• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 47 subscribers
  • Views 180 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

OPA2191: Cross regulation issue with OPA192 in 3-phase 400V measurement using shared neutral point

Rock Shao
Expert 2995 points
Part Number: OPA2191
Other Parts Discussed in Thread: OPA192, INA333

Tool/software:

Hi E2E,

Problem Description:

[Application]
We're using three OPA192-based differential amplifier circuits to measure phase-to-neutral voltages (L1-N, L2-N, L3-N) of a 3-phase 400V AC system. Each amplifier senses voltage via resistor dividers, with all three circuits sharing the same neutral point (N) as reference.

[Observation]
When enabling only one phase (e.g., L1 = 494V, others = 0V), the output is stable (e.g., ~7.8630V).
However, when enabling all three phases together, slight variations occur in each output (~0.0002V), even if the input voltages remain the same.

This issue is defined as poor cross regulation, where changes in one channel affect others.

Suspected Root Cause:

  1. Shared Neutral Point Drift:
    The "N" point is not a perfect ground. When all three sense paths share N as their negative input, any loading or imbalance can cause voltage drift, impacting all channels.

  2. OPA-based Differential Amplifier CMRR Limitation:
    The discrete differential configuration has limited common-mode rejection. It is more sensitive to neutral point fluctuations compared to an instrumentation amplifier. But the OPA219 CMRR 150dB max.

  3. High Impedance Design:
    We're using high-value resistors (e.g., R1 = 820k || 820k || 820k ≈ 273kΩ, R2 = 39kΩ) to reduce power, but this increases sensitivity to leakage, parasitics, and noise.

Current Design Snapshot:

  • Input: 3-phase 400VAC, each phase through voltage divider to OPA192 differential amplifier

  • All three amplifier circuits share the same N as reference

  • Gain ≈ 0.01585

  • Output range: ~7.8630V per phase (at ~494V input)

Request for Support:

We would appreciate any of the following:

  1. Confirmation on whether this cross-regulation issue is inherent to shared-N OPA-based structures

  2. Design suggestion using OPA192 to improve isolation/stability

  3. Recommendation on switching to INA (e.g., INA333) or other instrumentation amplifiers

  4. Any known TI reference designs for accurate 3-phase voltage sensing

Thanks, and regards

  • 0
    TI__Guru* 77481 points

    Hi Rock, 

    Rock Shao said:
    When enabling only one phase (e.g., L1 = 494V, others = 0V), the output is stable (e.g., ~7.8630V).

    This implies that the difference amplifier is attenuated at 7.8630/495.83 = 0.015858 V/V. or 39k/(3*820k) in attenuation. 

    Rock Shao said:
    However, when enabling all three phases together, slight variations occur in each output (~0.0002V), even if the input voltages remain the same.
    Rock Shao said:
    Confirmation on whether this cross-regulation issue is inherent to shared-N OPA-based structures

    Yes, this is a typical due to imbalance or mismatched line impedance. This voltage may get worst if the Neutral is not stiff and allowed it to float or shift due to return currents (unbalanced phase currents). Typically, the Neutral terminal is tied to Earth GND or at the main distribution panel. 

    Rock Shao said:
    Design suggestion using OPA192 to improve isolation/stability

    OPA192 has an excellent intrinsic CMRR. I do not know what resistor tolerances you are using in the difference amplifier, which it should be around 0.1% in tolerance or better. If you want to improve the system's CMRR at OPA192, you need to use matching resistors to improve CMRR. Even with a high performance OPA192, small cross-cross talk is still there. In addition, the 3-phase system is shared the GND paths that can inject common-mod noise into the measurement.

    OPA192 difference amplifiers should be stable. You may consider to limit the op amp's BW and reduce the op amp's total noise, since the measurement is occurred around 400Hz.  

     Difference Amplifiers—the need for well-matched resistors 

    Rock Shao said:
    Recommendation on switching to INA (e.g., INA333) or other instrumentation amplifiers

    You can try the following difference amplifier IC and see if there are improvements. 

    https://www.ti.com/amplifier-circuit/difference/overview.html

    Instrumentation amplifier topology (1 + Gains) won't help you since you want to attenuate the HV line by 1.58%. The difference amplifier is a better choice. However, 

    1/CMRR_total = 1/CMRR_opa192 + 1/CMRR_resistor_matching_network

    So if the resistor matching in the difference amplifier is poor, then the overall CMRR will suffer (even though you are using the high CMRR OPA192). 

    Rock Shao said:
    Any known TI reference designs for accurate 3-phase voltage sensing

    Please see the link below. 

    https://www.ti.com/tool/TIDA-01541

    https://www.ti.com/lit/ug/tidu181b/tidu181b.pdf?ts=1753415385198

    https://www.tij.co.jp/jp/lit/ug/tidueg8a/tidueg8a.pdf

    Please let us know if you have other questions. 

    Best,

    Raymond