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.

OPA387: Current Sense Circuit Input Side Low Pass Filter Design

Part Number: OPA387
Other Parts Discussed in Thread: TINA-TI

Hello,

I am using OPA387DBVR for my current sense design and there is a lot of noise on the input and output side. I saw there is a internal low pass filter but its frequency is too high. I need to filter above 1 kHz because my input is DC.

I am using a 100 mOhm shunt resistor and STM32F03 microcontroller ADC for measuring.

My current design is:

Is there any documentation about OPA387 filtering or can you suggest me a low pass filter circuit design?

  • Hi Metin,

    The total integrated noise for your application is about 29uVrms - see below.

    If you want to improve it, you need to add a low-pass filter at the output - see below (8.3uVrms).

    If you see any higher noise, it is probably because you operate Vout outside of OPA387 linear range of 0.1V < Vout < 3.2V (see below AOL test conditions).

    This means that for OPA387 output to be within its linear range, its input must be between 2mV and 56mv (see below). 

    BTW, since you reference the output to ground, you may not need to use difference amplifier unless you worry about the trace resistance - see below.

    Below I have attached Tina-TI schematic for your convenience.  You may download a free version of Tina by clicking on following link: https://www.ti.com/tool/TINA-TI

    Metin OPA387 Noise.TSC

  • Hi Marek,

    Thank you for your detailed explanation.

    I've had a few things on my mind. First of all, there will be situations where the distance between theöt resistor and the op-amp is long. So the resistance of the line will be high. Should I continue to use a difference amplifier in this case?

    Also, since the distance between the shunt resistor and the op amp is long and the system is in a noisy environment, should I add an input filter?

    Finally, should I remove the 22nF capacitors parallel to the 56kOhm resistors? In the simulation I got a lower total noise graph output with them connected to the circuit.

    Best Regards,

    Metin Enes Yıldırım

  • Please see my answers below:

    I've had a few things on my mind. First of all, there will be situations where the distance between the resistor and the op-amp is long. So the resistance of the line will be high. Should I continue to use a difference amplifier in this case?  Yes, if you worry about the voltage drop across the trace resistance, you should continue to use a difference amplifier.

    Also, since the distance between the shunt resistor and the op amp is long and the system is in a noisy environment, should I add an input filter?  You may - for most parts it gets filtered out at the output.

    Finally, should I remove the 22nF capacitors parallel to the 56kOhm resistors? In the simulation I got a lower total noise graph output with them connected to the circuit. Yes, there is a slight improvement in the noise by keeping 22nF feedback caps (5.88uV vs 7.75uV) so it's up to you to decide.

    But you will get even lower noise by increasing the output cap - see below.

    All in all, the range of your circuit will allow Ishunt measurement between 20mA and 570mA - see below.

  • Hi Marek,

    Thank you for your detailed explanation. Can you suggest me an input filter for this design? Am I need to use a differantial filter?

  • Metin,

    1. Marek is on travel, so I will help in his absence.
    2. I attach a design for a differential input filter on the diff-amp Marek suggested.  Normally, a differential filter is used when you can pick up a common mode noise (i.e. noise common to both inputs).  In your case, you are using a differential amplifier, but the non-inverting half is near ground potential.  So, a differential filter may not be needed.  If you think there is a potential pickup of a common mode noise you should use the differential filter.  If not, you can use a filter on the inverting channel only.  See slide 2.
    3. Here is a link to a document on filtering differential amplifiers (https://www.ti.com/lit/an/sboa016/sboa016.pdf ).  They are a little tricky to filter as compared to instrumentation amplifiers as the input impedance is set by resistors.
      1. One key point in the article is that capacitor tolerance can convert common mode noise to differential noise.  The article suggests using a 500x larger differential capacitor than the common mode.  Other documents recommend a Cdif = Ccm x 10.  The main point is that tolerance in the capacitor values will cause a different cutoff frequency on the inverting vs non-inverting input.  This translates a common mode noise to a differential noise.  Differential signals will be amplified.  Choosing Cdif = Ccm x 500 (or x10) sets the differential cutoff much lower than the common mode so that any translation of common mode to differential will be attenuated by the common mode filter.
      2. I show the 10x and 500x filter in my design summary.  The app note was written 30 years ago, so I think capacitor tolerance has improved.  The disadvantage of making Cdif = 500 x Ccm is that the differential noise cutoff is now at a higher frequency so less noise filtering happens.  The advantage of  making Cdif = 500 x Ccm is that it is more tolerant of capacitor component tolerance. See slides 11 and 12.
    4. Below is a power point that summarizes the results.  Also below are the TINA source files.

    opa387-filter.zip

    Diff-amp-filtering.pdf

    I hope this helps.  It's a lot of information.  If you want the simple approach you could just keep the simple filter.

    Best regards, Art