Request for Assistance with Improving TX Performance with 5GNR BW 10MHz.

More details When we push the power level to around 24 dBm, the ACLR is not good,

Please identify the problem if its related to CFR and/or DPD? Please share the ACLR plot before applying CFR. 

Please share the CFR configuration and CFR statistics and CCDF plot before and after applying CFR.

Please test only at TRX TX out for CFR performance and once you are clear on the expected CFR performance, you can integrate the PA and then test DPD.

Please share DPD tracking configuration, DPD statistics as well.

Thanks for the response. The first picture is what without CFR and DPD as Paul posted.

Looking at the picture with CFR and DPD ACLR Plot at 24dBm, i suspect that the CFR may be doing something wrong as i can see spurious. Lets isolate the issue between CFR and DPD by reviewing the statistics and other measurements as requested in my previous comments.

Hello Ramarao,

When CFR is disabled, the result becom better.

The CFR configurations are:

cfrMode = 0,
cfrTxDelay = 511, 
cfrPeakThreshold = 0.48,
cfrEngine1PeakThresholdScaler = 0.98, 
cfrEngine2PeakThresholdScaler = 0.985,
cfrEngine3PeakThresholdScaler = 0.99,
cfrCorrectionThresholdScaler = 0.98,
cfrInterpolationFactor = 2,
cfrEngine1MaxNumOfPeaks = 5, 
cfrEngine2MaxNumOfPeaks = 5,  
cfrEngine3MaxNumOfPeaks = 5
cfrEngine1Enable = 1,
cfrEngine1BypassEnable = 0,
cfrEngine2Enable = 1,
cfrEngine2BypassEnable = 0, 
cfrEngine3Enable = 1,
cfrEngine3BypassEnable = 0
cfrHardClipperEnable = 1,
cfrHardClipperThreshold = 0.55

And I'm using the CFR pulse from Request for Assistance with CFR Pulse File and Threshold Settings for BW 50MHz 5GNR Signal - Q&A - Design Support ADRV9026/ADRV9029 - EngineerZone (analog.com).

I have generated the pulse with 4x interploation. Please use Interploation factor of 4 in the configuration as well. As mentioned earlier, please check CFR performance alone and then check DPD performance. Please make sure that the CFR is working as expected and then move forward with DPD.

Please share the CFR Statistics to see if any peaks are skipped out of engine 3.

I got follow CFR statistics:

It is strange that cfrEngine2PeaksSkipped is 0.

CFR statistics look to be fine. Can you please compare CCDF, ACLR plots with and without CFR at Transceiver TX out? If you don't see any problem, you can go ahead with DPD.

Hello Ramarao,

Please refer to below results of transceiver output:

To me it looks like you are using 20MHz Pulse file and hence the Spectral distortion. Can you load 10M pulse file?