• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 1 answer
  • Subscribers 47 subscribers
  • Views 769 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.

INA281: Propagation

Jacob Hays
Prodigy 30 points
Part Number: INA281
Other Parts Discussed in Thread: INA240

Hello,

I am working on a board that utilize high side current sensing to record motor current.  I have been observing some significant delay between my MOSFETs turning on and the current sense amplifier output responding.  Along with this when the output first moves it overshoots the expected output voltage by quite a bit.

I have attached some captures that illustrate this behavior.  These captures were done with a simple 50ohm resistor as the load rather than a motor.  In this configuration I have a 1mohm sense resistor with approximately 1A of current, the expectation would be 50mV output as I am using the 50V/V version of the INA281.

I have read about overload recovery, but I am not certain that is what's happening.





  • +1
    TI__Mastermind 33160 points

    Hello Jacob,

    Besides the overload recovery you could have issues with CM change as well as bandwidth change. 

    To determine there is a overload recovery you could add an offset on the input voltage if you have a resistor in series with one of the inputs.  This will create an offset at the input of about 20µA*Resistance used.  I would suggest 20mV/50=400µV at input. which is Resistor of 20Ω on one input.  This will create an offset and a small gain error but create a positive offset to determine if it is an overload recovery issue.

    Also be aware the bandwidth is lower at lower input voltages see below.

    VCM changes at the input could couple to the output.  The CMRR at high frequency shows this but the INA240 works better for large changes in VCM.

  • 0 in reply to Javier Contreras4
    Prodigy 30 points

    I am assuming I should add this resistor on the IN- input?  I am beginning to think it could be a bandwidth issue as well, as I did some more testing I noticed as the current increased (and thus the output voltage increased) the delay between the FET turning on and the output voltage shrank, although even at a 500mV output the delay was about 1.5us.

  • 0 in reply to Jacob Hays
    TI__Mastermind 33160 points

    Jacob,

    I think it would work for either but I would place it on IN- as the current decreases with Vsense.  This plot is seen on the datasheet..  

  • 0 in reply to Javier Contreras4
    Prodigy 30 points

    Placed a 20ohm in series and the delay is still present.  I'm not sure that VCM is changing that much as one leg is hard tied to supply voltage (48V) and then the other is placed on the other side of the sense resistor, when the FET turns on we will see 1A flowing, so IN- should be approximately 47.999, so IN- will simply be toggling between 48V and 47.999V.

    Also can you explain how the bandwidth limitations would cause a delay like the above?  I would expect a difference\distortion in rise time but there should be no delay, but I could be mistaken.

  • 0 in reply to Jacob Hays
    TI__Mastermind 33160 points

    Hello Jacob,

    What voltage do you have at the output with input differential at zero.  I would hope it is above 20mV after adding the resistor to show it is above the low saturation point.  I had previously got a step response from -2mV input voltage to 40mV step at the input and I did not see a large delay.  See scope shot below.  This is even with a -2mV input.  Is there something else in your circuit that could be charging during this transition causing the current either not to go through Rshunt such as a capacitor. on the IN+ side.

  • 0 in reply to Javier Contreras4
    Prodigy 30 points

    I will check on the output voltage, I do not see anywhere else the current could be heading.  See attached, IN+ and IN- are tied across a 1mohm resistor on another page, can provide if needed.

  • 0 in reply to Javier Contreras4
    Prodigy 30 points

    I have done some further digging, the bias resistor did seem to resolve the issue.  Although I had to bump up the resistance up a bit when using an inductive load (motor), I am imagining with the 20ohm in there the inductive spikes were still causing it to saturate and/or overload.  I may move to another part that is bidirectional in nature (any suggestions would be great) or may add the resistance and deal with the offset on the current design as I already have procured for this build.

    Thanks so much!

  • 0 in reply to Jacob Hays
    TI__Mastermind 33160 points

    Jacob,

    We have the INA240 which can handle transients better and is bidirectional but is not as fast but the bandwidth is good over the linear range.  The resistor is a band-aid that causes offset and gain error unless you can calibrate those out.  Also the IB at 0V input can vary from 10µA to 30µA for the INA281 making the resistor value need to be designed for the 10µA condition to cover the variation of IB.