ADRV9029 stability erro

Q1: Follow this link, https://wiki.analog.com/resources/eval/user-guides/adrv9029/dpd_model_optimization

Q2: There are no. of metrics as part of the stability, refer DPD UG. To summarise, if the Tu Power/ Tx power/ ORX power are saturated, you could see the stability error.

Q3. This could be due to high power going into ORX port which is saturating the ORX or a bad matching at the ORX port.

Can you share us the DPD tracking config set parameters?

Please refer the below link as well for troubleshooting DPD errors.

https://wiki.analog.com/resources/eval/user-guides/adrv9029/dpd_error_troubleshooting

Hi all

I have same Error

1. As I checked, Testing DPD with TES GUI and EVALUATING ADRV9029 DPD [Analog Devices Wiki] show 2 different procedure to test DPD (one involves turn off Tx, while the other don't). Which one is the correct/ latest one?

2. As I tried so far, the STABILITY_ERROR persists for ALL pre-built DPD models (swept using Python scripts), so I think I set something incorrectly. So far, I can guarantee that Tx, Tu and ORx power is far from saturated, as in the following pictures.

3. I don't believe so, as the current average power at ORx port is -27dBm, with 12dB PAPR so it would be -15dBm peak, still far from saturated.

4. Also, my current average power at Tx is -16dBm, with 12dB PAPR so it would be -4dBm peak, also far from saturated.

Q1: If you are using GU version > 5.0.0.68, please use the one uploaded here.

 Testing DPD with TES GUI 

Q2: Can you change the M-Table threshold to ~-25dBm or can you increase the TX power to -16dBm (currently -20dBm); Change the peak search window size to 491520.

What is the carrier BW, if you are using NR 100MHz, try enabling direct learning and check?

Also, try with Regularization value of 25.

Q3: You can try increasing the ORX gain index by another 10dB and check the DPD performance

Q4: But from the DPD statistics, the Tx power is showing -20dBFS?

Also, try checking the DPD performance with 3dB backoff power from the PA max. rated power. 

1. In that case, does the sweeping Python code work? I don’t see it turning off Tx at any point.

2. Tx power is -16dBm currently (image in Q4). BW is 20MHz (bottom of image in Q4)

3. I remember it’s 255 already.

4. What is the expected value?

Then please try to increase the power going into ORX port and check the DPD performance

Also, try adding an attenuator of 6dB and check the performance

The Peak ORX power can go upto -3dBFS.

So far, I did all your recommended settings, but still nothing is solved.

This is the block diagram of my model:

This is my signal at Tx and ORx, respectively (power showed in the bottom of the images):

However, with all your recommended setups, the error is still not solved:

What should I do next?

Ramarao said:

What is the carrier BW, if you are using NR 100MHz, try enabling direct learning and check?

Have you tried direct learning?

Ramarao said:

Also, try checking the DPD performance with 3dB backoff power from the PA max. rated power. 

Have you tried this?

Can you check with a different model file once?

1. Not yet, as I'm using a GaAs PA and LTE-20MHz signal. Will try tomorrow.

2. Already. The image for ORx constellation is not too distorted.

3. I try by hand about 10 of them. None works. In addition, I don't understand which step did the GUI train the DPD? As loaded from the file, every coefficient is 0.

I tried 20Mhz and 100Mhz

i found 20Mhz with DPD working without problem  and i obtained around 18db correction:-

but 100 MHz no more 10db correction