• State
  • Locked Locked
  • Replies 5 replies
  • Subscribers 46 subscribers
  • Views 3065 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.

INA219 Shunt Voltage Reading

Justin Forrester
Expert 1015 points
Other Parts Discussed in Thread: INA219

Hi,

I am working with the INA219 and I am getting incorrect shunt voltage readings.  They appear to be double what I would expect given my current load and what I measure across the resistor with a voltmeter.

My test setup is as such: 2m Ohm shunt resistor, 100mA load, bus voltage of 11.5.  This is similar to the example setup used in the datasheet, so I initially used those same configuration parameters - Configure: 019Fh, Calibrate: 5000h.  I have since changed the configure register to 1FFFh, which is 16V Range, 8 PGA, and 128 sample averaging, hoping to get more precision.  I get the same results.

The INA219 reads shunt voltage of 0.4mV, bus voltage of 11.484V, and reads the current at 225mA, whereas with my multimeter I am getting 0.2mV, 11.48V, and 100mA.  What could be causing this problem?

 

Thanks!

Justin

  • 0
    TI__Genius 12325 points

    Hi Justin,

    One of the major sources of inaccuracy for current shunt monitors is differences in any resistances in series with the shunt voltage pins on the device. This can come from different trace lengths to each pin on your PCB or the tolerances of any series resistors being used for filtering or transient protection. The first step in tracking down the source of your inaccuracy would be to measure the voltage at the pins of the INA219 itself to see if it matches the expected value given your load current and shunt resistor value.

  • 0 in reply to John Caldwell
    Expert 1015 points

    Re-measuring the shunt voltage with a multimeter now shows it matching up with what the INA219 is showing, which is around 0.4mV.  These two match up now, but calculating the voltage drop across 2m Ohm resistor with a 100mA load gives me 2mV, which is half of what I am seeing.

    Furthermore, the current I am measuring is still different than the current the INA219 is reporting.  I measure 100mA, whereas the INA219 reports 220mA.

  • 0 in reply to Justin Forrester
    TI__Genius 12325 points

    Justin,

    The INA219 calculates the current from the voltage that appears at its pins and the shunt resistor value. Therefore, if the INA219 is measuring .4mV at its pins and the shunt resistor value is set at 2mOhm, the INA219 will report that your load current is around 200mA. The fact that the measured voltage at the pins is different than the voltage measured across the resistor tells me this is most likely an issue with your PCB layout. Is the INA219 as close to the shunt resistor as possible? Are the traces that connect the INA219 to the shunt resistor the exact same length? If you are using any input filtering, as shown in Figure 20 of the datasheet, have you ensured that the resistor values are well matched? These resistances should also be 10 ohms or less to maintain the part's accuracy.

  • 0 in reply to John Caldwell
    Expert 1015 points

    You may very well be right about the PCB layout.  In the next revision we will adjust the traces to the shunt resistor so that they are as close as possible to the pins on the INA219 and have equal lengths.  Right now one trace is fairly close, while the other is longer and goes through two vias.

    We are not currently using the input filter, as the datasheet said this was not needed in most applications.  When would you suggest it be used?

    Thanks for the help,

    Justin

  • 0 in reply to Justin Forrester
    TI__Genius 12325 points

    The input filtering network is really to prevent noise or transients from triggering the part's internal ESD structure. However, it is not necessary in all applications and the series resistors do introduce a bit of inaccuracy in the measurement.

    It definitely sounds like the pcb layout is the source of the inaccuracy you are seeing. Both the excessive trace length and vias contribute to a larger series resistance in one of the inputs. The bias current drawn by this input will cause an additional voltage drop that will appear in series with the shunt resistor voltage at the pins of the part. We really recommend equal length traces, devoid of vias, connecting the part to the shunt resistor.

    Good luck with the next revision!