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.

ISO7221C: Stacking Multiple Isolation Boundaries to achieve higher working voltages

Part Number: ISO7221C
Other Parts Discussed in Thread: ISO7721, ISO7821

We have several customers in the instrumentation and medical space that require isolation to relatively high WORKING voltages, such a >5 kV.

One approach to this has been to "stack" two isolation boundaries such that there is an intermediate region between them.  The thought being that the total working voltage is the sum of the working voltages of the isolation boundaries on either side of the intermediate region.

I see a problem with this- the common mode voltage potential of the intermediate region doesn't seem to be defined- in order to avoid stressing either boundary, it seems like the intermediate region needs to be actively driven (defined) to a common voltage potential half-way between the adjacent regions.  Due to board layout or EMI issues, without some circuit to set the potential of the intermediate region, one of the isolation boundary could get over-stressed, fail short, and then cause a cascading failure of the secondary isolation boundary and a complete isolation failure.

What is the right way to stack isolation layers to achieve working voltages above that achieved with a single layer?  How do we drive the intermediate region properly to avoid stress?  What about CMTI in this situation?

Please point us to whatever useful resources you might have on this matter; we have several customers asking this and this issue comes up alot.

  • Hi Antenna_Head,

    Using multiple digital isolators in series is possible to achieve higher isolation ratings, however as you mentioned, determining how voltage stresses will continually divide between the isolators requires a known intermediate voltage.

    Using two or more devices in series without an intermediate voltage has yielded about even distribution of stresses in our experiments.

    Which working and isolation voltage levels are our customers aiming to achieve?


    Please note that ISO7821 and ISO7721 are recommended over ISO7721C.


    Thank you for posting to E2E,
    Manuel Chavez

  • When you bridge the isolation boundaries with equal-sized resistors, you get a voltage divider.

    You probably need multiple resistors in series to get the required voltage rating, and the question is how much current is allowed to flow.

  • Agreed; the amount of common-mode current that is allowed would have to be a spec.

    The problem with a straight voltage divider is it creates a time constant against the stray capacitance across the boundaries.  Probably a capacitor in parallel with each resistor will contribute a zero to cancel the this low-pass pole.

    Clearly there seems to be some devils in the details.

  • Hi Antenna_Head, Clemens,

    A critical spec for even voltage division when using isolators in series is equal impedance of each part. If Cio values are different, voltage splits are expected to be proportionally different as well. External components could hinder this due to component tolerances and parasitics along with reduced isolation creepage and clearances.

    In our tests even oscilloscope probes across one or both of the isolators affects voltage division.


    Best,
    Manuel Chavez

  • Hi Manuel, (and others for the good conversation)

    TI has digital isolators with a single series cap for isolation and others with two series capacitors for the isolation barrier. Is the dual series cap what is allowing us to claim reinforced isolation? Is that also related to being able to claim higher working voltages and peak isolation voltages? Or are the higher working voltage just due to the capacitor(s) being higher voltage caps regardless of there being one or two?

    Thanks,

    John

  • Hi John,

    Thanks for joining the conversation! Using isolation capacitors in series increases our isolators' isolation voltage ratings, including working voltages, similar to how using resistors in series increases total resistance.

    Basic and reinforced (VDE) or single protection and double protection (UL) isolator ratings are given by certification agencies based on isolator performance at certain voltage levels for isolation specs. What each certification qualifies for reinforced or double protection varies by agency and standard, so it is our device's performance to certain standards allowing us to claim reinforced isolation as noted in the Safety-Related Certifications section of isolator datasheets. More information on significant certifications for digital isolators can be found on this blog.

    Using isolation capacitors in series is not the only way to increase isolation performance, and the technology within our isolation devices and capacitors has improved, but since datasheet ratings come from device-specific certifications, a series-capacitor design can still be a "basic isolator" like the ISO77xxB.


    Please let me know if there are additional questions.


    Thank you,
    Manuel Chavez

  • Hi John, all,

    To complete a point mentioned in my final paragraph above, both improving the voltage capability of each isolation capacitor and using series isolation capacitor topologies have helped TI's isolators achieve the high isolation voltage levels they are rated and certified for today.


    Thank you,
    Manuel Chavez