bandwidth? In this case, 12kHz or 24kHz? Or is it a different calculation?

The RSSI is calculated using the samples and the number of samples are determined by the RSSI interval, higher the number of samples the more accurate the RSSI measurement. The RSSI calculate the the energy within the bandwidth, in this case its 12 kHz

Question: "why does the slicer matter then?"

Ans: Slicers are used to prevent signal data from becoming saturated when the signal power is high, while also ensuring the system remains sensitive to weak signals. The system determines the optimal setting for these slicers based on the RSSI, but you can also adjust them manually.

Question: "Is there a rule of thumb for what power level is recommended to feed into the ADC? "

Ans: The minimum power level that the ADRV9002 can detect isn't a fixed value from the component itself; it's a system-level performance metric. This is because the lowest signal a receiver can successfully pick up is determined by the overall system's noise floor, which depends on factors like bandwidth and the noise figure of the entire system, not just the chip on its own.

Regards

Rahul

RahulMushini said:

Question: "why does the slicer matter then?"

Ans: Slicers are used to prevent signal data from becoming saturated when the signal power is high, while also ensuring the system remains sensitive to weak signals. The system determines the optimal setting for these slicers based on the RSSI, but you can also adjust them manually.

Question: "Is there a rule of thumb for what power level is recommended to feed into the ADC? "

Ans: The minimum power level that the ADRV9002 can detect isn't a fixed value from the component itself; it's a system-level performance metric. This is because the lowest signal a receiver can successfully pick up is determined by the overall system's noise floor, which depends on factors like bandwidth and the noise figure of the entire system, not just the chip on its own.

Thank you Rahul.

I want to calculate the Noise Figure of my setup with ADRV9003. The setup includes some cables, connectors and a splitter.
I apply a CW from signal generator at -60dBm. ADRV9003 at maximum gain (0dB Frontend Attenuation).
The RSSI measurement I see is 55.2dBFS which means the real power is 8.6dBm - 55.2dB = -46.6dBm
The power before the TIA gain is -46.6dBm - 20dB = -66.6dBm.
So the LOSS between the signal generator to the ADRV is 6.6dB. 
The formula I use is:
NF(dB) = -Noise(dBFS) + Signal(dBFS) + Signal(dBm) - 10*log_10(BW) + 174 . (When BW is 12kHz as you mentioned).
I don't take into account for the external attenuation I have like cables, connectors and the splitter. So that the Noise Figure I measure includes these components.
The measurements:
Noise(dBFS) = 109
Signal(dBFS) = 55.2
NF = -109 + 55.2 - 60 - 10*log(12,000) + 174 = 19.4dB
NF_ADRV9003 (When I take into account the external loss) = 19.4 - 6.6 = 12.8dB (Which makes sense with the datasheet)

Assume that my receiver has a requirement for SNR to be 5 [dB] and BW of 3 [kHz]. The sensitivity should be:
Receiver_sensitivity(dBm) = -174 + NF + 10log(Receiver_BW[Hz]) + SNR_MIN
For my setup: -174 + 19.4 + 10*log(3000) + 5 = -114.8dBm

Does it mean that any signal stronger than -114.8dBm will receive at my receiver? Assuming Frontend Attenuation is 0dB, and Slicer is 18dB.
Do I need to make any other unique configurations to achieve that?

Thank you again.
Avner

Hi,

yes, any signal stronger than -114.8 dBm should be successfully received and processed by your receiver, assuming all other conditions (like interference) are ideal.

Regards

Rahul