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.

TINA/Spice/INA139: Simulation Models do not work correctly in bipolar configuration

Part Number: INA139
Other Parts Discussed in Thread: TINA-TI, INA169,

Tool/software: TINA-TI or Spice Models

TI,

I am having issues with the TI Models for the INA139 and INA169 Current Sense Monitors. Simulation files for each type are attached.

Neither model will work correctly in a bipolar configuration to measure negative current and also the INA139 will not converge properly in this configuration. I would expect the output of either model provided to be 1V no matter which direction the current is flowing, per the datasheet.

Can you assist with this? Much appreciated.

Bipolar_INA139.TSCBipolar_INA169.TSC

  • Hi BGX,

    you violated the Voutmax specification mentioned in section 7.3.1 of datasheet. Decrease the output resistances in your simulation from 10k to 1k and it will work.

    Kai

  • Kai,

    I dropped the resistances in the two attached simulations in the original post, same issue.

    The Voutmax specification in 7.3.1 :

    Voutmax = (V+) - 0.7 - (Vin+-Vin-)

    or Voutmax = Vin- - 0.5

    where V+ = 5V

    where Vin+-Vin- = 48-(48/481)*480 = 0.0997920998 approximately 0.1

    5 - 0.7 - 0.1 = 4.2 V

    or 5 - 0.5 = 4.5 V

    lower of the two = 4.2V

    The output should be Vout = 0.1*1*10 in this case or 1V < 4.2V so the compliance was not violated.

    Copied more or less from Figure 15 on Page 16 ( 8.2.4 Bipolar Current Measurement )

    -B
  • Hi BG,

    Kai

  • Hello BGX,

    Thanks for your interest in our INA139 and INA169.  According to the transfer function for the part, if it were powered from the common mode, I would expect your output from the voltage divider to be 9.335V .  However, since your supply is tied to 5 V, the device output is limited some value of 5-1.2V or something smaller.  As Kai mentioned you could change the resistors on the output of your INA169 or INA139.  Alternatively you could tie the device supply to the 24V bus.  If the bus will change and exceed 40V or 60V, we do have a work around (http://www.ti.com/lit/an/slla190/slla190.pdf)

  • Kai,

    I can reproduce your results if I rebuild the circuit. But for some reason I'm getting the same behavior regardless if I use a Sine AC source +24 to -24 V 100Hz.

    Notice how it will not not produce a Vout during the swing negative.

  • Bipolar_INA139_2.TSCOr rather positive in this case, since the source is moved to the other side.

  •   

    I realize why there was some confusion initially when reviewing the simulations. My original circuit for the INA139 (which kai used) did not save my change to the R2 resistor which appears as 24 ohms in the original post instead of the 240 ohms as intended. The INA169 sim correctly shows 480 ohms.

    That confusion out of the way, I used Kai's circuit that appears to work, rebuilt it exactly, and replaced the DC source with an AC one.

    The post with my waveform is what best illustrates the issue I am having. I can get them to work independently by manually altering the source --- but it does not appear to work properly when both are present. 

    I would like to confirm that at least one of you can reproduce the issue with the AC source, where the waveform for at least one current monitor (usually during the negative swing of the A/C) does not produce the expected result.

    Hopefully that clarifies the issue I am seeing.

    -B

  • Hi BG,

    Vg1 must not go negative, because the common-mode input range of INA139 is +2.7V...+40V. Bipolar operation does not mean that the input voltage is allowed to go negative, but only that IN+ is allowed to become negative in relation to IN-. That the output can never become negative can also be seen from the operation of diodes D1 and D2.

    Kai
  • I'm still missing something here.

    Both INA's are powered from the same (separate 5V) source. It is expected that the output of the INA cannot go negative.

    Where you lost me is Vg1. If Vg1 goes negative, this is equivalent to switching the polarity of the source, is it not? If that happens then the IN+ becomes negative with respect to IN-.

    I suppose I could artificially do this by shifting the level up. (by referencing the INA's to V-)

    It looks like doing so almost works. Much closer to what i was hoping to achieve.

  • Hi Bg,

    yes, applying a negative voltage to the GND pin of INA139 makes the circuit work:

    The simulation runs a bit stubbornly. Some parts need to be redrawn to make it work.

    To be able to reference the output voltage to 0V of circuit, the supply pin of INA139 must see a positive voltage. I have chosen 5V. And to make the total amplitude of VG1 lie within the common mode input voltage range of INA139, the GND pin of INA139 was pulled down to -15V.

    This example demonstrates how useful it is that the INA139 has a current output signal and not a voltage output signal.

    Kai

  • Hello BGX561, Kai,

    Looks like you all are headed in the right direction.  Thanks for your support on this Kai.

    As Kai noted the Common Mode voltage at the input pins VIN+and VIN- needs to stay within the device specification of 2.7V to 40V for the INA139 and 2.7V and 60V for the INA169.  TINA can be rather difficult to work with at times. One alternative method you could use for testing your implementation is posted in the following schematic, I also have attached the TINA file for your convenience.  In this particular simulation, I don’t have the shunt resistor, but I do have a voltage corresponding to what could be across that resistor depending on your load. For this simulation, if you duplicate it in a new separate TINA file, you could possibly run into convergence issues. To overcome these, change the parameters in red seen as seen in the table below. This table can be accessed in TINA by clicking “Analysis” and then “Analysis Parameters”. 

    4278.E2E_INA139_bidirectional.TSC

  • Hi Patrick,

    thanks for the hints!

    Kai