• State Resolved
  • Locked Locked
  • Replies 11 replies
  • Subscribers 19 subscribers
  • Views 1435 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.

CC1120 sensitivity @169MHz

Xavier TATOPOULOS
Prodigy 110 points
Other Parts Discussed in Thread: CC1120

Hello,

We are testing CC1120 EVK at 169MHz. We currently have the following performances in sensitivity :

- 1.2kbps 2-FSK : -119dBm

- 4.8kbps GMSK : -112dBm

- 2.4kbps 2-GFSK : -117dBm

- 38.4kbps 4-GFSK : -101dBm

Do you confirm this performances ? It seems that we have 3-4 dB less than announced in datasheet...

Thanks for support. Regards,

Xavier

  • 0
    TI__Expert 3595 points

    That is low.  What is the deviation, filter bandwidth, and are you correcting for crystal frequency errors?

  • 0 in reply to Chris M
    Prodigy 110 points

    Thanks for quick answer.

    Below the configuration applied :

    For the crystal compensation, AFC is automatically turned on by Smart RF Studio.

    Regards,

    Xavier

  • 0 in reply to Chris M
    TI__Guru**** 317180 points

    For 1.2kbps, 4kHz deviation, 2FSK you should manage -123dBm.

    How do you do your measurements?

    The best method is to use a signal generator that sends packages

    If that is not possible, set a signal generator to send a '1010' pattern and set the CC1120 in synch serial mode.

    I have found that measurements with 2 CC1120EMs with an attenuator between gives inaccurate results becasue it is difficult to measure the exact attenuation.

  • 0 in reply to TER
    Prodigy 110 points

    We do the measurements using 2 EVK connected through variable attenuator, and then we use Smart RF Studio in expert mode with Packet Tx and Packet Rx functionalities.

    Regards,

    Xavier

  • 0 in reply to Xavier TATOPOULOS
    TI__Guru**** 317180 points

    Please also note that the RSSI value reported back from Studio is around 5dBm off for CC1120 in version 1.5.0. The RSSI offset is 101, not 96 as v1.5.0 uses. This will be corrected in v1.6.0 that will be released shortly.

    If you have a signal generator it is highly recommended to use this and if you don't have that possiblity double check what your attenuation really is. I found when I did similar tests that a variable attenuator has a insertion loss that has to be taken into consideration and you need to know the exact output power on the sender.

  • 0 in reply to TER
    TI__Expert 3595 points

    Also a power meter can be used to get the exact value of the power going into the chip.

  • 0 in reply to Chris M
    Prodigy 110 points

    Thanks for support.

    I think the main difference is due to error rate definition : it is not written in the datasheet that sensitivity is for 1% BER, and not PER.

    It can make a big difference in terms of measurement (at least 2-3dB).

    We will check in this way and keep you inform.

    Xavier

  • 0 in reply to Xavier TATOPOULOS
    TI__Guru**** 317180 points

    With 3byte payload 20% PER equals 1% BER.

    I noted that you wrote that AFC is activated in Studio. To use the calculated freq offset estimate you need to send a SAFC strobe. So for the narrowband test case please check that you sender and receiver is actually operating on the same frequency.

  • 0 in reply to Xavier TATOPOULOS
    Intellectual 320 points

    Hello All

    We did a similar test with 2 x EVK+CC1120 @ 868MHz, and Var. Attenuator in between.

    We got a 6-7 dB worse sensitivity then expected (-100dBm instead of -107).

    Q1: How to explain the results?

    Q2: How do I add FEC? Is it "built in" to the data?

    Regards

    HN

  • 0 in reply to HNRFGUY
    TI__Guru**** 317180 points

    Q1:First of all, if you can measure sensitivity as described here: http://e2e.ti.com/support/low_power_rf/f/155/t/133399.aspx, please do that. Have you measured the power level from the CC1120 and the exact attenuation or if you use RSSI, taken into account the error in RSSI reporting in v1.5.0 of Studio?

    Q2: FEC is not implemented on CC1120

  • 0 in reply to TER
    TI__Expert 4435 points

    You might try what we done in our lab at Ti,  but have found that their is to much back ground noise to very low sensitivty meaurements.  Try placing your DUD in metal boxes or a metal heat chamber.  This will reduce background noise due to wifii and others sources.  You should see more consistant values.

    Regards