3-wire 3-phase (without neutral) energy meter design using ADE9112 /ADE9113

Hi, 

The ADE9112/ADE9113 are simultaneously sampling ADCs. They produce the 24-bit ADC samples used for metrology algorithms. It would be the job of the MCU to produce RMS voltage, RMS current, active/reactive/apparent power, and energy per phase, anything else needed from the ADC samples. 

You need to have one ADE9112/ADE9113 for each phase. Each ADE9112/ADE9113 has a high gain ADC channel for the shunt measurement, and then 1 or 2 unity gain channel(s) for the voltage measurement. Sounds like the ADE9112 has enough ADC channels for your use case. If a 2nd voltage measurement was needed, the ADE9113 is available. The phase-phase voltage is needed, so instead of setting the voltage divider from phase-neutral like it is in the ADE9112/ADE9113 datasheet examples, its going between two phases.   

The EVAL-ADE9113 can be used. Check out "Figure 8. Typical Setup for the EVAL-ADE9113EBZ for 3-Phase, 3-Wire Delta Distribution Systems" for an example of phase-phase voltage measurement.

https://www.analog.com/en/resources/evaluation-hardware-and-software/evaluation-boards-kits/EVAL-ADE9113.html

Regards,

Jason

Hi Jason,

Thanks for your quick reply.

As per our understanding, we plan to use one ADE9113 or ADE9112 per phase for our 3-phase, 3-wire system. Each device will measure its respective phase current (via shunt) and phase-to-phase voltage.

Our proposed approach:

• IC-A: IA + VAB (VA–VB)

• IC-B: IB + VBC (VB–VC)

• IC-C: IC + VCA (VC–VA)

However, in the EVAL-ADE9113 document (Figure 8), only Phase A and Phase C voltages are shown.

Could you please confirm if this mapping is correct, and how Phase B voltage should be measured or referenced in this configuration?

Thanks.

--Alpesh

You need an ADE9112/ADE9113 for each shunt measurement. Note the ADC's AGND is essentially fixed to one side of the shunt, which is why you need an independent isolated ADC for each shunt. The voltage measurements use a voltage divider to create a small signal close to the shunt's voltage potential to keep within the process capabilities of the chip. 

The example in Figure 8 would have the MCU solving for IC instead of directly measuring it. If you need to measure it directly for the application, add another ADE911x like above. 

jnaso said:

You need an ADE9112/ADE9113 for each shunt measurement. Note the ADC's AGND is essentially fixed to one side of the shunt, which is why you need an independent isolated ADC for each shunt. The voltage measurements use a voltage divider to create a small signal close to the shunt's voltage potential to keep within the process capabilities of the chip. 

Hi Jason,

Thanks for your quick support.

Could you please elaborate further on how the phase-to-phase voltage divider should be connected in a 3-phase, 3-wire configuration?
I’m a bit confused about the AGND reference, because when adding the voltage divider between phases, it seems that each divider could make the other phase act as AGND. like /resized-image/__size/640x480/__key/communityserver-discussions-components-files/357/pastedimage1759857941840v1.png_2D00_494x472.png

Could you please explain the correct connection method for the divider and how AGND should be handled in this case?

Thanks

-Alpesh

Sorry, the pic is coming up blurry. I can't see the details well, but it looks like anti-aliasing filters are missing from the ADC input pins.

The ADC's AGND pins will be connected to one side of the shunt, typically through a ferrite. The isolated ADC side of the ADE9112/3 is going to be at phase voltage. It might sound funny, but you want that pseudo-differential connection to the shunt.

You then create a voltage divider between that phase and another phase. You will be measuring phase-phase voltage. Typically going to be measuring across a 1kΩ locally and have a 1MΩ connecting to the other phase (assuming ~240V here). That gives you a voltage waveform scaled for the ADC's dynamic range at a potential close to the local phase voltage. 

On the eval board page. Go to "Evaluation Design File", download the zip and have a look at the 02-074550-01-b.pdf file there. That is the schematics for the eval board with the ADE9113. Its an example of the voltage divider and  

EVAL-ADE9113

Hi Jason,

Thanks for your detailed explanation and guidance.

PDF
We’ve now prepared our schematic as per your recommendations — each phase uses one ADE9112, with the AGND connected to the low side of the shunt (through a ferrite) and the voltage divider implemented between the respective phase lines for phase-to-phase measurement.

Could you please take a quick look at the attached schematic and confirm that our approach aligns with your suggested configuration? Also share if any other thing is missing from our end.

We just want to ensure the divider referencing and AGND connections are implemented correctly before finalizing the layout.

Thanks again for your excellent support.

Thanks,
--Alpesh

Presumably you want all the ADE9112 sampling simultaneously. They will need a shared clock source to do that.

The ADE9112 also has a daisychain SPI port so they can be serviced by the MCU as a single device. Then you don't need the multiple CSb and DREADYb pins attached to MCU resources also. 

Hi Jason,

We plan to use the SPI daisy-chain connection with all ADE9112 devices.

Could you please confirm if you have any additional comments or recommendations regarding the phase-to-phase voltage divider implementation and AGND connections for each device?

We’ve implemented the dividers and AGND connections as per your earlier guidance

Thanks again for your continued support and guidance.

Thanks.

--Alpesh

jnaso 

Is there any feedback on this ?

Thanks.

--Alpesh

 jnaso 

Is there any feedback on this ?

Thanks.

--Alpesh