AWR2544LOPEVM: Received Power Calculation

Hello Shlomi,

could you reference this E2E thread and document for how to convert from dBfs to dBm?

(6) AWR1243: dBFs scaling in Radar Studio Post Processing in complex 1x, complex 2x and real mode - Sensors forum - Sensors - TI E2E support forums

reference the document called 'RadarStudio_matlab_scaling.pdf' in the thread.  Be sure to account for how many bits you use in the ADC.

Could you check if, using this method, the dBm value makes sense?

could you also verify your formula with the formula shown in this training video:

https://www.ti.com/video/series/mmwave-training-series.html?keyMatch=radar 

Regards,

James

Hi James,

Thanks for your reply. 

In this thread, you can see that Nathan describes the formula for calculating received power and the conversion from dBm to dBFS, which is a 10dB decrease according to the last table in your attached PDF.

In the Python code I attached in the previous message, the same calculations Nathan did are implemented, but the measured and theoretical numbers are still significantly different.

Can you confirm that all the parameters and the received power formula in the Python code are correct?

Thanks, 

Shlomi

Hello Shlomi,

Your formula is: 

Pr_lin = Pt * Gt * Gr * lambda_m**2 * sigma / ( (4*np.pi)**3 * R_m**4 ) * RX_gain

from the video I mentioned, the correct formula for the power captured by the RX antenna is:

i.e. 
Pr_lin = (Pt * Gt * sigma * Grx * lambda**2)/((4*np.pi)**3 * R_m**4)

Which looks the same as your formula, except for the Rx_gain, so your formula is after the receiver, while the formula from the video is before the receiver.

Are you sure you have configured the RX gain for 30 dB?  

Are you using all the available RX and TX channels?

Regards,

James

Hi, 

Sure, Rx gain is configured to 30dB and I'm using 1 TX / 4 RX.

The plot of 1D FFT Amplitude is after the receiver (according to my experience with mmWave studio 2 and this thread), hence, adding 30dB is the correct form to the received power.

Let me know if there is inconsistency between the studio versions for this plot.

Thanks, 

Shlomi

Hi Shlomi,

At first blush, I can see why you are confused, as my numbers also agree with your result:

After speaking with our antenna expert, it seems very unlikely that the antenna gain is so low that you would have an additional loss of (-7.1 - 31.7) = 24 dB in your system. 

May I ask you how you are getting the RCS of your target?  

Regards,

James

Hi James,

Thanks for the elaboration.

My target is a square trihedral corner with an RCS of 21.78 dBsm. It was measured multiple times using the AWR6843AOPEVM, and as mentioned earlier in the same thread, the theoretical and measured results were consistent, confirming the corner's RCS of 21.78 dBsm.

The formula for the RCS of this corner is 12*pi*a^4/lambda^2.

With the current chip, AWR2544LOPEVM, the RCS is 2.1 dB higher due to the shorter wavelength. Consequently, the power after Rx should be -5.6 dBFS, giving us a 26 dB difference.

Do you have any other ideas to explain the missing dBs?

Thanks,
Shlomi

Hello Shlomi,

I have asked my systems team and they confirmed that the delta between dBm and dBFS should be 10 dB.  

BTW are you sure about the corner reflector formula?  the one I have typically used has a formula of (4*pi*a^4)/(3*lambda^2), where a is the side edge of the three isosceles triangles used to make the reflector.  i.e. my formula has a factor of (4/3) while your formula has a factor of 12, which is quite a big difference. 

Has anything happened in your system that could lead to loss?  e.g. has the interface between the device and the antenna been disturbed?

Regards,

James

I'm using a square trihedral corner, whereas your formula relates to a triangular trihedral corner.

There hasn't been any event that could cause such a significant loss.

If you have a similar corner with an RCS of 20 dBsm, you could perform a similar experiment using my chirp configuration and share the results you obtain.

Thanks,

Shlomi

Hi Shlomi,

I can try that over the next few days, if you give me a little bit of time.  BTW have you shared your chirp configuration?  I don't see it above. 

