• TI Thinks Resolved

ISOW7841: High EMC emissions at 64, 207 & 316MHz

Prodigy 20 points

Replies: 4

Views: 100

Part Number: ISOW7841

We are currently using the ISOW7481 & ISO7481 to isolate an ADC circuit with the ADS1298, and we discovered that the EMC emissions of the Isolators are too high in the frequencies 64MHz, 207MHz & 316MHz.

This is unexpected as the chip has Robust electromagnetic compatibility (EMC) and low emissions and the design and layout guidelines have been followed. 
Please find below copies of the ISO configuration and spectrum analyzer graphs.

Any ideas on what could be causing this?

Thanks in advance

  • Hi Daniel,

    Thank you for posting, and welcome to E2E! We appreciate the schematic and measurements you've shared in this post.

    Some emissions are expected from isolated systems, especially when power is isolated as with ISOW7841, however this can be minimized following low-emissions design and layout guidelines, as you mentioned. Below is an emissions spectrum with expected emissions levels at Vcc = 5V and Vcc = 3.3V for ISOW7841:

    Guidelines we recommend for low emissions can be found in our Application Report, Low-Emission Designs With ISOW7841 Integrated Signal and Power Isolator (SLLA386C). Is it possible for you to share PCB layout images for verification? If desired, these can be shared via Personal Messages or email instead of on this public forum. We can analyze and share additional test setup and system-level recommendations to reduce emissions caused by ISOW7841.

    Please also let us know which emissions standards, if any, this system is intended to comply with.

    Thank you!
    Manuel Chavez

  • In reply to Manuel Chavez:

    Hi Daniel,

    Manuel has already shared you a link to an application note that includes many suggestions to help reduce radiated emissions. He has also shared you emissions test results expected on an EVM with clean test setup. Unfortunately, emissions are largely dependent on PCB size, connected cables and the actual test setup and hence, it is possible to see higher emissions than what is observed on TI EVM on a clean test setup.

    I see that you are seeing emissions of about 54dB at about 60MHz and other lower failures at other frequencies. Based on whether you need to meet CISPR 32 Class a or Class B, the amount of failure can be different.
    I am adding below a few specific suggestions from the application note that include device part numbers to make it easy for you to improve emissions on your board.

    To reduce emissions around 60MHz, 180MHz & 320MHz (harmonics), we recommend you to use common-mode chokes (CMC) at both input and output power supply pins of ISOW as shown in the image below.

    The ferrite beads on I/O lines are optional, you can also choose to use >1kΩ resistors instead if the datarate is not very high. Please make sure all the I/O signals are referred to the GND points after the CMC and not to ISOW device GND pin. ISOW device GND & GNDISO pins should stay local to device and all the signal I/Os should only refer to the GND point that is after the CMC.

    The CMCs should have an impedance of >1kΩ at 60MHz, I am listing below some of the example CMC part numbers that we recommend.

    1. CMC - 200mA, 15kΩ - TDK - ACT1210-510-2P

    2. CMC - 200mA, 20kΩ - TDK - ACT45B-510-2P

    3. CMC - 200mA, 10kΩ - Wurth - 744235510

    Let us know if you have any questions, thank you.

    Koteshwar Rao

  • Hi All,

    Thank you, Manuel and Koteshwar for your answers and recommendations. 
    We are going to implement the changes in the schematic and layout. Hopefully, the Emissions will be reduced.

    Kind regards

  • In reply to Daniel Ventura:

    You're welcome, Daniel!

    Please let us know if any additional assistance is needed.

    Manuel Chavez