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.

INA240: Current Measurement is not accurate with unwanted resonance at the output

Part Number: INA240

Dear All,

I am building a 800W pushpull converter operating at 40kHz switching frequency with primary current going till max 100A. Converter input voltage on primary side can vary from 30-42V.

I want to implement peak current mode control for this converter. So, I used a shunt resistor of 1mOhm in the high side to sense the primary switching current. In order to amplify the sense voltage I used INA240A1 (20V gain). But the IC output is not replicating the primary current waveform. Please refer to the following figures for the type of errors I am getting at the output of INA240A1 IC.

In all the following figures, green waveform represent IC output and yellow waveform represent actual current.

Shunt Gain = 0.002 and INA240A1 gain = 20. Total gain = 20*0.002 = 0.04

 

Figure 1. Error in IC output - With current measurements @ cursor value = 28.8A

Figure 2. Error in IC output - With IC output measurements @ cursor value = 1.12V

Figure 3.  IC output is deviating from its corresponding current waveform with peaky middle parts

Figure 4.  IC output is deviating from its corresponding current waveform with peaky middle parts and slow falling at the edges.

 

 


Figure 3 & 4 are of big concern as the IC output sometimes not following the current waveform making it non suitable for current control. I didn't understand why the IC output is no falling fastly as the original current waveform and following that it is giving peaky values in the middle. Any suggestions/help in this regard would be greatly appreciated.

Thanks in advance!

Kashyap Gundlapalli 

  • Hi Kashyap,

    The behavior you're seeing is normal.
    There will always be a delay in the part tracking the input, which can result in both a phase shift or delayed output and trying to act upon such signal will cause loop instabilities. 

    Would you be able to tell me more regarding your application requirements?

    Would average current control be an option for you? If not, what's the current PWM controller in use?

  • Hi Carlos,

    Thank you for the reply!

    Currently, no current control is implemented. It is running with DC link voltage control and the loop is operating at quite low frequency i.e 80Hz.

    Average current control is an option but I am also implementing hardware short circuit protection based on current sense output to protect pushpull devices and the response time of this protection circuit is 2.6us. My only concern with average current mode control is that the current sense should be fast enough to track the short circuit on primary side.

    Can you please tell me which parameters in the datasheet that you are referring to when you mentioned that it is IC's normal behavior? I thought the slew rate could be an issue at falling edges of actual current (figures 3 & 4 - blue circled areas of figure 4 as shown below) but I couldn't relate any datasheet parameter to the red circled area of Figure 4 as shown below:

    Kindly, suggest the possible reason for this phenomena and how to avoid it. 

  • Hi Kashyap,

    Looking at the oscilloscope plots, it seems that the yellow trace, just before the red circled area, is a switching event and there is a common mode change. The datasheet specifies that such event may take up to 9.6uS to settle.

    It appears the scale the oscilloscope is set to is 4uS/DIV, if so, the event circled in red is within the specified limits. Nevertheless I would like to clarify on a few more things before I can make any suggestions:

    • Oscilloscope scales (TIME/DIV)
    • Plot showing the bus voltage and the scale used
    • Is there a common mode event and, if so, what is the amplitude? Can this be plotted?
    • Can you provide a diagram showing where the current measurement is taking place and how your switches/load are organized?

  • Hi Carlos,

    Yes it is a switching event. As far as I understand it cannot be common mode change as the IC only see the differential voltage (appearing across it's input terminals) changing from finite value to zero in that instant. Please refer to the below circuit diagram of our system. 

    In the datasheet part you referred to, bandwidth is given as 400kHz where as settling time is 9.6us. Aren't these both parameters interrelated? Can you please explain the difference between bandwidth and settling time parameters mentioned in the datasheet?

    Following are the clarifications you asked for - 

    • Oscilloscope scales (TIME/DIV)
      • Figure 1: 1us/div
      • Figure 2: 1us/div
      • Figure 3: 10us/div
      • Figure 4: 4us/div
    • Plot showing the bus voltage and the scale used
      • Can you please tell me which bus voltage you require in the plot?

    • Is there a common mode event and, if so, what is the amplitude? Can this be plotted?
      • What do you mean by common mode event here? Kindly, explain

    • Can you provide a diagram showing where the current measurement is taking place and how your switches/load are organized?
      • Circuit is provided above. Let me know if you need more details.

    Thanks!

    Kashyap

     

  • Hi Kashyap,

    Thank you for bearing with me on this one, I haven't received a notification to your previous reply.
    I am now looking into the issue and will get back to you shortly.
  • Hi Kashyap,

    Bandwidth means how fast one can follow the input signal and with what attenuation, and this refers to a continuous signal, such as a sine wave, whose output is phase shifted (lags) from the input

    Settling time means how long the device will take to recover after a step change on the input voltage, such as a sharp transition from a high to low state or a differential reading. This is due to the internal circuitry.

    Given the requirements of your circuit I would suggest posting your question on the high speed opamps section.
  • Hi Carlos,

    Thanks alot for the explanations and suggestions!

    As suggested by you, I have posted my question in high speed op-amps section and here is the link of my post - 

    http://e2e.ti.com/support/amplifiers/precision_amplifiers/f/14/t/593398

    Best Regards,

    Kashyap Gundlapalli

  • Hi Kashyap,

    Thank you for your feedback.
    I might suggest adding a link to this thread, so that whoever takes care of your post is aware of some of the issues you experienced beforehand.
  • Hi KG,

    Seems to me you would have far better results to measure the secondary load side of the inverter e.g. Put shunt in series with load near ground side. About oscillations on INA output, Tina analysis results seem indicate with low value shunts 1m-0.5m the REF inputs have to be well isolated, perhaps use REF/2 from a precision voltage source. Input inductive ringing is especially true VS=3v3, perhaps not as prevalent VS=5v.

    For that reason we choose to use the A2 device and very low value shunt 0.5m and soon to test 0.2m shunt. Seems 50v/v gain offsets the low shunt impedance and inductive current ring at the INA input. It may be the setting time is faster on the A2 since it is not listed in the data sheet table Carlos posted.
  • That ringing also occurs from FET dv/dt might enter through either REF input and further aggravate the INA output issue.

    BTW: TI has introduced 80v GaN FET device that has a zero Trr, requires no body diode. The video training lab touts the absence of FET turn off dv/dt. GaN FET with integral gate driver is highly efficient since there is no Trr.

  • I think delay is very routine thing and it occurs while monitoring. If you are still looking for more efficient solution then use intellisens i2 probe, it can measure upto 100A. Check its specs if they match your requirements.

    www.taraztechnologies.com/.../