The Error Vector Magnitude (EVM) performance of the Nordic7002EK in the 5G frequency band is not good

[zyan]

Hi all,

I am using the Nordic 7002 EK board with the Nordic 5340 DK board.

When I conducted the production testing (using IQ xstream), I found that for TX-HT-MCS7 in the 5520 channel, the TX power was 14.5 dB but the EVM was only about -20 dB (usually it is about -27 dB).

I checked the OTP area to see if I could decrease the TX power to improve the EVM, but only the MAC address or the CALIB can be changed.

Is there anything I can do to improve it or is this the limit for the EVM test?

Thanks & regards

[JONATHAN LL]

Will look in to this and provide more info later on. 

Can you share what NCS version you are using? 


Regards,
Jonathan

[zyan]

Hi,Jonathan

Thanks for your quick reply. I used the radio_test sample and my NCS version is 2.6.0.

[JONATHAN LL]

Hi,

Can you share more info for us to reproduce: 

1. Test results with latest NCS (2.7.0 rc) - please specify the NCS SHA used in your testing

2. Test setup, NCS sample used for testing and detailed test procedure will really help us understand better (hopefully all cables used are calibrated?)

3. Can you share the hex file that you are using for nRF5340DK+nRF7002EK board?

4. Any VSA EVM measurement plots will be very useful.

[zyan]

3312.zephyr.hex

Hi, here are the details about the power and EVM values I measured with the 2.6.0 SDK and the 2.7.0 RC2 SDK, along with the .hex file I used.

[JONATHAN LL]

Thanks, I will have forwarded the info to our experts and will update you if we need anything more or have some results or comments. 

Regards,
Jonathan

[responsible_duck]

Hi JONATHAN LL just to add, we are seeing similar behavior (unrelated to the original poster or project). 

[JONATHAN LL]

Can you (responsible_duck ) create a new ticket here on Devzone? Link to this ticket as well in that new ticket and we will follow up.

[JONATHAN LL]

Hi zyan , 

1. All our measurements refer to output of RF pin. Are the power numbers measured (by you here) refer to output of RF pin? Or measured at the output of SMA connector?

2. What is the command sequence used for these measurements? Please share the command sequence.

3. Please confirm that hex shared by customer is built with NCS release 2.6.x codebase.

[zyan]

7180.zephyr.hex

Hi, Jonathan

Thanks for your replay.

1. We measured at the output of SMA connector.

2. We use the commands from wifi radio test, here is an example about how we test  G mode in channel 2412.

wifi_radio_test init 1
wifi_radio_test tx_pkt_tput_mode 0
wifi_radio_test tx_pkt_rate 54
wifi_radio_test tx_power 18
wifi_radio_test tx 1

3. Sure, here is the version of my build environment, and I upload the radio_test 's .hex file for 5340DK & 7002EK if you need it.

[JONATHAN LL]

Thanks! have forwarded internally and will update when i have something. 

[JONATHAN LL]

Hi zyan ,

1. Our measurements in product spec are taken at the output of RF. Please compensate for PCB losses or any other losses between RF pin and SMA connector.

2. Get power measurements without issuing tx_power command (i.e., wifi_radio_test tx_power 18 in your command sequence below) . This will provide measurements with default tx power settings. Use NCS 2.7.0 codebase for these measurements.

[zyan]

Hi，Jonathan

1. We added the losses between RFpin and SMA connector.

2. I downloaded the latest sdk 2.7.0 and tried again, it did not work, here is the data I got.

And I want to ask the following questions:

1. Does the nrf5340DK combined with the nrf7002EK perform worse than the nrf7002DK in testing? I think the nrf7002DK might perform better.

2. Do you have reference data for the EVM test of the nrf5340DK  with nrf7002EK?

3. Is there a way to adjust the power settings in the register or OTP for the nrf7002 module? This adjustment could potentially resolve the issue and help us with out production testing.

If this performance is the best that the module can do, we might accept it.

Thanks for your help!