Regards,

James

Hi James, 

Chirp configuration:

• Start frequency 77GHz
• Frequency slope 65 MHz/usec
• Sampling rate 25MHz
• Number of samples 128
• Chirp loops 31
• Idle time 5usec
• ADC valid start time 5.12usec
• Ramp end time 11usec
• Frame periodicity 0.82msec

Thanks, 

Shlomi

Thanks Shlomi, 

Please give us a day or so to try this on our end.

Regards,

James

Hello Shlomi,

We tested AWR2544LOPEVM with you chirp configurations at a 2-meter distance with a 40 square meter RCS at boresight and were able to obtain the following results. The results and sensor configuration are shown below.

It does appear there is some lost within the system, however, the loss we are getting is still less than yours. We will check with our systems team tomorrow to see if there is something that is not accounted for.

Thank you,

Gregory

Hi Gregory,

Thanks for sharing your experiment results.

As we discussed earlier, received power is calculated by the following formula:

Pr_dBFS = Pt_dBm + Gt_dB + Gr_dB + sigma_dBsm + 10log(lambda^2 / ((4pi)^3 * R^4)) + Rx_gain_target - dBm2dBFS_factor

If we do the math with your RCS (40m^2 ~ 16.02dBsm) and R = 2m, we expect to get Pr_dBFS = -9.19, which is around 16dB higher than your experimental results.

1. Do you agree with the expected received power calculation?

2. How can you explain this 16dB difference?

I conducted another experiment using the same corner (see image below), with a 23.96dBsm (the gain was accurately ensured by AWR6483AOP).

According to the received power formula, the expected value is -3.32dBFS, whereas the measured value is -27.59dBFS. This indicates a disparity of about 24dB, which is 8dB higher than your observed difference.

How can you explain this?

Thanks,

Shlomi

Hello Shlomi,

We are following up on this.  We have been informed that you need to also need to subtract ~4 dB from the link budget for the transition from the package to the antenna (2dB each way).  This makes up for some of the additional loss.

As to why your result is different from ours, its a bit difficult to say.  How accurately are you controlling the elevation angle?

We are also following up with our internal team to better understand any additional factors we should consider.

May I also ask why you are concerned about this calculation?  I ask b/c I would like to know if there is any concern you have that I'm not aware of.

Regards,

James

Hi James,

I need to detect a small target, so antenna gain is critical; every dB matters.

Elevation angle is pretty accurate, the corner is very close to the radar. 

Additionally, your measurements show a 12dB discrepancy from the theoretical received power (it was 16dB, but I subtracted 4dB from the package to the antenna).

I'm still having around 20dB difference between measured and theoretical results.

Could you provide the correct formula for the received power for this chip? 

If possible, could you also provide the correct formula for the noise floor?

Thanks,
Shlomi

Hi James,

Thanks for your reply.

I see that the analytical and measured results in your experiment are consistent.

Could it be a calibration issue? Do you have any guide for calibration with mmWave Studio 3?

Thanks,

Shlomi

Hello Shlomi,

We did not do any additional calibration to get the result.  I don't believe that should be an issue.  When you include the 2dB package loss (x2 for round trip), and 13 dB conversion factor, and also carefully check the elevation alignment do you get a more consistent result?

Regards,

James

Hi James,

Yes, I have rechecked everything, but the measured and theoretical values are still inconsistent.

I'm using Rev A of the AWR2544LOPEVM.

Which revision are you using?

Shlomi

Hello Shlomi,

We used Rev A.

Have you ever dismounted the antenna?  do you have a second board to compare to?

Regards,

James

Hi James,

I've never dismounted the antenna and don't have a second board for comparison.

I was thinking of buying another board to compare. Do you have any other suggestions?

Thanks,

Shlomi

Hello Shlomi,

It's definitely a good idea to get a second board for comparison.  You could also send us a picture of your setup.  Maybe with this we can point out some issue.

Regards,

James

Hi James,

Could you please provide your email so I can send you the setup pictures? If we find anything crucial, we'll share the conclusions with the forum.

Also, could you send a picture of your corner?

Thanks,
Shlomi

Hello Shlomi,

You can contact me at James dot Murdock at TI dot com.

Regards,

James

Hi James, 

I just sent the email.

Thanks,

Shlomi