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.

INA303: Low side configuration and 0V REF adds a delay

Part Number: INA303

Hi,

I am currently working on a design and I am using the INA303A2. I an using the evaluation module INA303EVM (the one with 50V/V gain). Here is the schematic I am using, with some notes.

I have a 1mOhm sense resistor to measure the current. I am using this in a low side configuration, so IN- is connected to GND. I cannot figure out why, but using the INA303 in a low side configuration, as well as using a 0V reference voltage creates a delay on the output. For instance : 

Blue is the output measured directly on the evaluation board (TP122) and teal is the voltage across the shunt resistor. As can be seen, it takes about 60us for the output to response. Please not that the delay vanishes if I use a reference of Vs/2 = 2.5V. I have not yet tried other values of reference voltage.

Interestingly, if I remove the shunt resistor and apply a waveform generator directly on the IN+ and IN- pins, the problem also vanishes. But as soon as I load the waveform generator enough, the common-mode voltage drops and the delay is back again. Also, increasing the frequency of the waveform generator decreases the delay.

All this leads me to think that this is a bias current issue. It seems to me like there is a minimum common-mode voltage needed to bias the inputs, or a non-zero reference voltage. I did not expect so however, as the datasheet states that the device can be used in both low-side or high-side configuration. The datasheet also specifies that the common-mode input range is -0.1 to 36V. But again, this is for a Vs/2 reference...

I will probably end up using a non-zero reference voltage, but I would like to know why this happens, and what minimum value of reference voltage I can use to avoid this delay. I'm using the INA303 for overcurrent detection, and ultimatly short-circuit protection, so a 60us delay to me is not acceptable. 

I have already tried : 

- Adding some filtering on the shunt resistor  inputs (10ohms with 1000pF)

- Using a 2.5V reference as mentionned earlier. This does solve the problem. I haven't tried any other values.

- Changing the Ilimit or delay of the alert (Obviously this did not help) 

Any help is greatly appreciated.

Thank you,

Olivier

  • Hello Olivier,

    Thanks for considering to use Texas Instruments in your design. If I understand correctly, you apply a 5A load across a 1mohm shunt, and you see a delayed 185mV output when you have a grounded reference. Even after accounting for gain error, I would expect your output to be greater than what you are measuring. Do you get the same output level when the reference is set to mid-supply? Did you place your input filter at R32, R72, and C32? Also are you measuring directly at the output pin or after a filter on the output? Are you able to monitor the quiescent current of the supply? I suspect the device may be damaged. You may want to consider replacing the part on your evm with a new part and check if the performance improves.
  • Hello Patrick,

    thank you for your reply. The scope screenshot I attached was actually with a different setup. I had a 5.6ohm shunt resistor and I used a waveform generator. Nevertheless, I have the same problem with the real setup (1 mohm shunt). Thank you for your suggestion, I will try replacing the part and see if I still have the issue. I do have the input filter installed. I am using two 5.6 ohms and a 1000pF capacitor. I am measuring the output directly, no filters.

    Kind regards,

    Olivier

  • Hi Patrick,

    unfortunatly, changing the part did not help in solving the issue. Here is a screenshot I took with a 1.25mOhm shunt resistor. I am now measuring the current through the shunt (green) as well as the direct output of the INA303 (teal). The following screenshot uses a 0V reference.

    I have the delay again. The less current I pass through the shunt, the longer the delay, as if there as to be a minimum voltage across the shunt for the INA303 to work properly. Is this delay expected with a 0V reference?

    Here is a screenshot with now a 2.5V reference (I am now mesuring the output in AC coupling) : 

    I also measured the quescient current. It is exactly 1mA. I disconnected pin 1 from the board and applied 5V directly to it with an ampmeter in series. It inscreases to 1.060 mA when I have 2A running through the shunt.

    Thanks,

    Olivier

  • Hi Olivier,

    with a reference voltage of 0V and a very small input voltage the output tries to fully go down to 0V. But it cannot manage this and saturates. I think what you see is sort of overload recovery time because the output of INA303 needs some time to leave the negative rail. Setting the reference voltage slightly above 0V (50m...100mV) should eliminate the delay.

    Kai

  • Hi Kai,

    yup, I tested with 50mV and the delay is gone. Thank you very much to both of you.

    Best regards,

    Olivier