AWR2944EVM: Phase noise performance

Hello,

Thank you fo rreaching out to Texas Instruments. There can be many reasons to the higher phase noise. Do you see this behavior on multiple EVMs or just the one?

-E. Shareef

Also, Could you move your location and confirm if the number stays the same? There is a possibility of another source in your vicinity. Changing the distance between your EVM and horn can also confirm this. 

Hello,

   Thank you for your response. I did the testing on two EVMs based on ES2.0 and ES1.0. There are the same performance of PN on these. I used the settings listed below.

I tested the continousStreaming mode, and got the better phase noise, which is near to -94dBc/Hz. But, I set slop rate to 0 in chirp mode and the phase noise is worse.

Ming

Can you please share your Frame config ? 

Here it is.

Set number of frames to zero. Please let me know how the PN changes. 

I use the one device to test TX spectrum in different modes.

The settings in CW mode is pasted below.

And the meausred result of PN looks great.

Then, the settings in chirp mode is pasted below. I increase period time to get the higher duty cycle.

And the meausred result of PN is worse. There are so many spurs near the carrier.

Ming 

Could you please try to go 79 GHz fc? In chirp mode does that still give the same phase noise? 

I am looping in a few team members to debug this 

-E. Shareef

The phenomenon is the same. I think it's irrelevant to carrier frequency. I suppose awr2944 has high slew rate in current when is operating in chirp mode, and the power rail in RF may has some ripple and noise. It leads to bad performance in PN.But, how to fix this?

Hello Ming,

Can you share details of your test setup? Are you workin gin a low noise environemnt as an Anechoic Chamber?

Thanks,

-E. Shareef

Hi Ehtesham,

     The setup of test is shown in this figure. We used a signal analyzer with mixer to capture frequency operating between 76~81GHz. We also did the test in an anechoic chamber. The meausured results are the same.

Ming

Thanks for the update Ming,

Were you able to run this test on a second board ( I don't know if you have access to one)

Also, could you runt the same test again activating one TX at a time?

Thanks,

Shareef

Hello Ehtesham,

   1. I found the same phenomenon in EVB ES1.0.

   2. I test the tx spectrum vary with the different number of working tx channel.

Only tx1 is on.

tx1 and 2 are on.

tx1&2&3 are on.

All tx are on.

All of them have worse performance with high noiise  in side-band.

Ming

Hello Ming,

The results you are seeing are not inline with what TI has observed on the device. In order to investigate further, I am planning on re-creating your test setup and try to take readings accordingly. I will respond with an update by the end of the week.

Thanks,

-E. Shareef

Hello Ming,

I was able to confer witha few colleagues that with a regular Spectrum Analyzer, you will not be able to read PN in chirped mode. The phase noise analyzer needs to ‘lock on’ to the signal and then analyze the noise. By the carrier constantly moving, the PLL loop cannot lock to the desired signal.

Hope this helps,

-E. Shareef

Hello Ehtesham,

    In our experiments, the duty circle is nearly 100%，and slope rate is set to zero. The spectrum analyzer can lock on to the signal. I think  FMCW may deterioate PN.

Ming