• State TI Thinks Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 64 subscribers
  • Views 527 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.

OMAP3530 EMI Noise at 1.2GHz

JemiahAitch
Intellectual 290 points
Other Parts Discussed in Thread: TPS65950


I've got a devil of an EMI issue. This is on a Logic SOM-LV running the 3530 and the TWL4030/TPS65950 PMIC. Holding a 1.2GHz antenna right over the PMIC I get a beautiful spike on my spectrum analyzer, centered at 1.2GHz with a pair of smaller side band spikes at 1.148Ghz and 1.227Ghz. I've also got a Beehive 100B H-Field Probe, their smallest loop antenna and I can localize it to near the PMIC and the caps for PLL1, PLL2, TWL_RTC, VINTDIG_OUT and the RTC. This is what I've tried so far to get ride of it:
- I've tried every permutation of the spread spectrum functionality of the DSS, DPLL, PDPLL and USBHOST modules on the processor.
- I've turned off sys_clkout1 and sys_clkout2 which are unused.
- Tried adjusting the ramp rate of the VDD1 module.
- Tried changing the clock source for the DCDC module.
- Tried adding some additional capacitance to the 32KHz crystal supplying the PMIC.

Nothing I've tried has had any effect. There's a long list things I could post related to the issue and I'd be happy to post it if anyone has any ideas.

Thanks,
Jemiah
.

  • 0
    TI__Intellectual 2820 points

    Hi JemiahAitch,

    We need some time.  Please let us come back to you on this question.

    Regards,

    Lawrence

  • 0 in reply to Lawrence Ronk
    TI__Genius 15100 points

    Hi,

    I have never heard of such issue with the PMIC. The only switching on the PMIC would be the DCDC at 3.2MHz (assuming audio HF is not used). The noise you see at 1.2GHz is very strange an dthere doesnt seem to be any obvious source for this on the PMIC.

    Do you see this on multiple SOM boards?

    What is your application and is there any other circuit on the SOM/motherboard that can possibly do this? You would have checked this, however, i thought of confirming.

     

    Regards,

    Gandhar.

     

  • 0 in reply to Gandhar Dighe
    Intellectual 290 points

    Thanks for the reply Gandhar. I've continued to attempt find the source and I've actually identified two distinct patterns of 1.2Ghz radiation. I've attached some images of these. One is very sharp with peaks separated by 60MHz (Sharp.jpg). The other is much more disperse with peaks separated by about 25Mhz(Wide.jpg). These images were captured with an E-field probe that has a 0.2mm sensitivity. I'm not sure these signals are coming from the PMIC but they seem to be strongest near the PMIC. The problem is that these signals seem to be coupling onto nets on the board that are unrelated. I've found strong emissions of 1.2 GHz on the I2C buses, the 1.8V rail,  uP_GPIO_2 (JTAG_EMU1/GPIO_31 on the OMAP) and the 32kHz crystal.  I've tried detuning all these nets and am in the process of adding some small caps in strategic locations to absorb the emisions. So far, I've had no success in shifting or attenuating the signals. Just for background, the core issue is a violation of the DO-160 specification at 1.2GHz. The Sharp signal is present immediately after power up. The Wide signal doesnt show up until the OS (Linux) has started to loaded. I'm working my way through the board init code to see what it's fiddling with at the moment the signal starts.  My current suspects are: the oscillation source for the switching power supplies in the PMIC and the USB clocks of the PMIC. I've been sifting the TRM but have yet to find anyway to change the 3MHz clock frequency of the switchers. Is there anything you can tell me there?

    Thanks,
    Jemiah

    .
    ..

  • 0 in reply to JemiahAitch
    TI__Genius 15100 points

    Hi,

    The clock frequency for switchers will change when in low power state. If you configure this state using the SLEEP feature then the switching frequency will be approximately half.

     

    Regards,

    Gandhar.