• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 48 subscribers
  • Views 274 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

INA253: 40MHz frequency peak at output

JosefMieslinger
Intellectual 2170 points
Part Number: INA253
Other Parts Discussed in Thread: INA240

Dear TI Amps Team,

my customer wants to use the INA253 for in-phase current sense in a solenoid application.

They did some testing and found that there is some 40MHz frequency peak at the output of the INA253 despite there is no input voltage (IN+ shorted with IN-, no PWM).

Can you please explain where the 40MHz could come from and how to surpress this?

Thanks and best regards,

Joe

  • 0
    Guru 140200 points

    Hi Joe,

    the INA253 contains a zero-drift chopped amplifier. This could make some noise. I would recommend to add a low pass filter to the output.

    Kai

  • 0
    TI__Mastermind 32225 points

    Hey Joe,

    So is the customer measuring the output noise voltage spectrum and are they seeing some peaking of energy at 40kHz?

    What are the common-mode (VCM) and power supply (Vs) voltages being used? Could the VCM or Vs be noisy?

    Even if VCM and Vs are perfect DC signals, there could still be a noise peak at some frequency. This frequency will relate to some internal clock and switching sources because this device contains a chopper amplifier as Kai answered. Chopper amplifiers actively chop the input offset voltage and thus require clocking.

    Figures 20 to 22 show peaks in input-referred noise spectrum, but they in the 100's MHz region.

    To suppress this you can use some RC low-pass filter on the output or an active-filter amplifier. If it is coming from the sources of VCM and Vs, then add an input filter at inputs pins or an RC/LC filter at the Vs pin.

    Sincerely,

    Peter

  • 0 in reply to Peter Iliya
    TI__Intellectual 2170 points

    Hi Peter,

    thanks for the quick feedback.

    The customer sees a frequency peak at 40MHz, unfortunately this is not covered by the figures 20 to 22 in the DS.

    Can you please confirm the switching frequency of the internal oscillator of the chopper amp?

    Do you have some figures about the amplitude of the switching noise of the chopper?

    The customer can measure this with oscilloscope several 10mV at the output.

    Thanks and best regards,

    Joe

  • 0 in reply to JosefMieslinger
    TI__Mastermind 32225 points

    Hey Joe,

    My apologies, I was referring to the INA253 datasheet which uses the same core as INA240.

    You can use figure 18 of INA240 datasheet to see a rise in the noise density around 200 kHz, which is the chopping frequency. There is an internal clock that will operate at MHz level, but this should not feed through the output at any significant level.

    We do no have data showing the amplitude of switching noise. Although it is possible to zoom into the signal and possibly see some noise at a specific frequency, although it should not dominate the overall noise of the output signal.

    Can you please comment on the set up and device configuration when the switching noise is being measured?

    Is the noise 10mVpp? Over what time scale is the measured?

    What is the output voltage during the test? Is the reference voltage be generated or is there an input differential sense voltage being applied?

    Are the VCM and Vs supplies noisy?

    Making noise measurements is always a delicate process. To measure the true noise of INA240, the VCM and Vs sources must all be heavily decoupled with plenty of capacitors (100nF + 1uF or 10uF caps). The output voltage should be offset into the linear region, which can be easily accomplished with one REF pin connected to VS and the other to GND to bias output to Vs/2. Next, the output should be measured differentially (Vout-VREF) with low noise probes that are as short as possible. Then placing everything inside a Guassian chamber will also help reduce EMI pick up. The chamber can be tied to GND and GND plane must also be stable 

    Sincerely,

    Peter