• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 33 subscribers
  • Views 113 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.

TCAN4550: TCAN circuitry causes Radiated Emissions test failure

mike f
Prodigy 120 points

Part Number: TCAN4550

Hi, we have a product that uses TCAN4550 that is well-tested and everything functions as intended. 

We have just begun Radiated Emissions testing and we are failing because of excessive emissions in the harmonics of 40Mhz (120, 200, 280 etc). We have pinpointed the 40Mhz TCAN crystal as the source of the emissions. We hope that by making small changes to our board we can reduce these emissions.

  • Our original design is a little lacking in coupling capacitors for the TCAN chip. is it possible that adding capacitance could reduce noise in the system?
  • We are seeing that our CAN wires radiate this same noise. We chose not use a choke between CANH and CANL due to packaging constraints. Assuming we are able to reduce the noise from the crystal, can we get away without a choke here? If not, is there a very small choke you could recommend?
  • We chose to use OSC1 as an external clock input and we grounded OSC2. The datasheet shows a crystal oscillator connected to both OSC1 and OSC2. Can you outline the differences between these two approaches? 
  • Is it possible to use a "damping resistor" on the oscillator to reduce emissions while still allowing the TCAN to function properly?
  • 0
    TI__Mastermind 35630 points

    Hi Mike,

    These are good questions and I don't think I have solid straight-forward answers because each application is different so you may need to do some experiments and see if you can get any improvement.

    If you can add decoupling caps, I would think that would only help.  The crystal oscillator's current will vary during the oscillation cycle, so stabilizing the local supply (VIO) near the oscillator should help reduce any supply variance at the switching frequency. 

    Also because CAN bit timing is by definition an integer multiple of the clock frequency, there is doing to be some emissions from the CAN data bits.  Any asymmetry between CANH and CANL will result in a common-mode spur that is usually filtered out with with a common-mode choke.  

    If you have termination resistors on your board, using to series 60.4 ohm resistors with a 4.7nF cap to GND in between them will help filter common mode noise as well.

    Adding some capacitance to GND on each of the CANH and CANL traces will help slow down the rise/fall times and filter out some of the higher frequency content that may also help reduce emissions.

    For the single-ended clock input, a dampening resistor may help, but perhaps some capacitance as well for the same reasons of slowing the rise/fall time and filter high frequency content.  When using a single-ended input instead of a crystal, the clock voltage levels need to meet the digital VIH and VIL levels but this is usually a square wave type of clock signal that has higher frequency harmonics in the signal to create the square waveform.  A crystal is a pure 40MHz sine wave with an approximate 1Vpp amplitude so it may have less harmonic energy, but it is not as easy to implement and needs extra components.

    For CMC recommendations, we usually recommend either a 100uH or 51uH choke such as ACT45B-101-2P-TL003, or ACT45B-510-2P-TL003 as examples.  The TCAN4550 was not designed to meet worldwide EMC requirements without the use of a choke, so I can't confirm whether or not you will meet whatever standard you are testing to without a choke.

    Regards,

    Jonathan

  • 0 in reply to Jonathan Nerger
    Prodigy 120 points

    Thank you for all the insight! I'll update with our experimental results

  • 0 in reply to mike f
    TI__Mastermind 35630 points

    You're welcome. 

    Regards,

    Jonathan