• State TI Thinks Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 3 answers
  • Subscribers 10 subscribers
  • Views 503 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TRF372017: Image Rejection

SM
Expert 1730 points
Part Number: TRF372017
Other Parts Discussed in Thread: DAC3484,

Hello,

    We are using the TRF372017 in conjunction with a DAC3484 to generate a modulated carrier. We are generating single tone cosine and sine and are feeding them to the I and Q of the modulator. We are getting equal sidebands at the moment (when there in concept should only be one). (f_LO +/- f_IF, both are present). 

Is there some setting that we are missing? 

Thanks for your help in advance,

  • 0
    TI__Guru 66014 points

    Hi SM,

    you can first check the waveform at the DAC3484 output are truly 90 degree phase shifted. After that, please also check the interface network is matched correctly for the common mode level requirement for both the DAC and the TRF part. Below app note will help:

    -Kang

  • 0 in reply to Kang Hsia
    Expert 1730 points

    Hello Kang,

             We checked the signals on an oscilloscope and they appeared in quadrature (the scope measured a phase difference of 97 degrees). We also checked the dc levels on them at R62 and R54 with respect to ground. One of them was at 1.37V and the other at 2.11V. Could this be affecting it?

    We checked the differential outputs (BBI_P and BBI_N and they are two sinusoids out of phase and the same for BBQ_P and BBQ_N)

    The register settings in use are:

    Reg 1: 60100409
    Reg 3: 0000000B
    Reg 5: 0D03A28D
    Reg 6: 9090100E
    Reg 7: D041100F
    Reg 4: 4A01000C
    Reg 2: 8880620A

    The traces on the PCB are length matched. 

    Lets say they are not in perfect quadrature for some reason, we should still see some level of image suppression. At the moment they are within 1-2dB of each other. 

    Following is our schematic for the same: the series inductors are set to zero Ohms. 

    Regards,
    SM.

  • 0 in reply to SM
    TI__Guru 66014 points

    Hi SM,

    SM said:
    We also checked the dc levels on them at R62 and R54 with respect to ground. One of them was at 1.37V and the other at 2.11V. Could this be affecting it?

    This is quite extreme. Given the DAC to TRF is AC coupled, there must be some mis-configuration or soldering issues at the TRF372017 side. The DC offset adjustment capability on the TRF372017 is in the mV range. Therefore, I suspect this is a soldering issue.

    SM said:
    Lets say they are not in perfect quadrature for some reason, we should still see some level of image suppression. At the moment they are within 1-2dB of each other. 

    I agree.

    I have also asked our colleague to look into this issue as well to brainstorm.

    -Kang

  • 0 in reply to Kang Hsia
    TI__Mastermind 18065 points

    When you see both carriers at the same level then it means one of two things: 1) the BB signals are not in quadrature or 2) one channel's signal is not getting to the modulator input.  You are reporting that the signals are indeed in quadrature which implies there is signal on both I/Q; hence, one channel must not be getting through.  As mentioned previously, the DC imbalance is excessive and problematic.  The voltages on all legs (if the internal Vcm is selected) should be hovering around 1.7V.

    As an experiment, disable one of the DAC outputs and monitor the modulator output.  This can help to deduce which leg is not working. If for example you disable DACA and there is no change, then that channel's signal is not getting through.

    As a side note, the DAC loading has an inductor to ground.  This will set the DAC Vcm to 0V.  That is a bit strange configuration and will create a HPF response and I am not sure if the DAC will "like" being DC shorted.

    --RJH

  • 0 in reply to RJ Hopper
    Expert 1730 points

     @RJ Hopper

    The idea of switching OFF one of the DAC outputs to find the non-working channel was good. We turned OFF both the channels and found that the I port of the modulator is not working and this is the port which has the common mode voltage around 2.13V. We checked the soldering and also reheated the device but there is no difference. 

    What I mean by I port not working is that when a signal was fed on the I port and no signal on the Q port there was no output from the modulator, while vice versa (no signal on I and signal on Q produced an output). 

    Regarding the shunt inductor to ground on the DAC output side, we will get remove it (don't remember how we got it there)

    Thanks again for helping me out and looking forward to your inputs,

  • 0 in reply to SM
    Expert 1730 points

    Hello,

        Looking forward to your inputs,

    Thank you

  • 0 in reply to SM
    TI__Mastermind 18065 points

    Ok, so you determined the I port was not working.  That explains why the Vcm is off and why there was no sideband suppression.  Something still must not be configured properly in that path.  Sometimes it is difficult to debug.  Perhaps you can break the connection to the modulator and use a probe to see if the AC and DC signals are expected.  My experience in cases like these is that there is a wrong resistor value or bad solder joint that is the root of the problem.

    --RJH

  • 0 in reply to RJ Hopper
    Expert 1730 points

    Hello @RJH @Kang

    Last few days we have tried a lot of things:
    1) Replaced the chip

    2) Replaced all the components in the I and Q paths and removed the shunt inductor as you pointed out

    there is no difference at all. The I path still doesn't work. 

    We strongly suspect that there is something wrong with our register settings.

    The register settings in use are:

    Reg 1: 60100409
    Reg 3: 0000000B
    Reg 5: 0D03A28D
    Reg 6: 9090100E
    Reg 7: D041100F
    Reg 4: 4A01000C
    Reg 2: 8880620A

    Please have a look. Looking forward to fixing this issue with your help,

    Thank you for your time,

  • 0 in reply to SM
    TI__Mastermind 18065 points

    As I recall, you previously executed an experiment where you turned off the signal at DAC for the I channel and found that there was no change in the modulator output.  Similarly, if you turn off the signal at the DAC for the Q channel the entire modulator output goes away.  This indeed points to a bad I channel.

    I cannot think of a register setting issue that would disengage the one channel in the TRF3720.  So let's confirm a couple of things first.  Have the DAC output a low frequency signal, say 30 MHz.  Then with a probe, you should be able to walk down the I/Q traces up to the coupling cap, past it, and to the pin of the device.  In this experiment do you truly see a quadrature 30 MHz tone all the way to the modulator pins for each channel?

    Also, check the Vcm before and after the coupling cap.  Before the cap, you should see the common mode voltage that the DAC wants to operate at.  After the cap you should see 1.7V on all four BB inputs.

    --RJH