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.

INA213: Increasing output voltage error with increasing common mode voltage

Part Number: INA213
Other Parts Discussed in Thread: SM72295

Hi:

We are using 2 of these devices in an MPPT battery charging circuit, one to measure battery current and the other to measure solar array current. The charger design is based on the TI PMP7605 reference design. We are using a 0.002 ohm sense resistor wired straight across the inputs on the high side of the bridge. The sense resistors are located exactly as they are in the reference design... we are using the INA213's in place of the current sense amps that are integrated with the SM72295.

What I am seeing is that the measured battery charge current is always within 5% of actual but the measured current from the solar array has an error that increases with increasing voltage across the solar array. For a given battery voltage (Vb), I can vary the duty cycle of the switching MOSFETs to vary the voltage that appears across the solar array connector (Vs) from Vb up to about 20v (and beyond if I wanted to but I am aware that the absolute maximum is 26v for the INA213 part). For Vs equal to, or slightly greater than Vb (around 12v), the measured solar array current is OK. As I increase Vs towards the 20v limit however, the measurement error increases to give me increasingly higher solar array currents with errors > 30% of the actual current through the sense resistor. It seems as if I am seeing some kind of gain error that increases with increasing Vcm or perhaps my input bias current is increasing with Vcm?

Thanks for the help

 

  • Hello Nicholas,

    Thanks for considering Texas Instruments for you solar power system.  If I understand correctly, your modified system looks something like figure 1 below.  Changing the common mode and power supply definitely can change how much error you see.  Calculations can be done for  the worst case error expected according to equations 1 and 2.  However, 30% error does seem rather high.  From some quick calculations, I took a root mean square sum of the CMRR error, PSRR error, gain error, and offset error.  I assumed a max common mode voltage of 20V, a max supply voltage of 20V, and a max current of 20A across your shunt.  From my calculation, I would expect you to see around 2.7% error in measurement.  So I must be missing some details here.  Could you show me in a basic diagram of the INA213 at the solar array shunt what output was seen with a specific supply, common mode voltage, and sense current?

    Figure 1

      

    Equations 1, 2

  • Thanks Patrick

    One thing I forgot to mention is that our Vcc to the INA213 is 3.3v

    I will get some specific voltage and current data for you tomorrow.

  • Hi Patrick:

    The diagram you provided is correct.

    As I mentioned above, Vcc is 3.3v. (Perhaps this isn't an optimal Vcc for this part?)

    As the circuit is operating at the moment, I measure 6.75 mV through the 0.002 ohm resistor for 3.37A (my actual measured current from the solar array is 3.45A) With a gain of 50, I should measure appx. 0.34V at the INA213 output. My actual measured output voltage however is 0.435v.

    Vcm is 15.2v.

  • Hello,

    Thank you for providing the Vcc value.  That still falls within the recommended operation range for the part, so it should not be an issue.  So if I understand correctly, you measured 6.75mV accross the shunt.  Did you measure directly or near the INA213 input pins?  (Trace mismatches and  measurements a few centimeters from these pins can produce larger perceived errors).  Actual current is 6.75mV x 0.002 ohms = 3.375A and your output from the INA213 was 0.345V, which would match 0.345V/50V/V/0.002ohm = 3.45A.    Or was the output actually 0.435V?  If the output from the INA213 was 0.345V, that would give an error of 2.22% (abs(.3375-0.345)/.3375*100).  Given your operating conditions that does seem reasonable.  There are several sources of error that would contribute to this, and if you would like to learn more about it, I would recommend watching  our video series located here: https://training.ti.com/getting-started-current-sense-amplifiers

    However, if you are a little pressed for time and just want a quick answer on to what kind of errors you may expect for all of your operating conditions.  I would go to the INA213 product page and enter conditions in the error curve tool, like what I have posted in the figure below:

  • Hello Nicholas,

    How is your debugging coming along? Do you have any further questions?
  • Hello Nicholas,

    I have not heard from you in a while. I presume you were able to resolve your issue, so I am going to close the thread. However, if you need to revisit this issue for more help, feel free post here again or start a new thread. Good Luck on your solar power system.
  • Hi Patrick:

    Sorry I dropped off, I got distracted by another issue.

    I need to do a respin of this board to address some other issues and so I need to make a decision on this part. As the board is operating on my bench now, I am observing the following.

    Circuit current as measured with a clamp meter: 4.41A

    Voltage across 0.002 ohm resistor: 0.00895v (0.00895 / 0.002 = 4.475A)

    Output voltage of INA213: 0.529v (It should be 0.00895 * 50 = 0.4475v)

    Common Mode voltage: 15.4v

    So, the output voltage error is currently ( (529-447) / 447 ) * 100 = 18.2%

    The circuit is properly laid out and routed with the current sense resistor located directly next to the part pins 4 and 5 with symmetrically routed leads of equal / matched length to the pins. I will try to attach the gerbers of the layer it is on

  • I attached a screen shot that shows how I have the sense resistor routed and it location relative to the device.

    Concerning the magnitude of this error in relation to Vcm, it's been a bit difficult to correlate these factors so, for now, I am downplaying that aspect of the problem.

    With Vcm at around 15v however, I should not be seeing the error in output voltage that I am. Any ideas are appreciated.
  • Hello Nicholas,

    That measurement is higher than expected considering you are well above the swing to gnd specification and the offset should only contribute 5mV. I am going to check with my team on possible causes, but there is one thing I would like for you to check in the mean time. This device has a REF pin, which I noticed you tied to the ground pin. However, how close is this connection to the ground plane? Perhaps, there may be a small differential between these pins and your ground plane or the ground point you are probing from for measuring the output voltage.
  • The REF and GND pins are tied together with a via close by and that via is routed through a trace to the plane (which is split). That could be part of the problem but the other INA part that measures battery current is routed identically and works fine.

    I tried tying the REF and GND pins to different ground points but couldn't affect the output.

    I split the ground plane under the SM72295 in order to separate the power ground plane (for the switching MOSFETS) from the digital ground plane and I tie them together near the power supplies. Is there a way for me to send you the design files so that I don't have to post them here? I would like to have the layout reviewed.

  • Hello Patrick:

    Where can I upload my design files fro review?
  • Hello Nicholas,
    You should receive an email from E2E about a request to connect. In that request, should be my email so you can send me your files.

  • For those curious about the final conclusion over what was causing the error, it was due to connecting the INA213 output directly to a SAR ADC.  Typically Texas Instruments recommends placing a charge bucket filter on the input of a SAR ADC, because the internal sample and hold circuitry of the ADC requires a quick charge dump that a typical low bandwidth amplifier cannot otherwise provide.  For more information  on this topic, you can watch a training video here.  To calculate the required RC value for this filter, you can use a calculator TI provides, located here.  This calculator actually performs many different functions, so for the RC filter calculation, refer to the following image below.