• State TI Thinks Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 1 answer
  • Subscribers 53 subscribers
  • Views 625 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

CC1200: How to reduce the TX spike on/near adjacent channels ?

user5217955
Prodigy 180 points
Part Number: CC1200


Hi,

the TX spectrum of my product is like this, and you can see two distinct spikes. But, the same product, once got this transmission spectrum, very smooth. (Although the frequency setting is different, it does not affect the result.)

My question: how to reduce the  TX spike on/near adjacent channel ? What can I do to get the previous result ?

Thanks and BR,

5633.ACPR.docx

  • 0
    TI__Guru**** 317180 points
    It could look like this measurement is done sending packets. Is that the case and is that according to the spec you are measuring against?
  • 0 in reply to TER
    Prodigy 180 points
    The test is a continuous sending mode, bit stream, not packets. These bits need to match PN9 code format.
    What I saved was the waveform after sending, in order to look clear. The waveform in the transmitting is same, has two peaks in/near the alternative channel.
  • 0 in reply to user5217955
    TI__Guru**** 317180 points

    I did a test with this setup:

    on a CC1200 EM and got the following result:

    As you see I don't get any spikes. Do we test in the same way? 

  • 0 in reply to TER
    Prodigy 180 points
    No, we did not use same way. I tested the real signal spectrum of our products.

    But your result is very helpful.
    1) Would you export registers of the wave and 150kbps's ? I will compare the difference of register setting.
    2) We normally uses 920MHz. The spectrum of 920MHz is similar with 868MHz ? If it is ,would you send me export registers of 50kbps and 150kbps ?
    Thanks
  • 0 in reply to user5217955
    TI__Guru**** 317180 points
    Before starting to look at different datarates etc it would be useful to find out why the sidelobes are so high.

    Are you able to use the exact same setting I used on your board? My question was more if except for the actual hardware used the test method was the same. I checked with the center frequency @920 and the spectrum is the same.
  • 0 in reply to TER
    Prodigy 180 points
    I want to use the exact same setting in my board, that is why I'd like you to export registers to me.
    The setting is same on xtal frequency 40MHz, symbol rate 50kbps, RX Filter BW: 104KHz, 2GFSK, deviation 25KHz, continuous TX, no whiten. The differences are carrier frequency, and the bit stream is not dual sync word.
  • 0 in reply to user5217955
    TI__Guru**** 317180 points
    You can get the exact same setting I used by using a register export in your end.

    Since I assume that you are running this test from your own code, have you verified that you are actually are not sending packets? (use zero span on the spectrum)
  • 0 in reply to TER
    Prodigy 180 points
    I did not set up the TI tool. Maybe I will. Would you please send it to me first this time?

    Yes, I am running the test from my own code. From my code and eye diagram (eye diagram needs continuous bit to form the normal eye), I think it is not sending packets, it is bit stream. I tried to use your way to check, but my test equipment does not support the mode. In my test equipment, if it is in packets transmit, the spectrum can't keep showing up.

    Is there any difference in spectrum between bit stream and packet transmitting ? As far as my code, they are same.
  • 0 in reply to user5217955
    TI__Guru**** 317180 points

    Does that mean that you haven't installed SmartRF Studio, and if so, where did you take the settings in your code from?

    And you should get a different spectrum sending packets vs sending packets, see below for what I measuring sending packets:

  • 0 in reply to TER
    Prodigy 180 points

    Yes, I haven't installed it. The code has a long history.

    Would you tell me why the spectrum of bit stream is different from packets data ? I don't quite understand it. Thanks.

  • 0 in reply to user5217955
    TI__Guru**** 317180 points
    If you first look at sending packets in the time domain: First you have no power, than x dBm for the duration of the packet, then no power and so on. Or in other words something that look like a square wave. In the frequency domain this looks a lot different compared to just sending continuously TX data since in that case the power is always on.

    That said the spikes you have looks a lot larger than I get which could indicate also issues with the schematic/ PCB. If you send me a friend request with the schematic (pdf) and layout (gerber) I can take a quick look and see if I see something.