Inquiry Regarding ADE9113 Isolation Ratings for 1000 VDC Measurement Applications

Hi,

The short answer is 'no' for 1000 VDC, neither basic nor reinforced. We would recommend using the non-isolated version in the same ADC family, ADE9103. Reinforced isolation could be accomplished by isolating the SPI interface with a device like the LTM2810-S. Alternatively, the MCU could be moved to the DC bus side. The systems energy output pulses isolated with an optocoupler, or perhaps isolation on the comms protocol between the metrology processor and the outside world.     

To explain the device limitations in more detail... An isolator is essentially two parallel structures. Each has its own failure mechanism we're looking to prevent. There is the internal signal isolation process, and there is the physical separation between the voltage domains enabled by the device package. We typically need to consider working voltage limitations and transient overvoltage limitations. 

The internal signal isolation process working voltage is rated under the IEC 60747-17 standard (VIORM and VIOWM quoted above). Lifetime testing is done to characterize those figures. The working voltage due to the package will be determined from the package creepage, the material group (both of those in the datasheet), the end application installation environment. Standards like IEC61010-1 and IEC62368-1 have tables to calculate working voltage limitations from package/environment characteristics.

Specs like VISO are ratings for a transient overvoltage. These are momentary stresses that cannot exceed devices ratings to avoid damaging the signal isolation process. Then there is package clearance and environment characteristics with look up tables in IEC61010-1 and IEC62368-1 for those limitations. 

Regards,

Jason

Thank you for your previous explanation regarding the ADE9113 isolation ratings.

Based on your response, I understand that the ADE9113 isolation barrier should not be used with a continuous potential difference of approximately 1000 VDC between the high-voltage measurement domain and the SELV domain.

To evaluate whether the device may still be suitable for lower-voltage applications, I would like to clarify the maximum allowable continuous DC working voltage.

Could you please advise the following?

1. What is the maximum continuous DC voltage that may be applied across the ADE9113 isolation barrier while maintaining compliance with its safety certifications and lifetime expectations?

2. How should the published ratings below be interpreted in terms of equivalent continuous DC working voltage?
- VIOWM = 405 Vrms
- VIORM = 573 Vpeak
- Reinforced insulation at 415 Vrms
- Basic insulation at 830 Vrms

3. For example, would the following continuous DC voltages be considered acceptable across the isolation barrier?
400 VDC/500 VDC/600 VDC/800 VDC

4. Is there a recommended maximum DC working voltage for:
- Basic insulation
- Reinforced insulation

5. Is there an application note, certification report, or engineering guideline that explains how the AC isolation ratings should be interpreted for DC applications?

My objective is to determine the maximum DC bus voltage for which the ADE9113 can be used as the isolation barrier between a high-voltage measurement domain and a SELV control domain.

Thank you for your assistance.

Hi KNH,

Unfortunately there is not a universal limit. Applications are typically bound by the applicable standard for its specific space and there are some differences between standards. There are also differences based on where it is installed. The pollution degree and the overvoltage category of the installation are examples. 

Regarding package limitations... The ADE9113 is material group II and has 8.3mm creepage. Let us assume installations are pollution degree II and <2000m. I'll use tables in IEC 60664-1 because those are often the source for the system standards. The system standards typically equate VDC to VRMS. Interpolating between the rows in IEC 60664-1 Table F4 for ADE9113 specifics, I come up with 579V for reinforced, 1158V for basic. Those numbers would be lower for pollution degree III or higher altitudes. 

The signal isolation process is interesting for DC. IEC 60747-17 is the specific standard for thin film isolators. That standard has a methodology for determining the working voltage at 50/60Hz which are the VIORM and VIOWM figures in the datasheet. We have done characterization of the ADE9113, stressing the barrier with DC using the same methodology defined for determining 50/60Hz lifetime. The characterized DC working voltages are much higher than those reported in the datasheet for AC.

There is then the practical issue of working voltage certifications. As far as I am aware, IEC 60747-17 doesn't specifically cover DC. The VIORM (which is a reinforced value) can be substituted for DC there. The working voltage limits of the barrier between basic and reinforced don't double like they do with the package. The difference is significant enough for basic certifications, so the reinforced VIORM figure would be used in a basic insulation application.   

Long winded answer for 400V/500V, either basic or reinforced would be achievable. Working voltages beyond that, use the ADE9103 and a discrete isolator. Series isolators for those higher voltages might also be a possibility from a safety perspective but be mindful that you are adding to the delay in comms. That might effect the maximum SPI rate or how many parts can be daisy-chained before additional timing complications are created.

Regards,

Jason

Thank you very much for the detailed explanation.

Your explanation regarding the distinction between package limitations (creepage and clearance) and the internal signal isolation process was extremely helpful. I also appreciate your clarification regarding the use of IEC 60664-1 Table F.4 and the reasoning behind the estimated working voltages for the ADE9113 package.

I have a few follow-up questions.

The system I am developing is evaluated in accordance with IEC 60664-1.

In addition, the PCB will be coated with a conformal coating that is intended to meet the requirements of Type 1 protection as defined in IEC 60664-3.

Therefore, I believe the insulation coordination may be evaluated using Pollution Degree 1 rather than Pollution Degree 2.

Under the following assumptions:

- ADE9113
- Material Group II
- Creepage distance = 8.3 mm
- Pollution Degree 1 (according to IEC 60664-3 Type 1 protection)
- Altitude below 2000 m

My interpretation of IEC 60664-1 Table F.4 is that the package creepage distance may support approximately:

2160 V for basic insulation
1080 V for reinforced insulation

Could you please comment on whether this interpretation is correct?

In other words, if the equipment is evaluated as Pollution Degree 1, would the package creepage limitation of the ADE9113 still be a concern for a working voltage around 1000 VDC,

or would the package-related limitations be considered acceptable in this case?

Regarding the internal signal isolation process, you mentioned that:

"The characterized DC working voltages are much higher than those reported in the datasheet for AC."

This is very interesting.

Would it be possible to share any characterized DC working-voltage data, lifetime data, application note, or engineering guidance related to the ADE9113 isolation barrier under DC stress?

If the detailed data cannot be shared, could you at least indicate the maximum continuous DC working voltage that Analog Devices has characterized internally for the ADE9113 isolation barrier?

I am trying to understand whether the remaining limitation is primarily the package insulation coordination requirements or the long-term reliability of the internal isolation barrier itself.

Thank you again for your support and detailed explanations.