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.

How to provide IEC 61000-4 protection to isolated analog 4-20mA loop inputs

Other Parts Discussed in Thread: ISO1211, ISO1212, STRIKE

I am designing an embedded industrial controller that must have 11 separately isolated 4-20mA two-wire loop inputs.  The power for each loop and the isolated circuitry is supplied by its own isolated 24V DC-DC converter.  Line-to-line protection for the current-loop inputs is easy, but I don't see how to provide line-to-protective earth protection.  The accumulated charge of multiple EFT strikes during the testing will accumulate on the isolation capacitance of the 24V DC-DC converter until reaching its isolation capacitance breakdown voltage.  Further, the current that flows during each transient is in the neighbourhood of 25 amps, although only for some 15 nsec, but then it is about 15 amps for about 50 nsec.

So I have the following concerns:

1. How do I bleed the charge off the isolation capacitance without violating the isolation requirement?  I can use a bleeder resistor to protective earth, but that has to be small enough to discharge the capacitance in the time between successive EFT strikes during the IEC 61000-4-4 test.  With a maximum isolation capacitance of 120 pF, the resistor has to be less than about 15K to get to 5 time constants.  To provide tolerance to mis-wiring during installation for this 240 VAC system, the resistor would need to have a 15 watt rating.

2. Is the large current spike a problem for the isolation capacitance?

3. Is there a better way to do all this?

Thanks for your time,

Peter

  • Hey Peter,
    Thanks for your question. Our team will take a further look at this on Monday and give you a more detailed answer then. In the meantime I would have a look at the ISO1211 and the techical documents related to it (look at the technical documents tab of the ISO1211 product folder).
    Best regards,
    Dan
  • Peter,

    Are you referring to our ISO121x products? If so, have you considered using a TVS diode to help prevent charge build up? For the diagram, L1 can be a sense pin, L2 can be the FGND pin and "Ground" can be protective earth.

    Respectfully,

    Lucas Schulte

  • Helo Lucas,

    Thank you for your answer.  Much appreciated.  My concern is that the TVS diodes are essentially "de-isolating" the isolated circuitry.  I had three reasons for wanting the circuitry isolated, as follows:

    1.  to make the system largely immune to sensor wiring errors during installation of the system,

    2.  to make it immune to differing Protective Earth potentials, because some of our sensors will be located a considerable distance away and in parts of the facility that are powered by different distribution transformers,

    3.  to prevent ground loops.

    Doesn't placing TVS diodes between L1, L2 and protective earth violate 1 and 2 above, and even to some extent violate 3?

    I'm a newbie at this; please bear with me.

    Best regards,

    Peter

  • Peter,

    Are you looking at a particular digital isolator? For example if you were looking at using the ISO1212, then section 9.2.1.2.5 of the datasheet covers how EFT is handled in the typical application. Our ISO121x family of devices are meant for isolating digital signals in 4-20mA current loops.

    The TVS diodes appear as an open circuit during normal operation and shouldn't factor into ground loops.

    Respectfully,
    Lucas
  • Hello Lucas,

    My current loops are being used for analog sensors, not switches, so my data are analog not digital. I am sensing temperature and pressure, both to better than 0.4%, about 8 bits of resolution, and have added 2 bits for a noise guard, making 10 bits with an ENOB of about 8.3 bits. The ISO121x family doesn't work for me as a result.

    The design has a field-side A/D converter (actually a uP with a 12-bit ADC to provide programmability as well), the current loop and the A/D are powered with a 24 volt DC-DC converter, the latter through a linear regulator. I have included current limiters for both the loop and for the uP and associated circuitry. The output of the uP is SPI through a digital isolator to the main processor.

    My problem with using TVS diodes to PE is that it is possible that:

    1. the installer may inadvertently connect 24 volts to the signal input directly, so the TVS diodes would have to be rated to handle that;
    2. there may be a potential difference between the PE grounds at the sensor and at the circuit which, if the sensor gets shorted to PE at the sensor, would cause a problem; and
    3. in future, we may be using sensors powered with an external supply, which creates a horror show of possible problems with differing ground potentials, differing PE potentials, etc. This would be hard to trouble shoot.

    I may be asking for the impossible, but I have to ask.

    Best regards,

    Peter
  • Peter,

    Unfortunately our specialty on this forum is digital isolation but I will do my best to help you out and notify an expert for analog applications. TVS diodes are chosen so that their breakdown voltage is targeted above anything that happens during normal operation. So a TVS with 60V breakdown would not trigger when 24V is introduced.

    Respectfully,
    Lucas
  • Peter,

    Is your system similar to the diagram above? If so I recommend the ISO77xx family of devices for the green blocks as it has the best EFT performance out of our digital isolator portfolio. Also, What EFT pass criteria are you seeking such as Class A or B? Additionally the bleed resistor will not be required for the isolator.

    Respectfully,

    Lucas

  • Hello Lucas,

    Yes, the part of my system that is troubling me is the blue box in the upper right corner of your figure. In particular, the protection indicated by the red arrow and the isolation capacitance of the isolated power supply and the isolation barrier, indicated by the blue arrow.

    The problem is as follows:

    If I do not provide a path to protective earth in the input protection (red arrow), then transients will charge the isolation capacitance of the isolated power supply and of the isolation barrier to unacceptable levels because there is no path to discharge that capacitance (I am looking for EFT, ESD and Surge to level 4 class B or better). The combined capacitance has to be rated for that, and the capacitance for the isolated power module that I have chosen will not handle that level (the full strike potential).
    If I do provide a path to protective earth, I have violated the isolation. There is the potential for ground loops and issues related to differing PE potentials at both ends of the current loop.

    I don’t know how to resolve this. For approach 2 above, perhaps the problem with approach 1 above can be mitigated by using a high-value bleeder resistor. Your thoughts, please?

    Adding a transient current suppressor (TCS) or transient blocking unit (TBU) ahead of the TVS diodes to limit the transient current does not solve this because:
    the ones that I have found have only a 40V withstand voltage, and
    this does not solve charge accumulation, merely decreases it, making it take longer to reach a critical voltage across the isolation capacitance. During the IEC 61000-4-4 test, the DUT will receive 75 2kV pulses in 15 milliseconds (IEC 61000-4-4 level 4 for data lines) repeated every 300 milliseconds. With Ciso = 120pF, the voltage across Ciso rises at about 104 volts per pulse (Q = CV) which means that the voltage across it will exceed 2kV after a mere 19 pulses or about 3.8 milliseconds.

    Do you have any idea how to address this? I am at a loss at the moment.

    Best regards,

    Peter
  • Peter,

    The ISO77xx for your application will help meet EFT, ESD & Surge to level 4 Class B without any issues and possibly Class A. It doesn’t require any additional components like a bleed resistor.

    Respectfully,
    Lucas