Post Go back to editing

EMI reduction of inverting input OP AMP architecture

Hi all,

I'm using an inverting opamp architecture with an AD8534 to sum the ac currents from two physically orthogonal sensing inductors. The magnetic sensors are for low frequencies 200Hz to 5kHz and the source impedance into the OP AMP is low - a few hundred ohms. The circuit is suffering RF rectification in the few hundred MHz region. I can't use a simple capacitor across the non-inverting input since the amp goes unstable. I can't add much series resistance and a capacitor because the passband response will be affected and at low resistance values, the amp is still unstable. I've tried ferrite beads on the inputs but that hasn't worked. I can't completely screen the sensors and that would probably stop them doing their prime job anyway.

I've looked at the EMI hardening app notes MT-070 and MT-096 but those suggestions don't work on the current summing input. Can anyone help with a solution or suggest an EMI hardened amp that I can try? The circuit works at 4V. Thanks.

  • Hello, 

    I will check what options we have for you and get back to you soon. 


    Thanks, 

    Kris

  • Hi Gareth, 

    Sorry for not getting back to you sooner. 

    Have you tried doing some shielding on your setup? You could try to put the PCB in a metal enclosure, then connect the enclosure to system ground. Use something non-conductive to isolate the PCB from the box. There are potentially many ways that the EMI could be coupling into your circuit, and some shielding might help to understand whether the EMI is coupling through the pins or externally. 

    Could you also provide a picture of your setup? Are you working with a PCB?

    Thanks, 

    Kris

  • Dear Kris,

    Thank you for your reply. I haven't tried shielding yet because I expect that to stop the inductors from sensing the magnetic field (the prime function of the product). The circuit is on PCB and is fitted into an ABS case. The inductors and/or the connecting tracks are probably acting as antennas and resonate in the region of 280MHz under the 80% AM 3V/m EM field test. I have been able to make some improvement by placing RF decoupling capacitors on the voltage reference end of the inductors, adding a series resistor to the non-inverting input and changing the OPAMP to a TI product that claims to have better EMIRR performance. I hope the latter will not prevent you from continuing to discuss the issue with me! We're still using the AD8534 for other signal processing in the product.

    The issue is how to reduce the amount of unintended RF pickup from being rectified at the IC input(s). As I mentioned before, the signal is current input and low impedance so I can't introduce much series impedance and I can't capacitively decouple on the inverting input since the IC becomes unstable (high gain at high frequency).

    As you can see from the attached pictures, this board is a prototype. Due to commercial time pressure, we have had to send the board data for PCB manufacture but we've compacted the IC input tracks as much as possible and added RF decoupling. I will reassess the modified design when the first batch are built in a few weeks's time but I still expect to have EMI issues.

    Are you able to offer any more suggestions?

    Thanks and regards,

    Gareth Pont

Before You Switch


Switching languages will make ADI Explorer unavailable. Resume your session by switching back to English and reopening ADI Explorer.