INA826 destroyed with 30V Differential?

first_name last_name13,

Based on my understanding of your initial circuit description, I have setup the attached simulation using TINA-TI. Please take a look at it and verify that it accurately represents your circuit you have built and tested. To help you troubleshoot your device failure, please let me know if your schematic differs from the simulation I have attached.

2211.INA826_test.TSC

I would also like more detail as to where you installed the 51k resistor (R1) in your circuit. You mentioned it being “across the reference” but please clarify the placement.

To help me better understand your testing setup and application, please define the maximum and minimum input voltages for V2 and V3, with respect to ground.

Thanks for looking into this. To answer your questions: the 51k resistor is across the Rg inputs.  Make that change and your schematic is a rough approximation of what is going on.  I've done some more probing and it seems that when the power supply turns on, there are some brief transients up to 34V on the positive rail, but that should be within spec.

If everything works as expected, V2 should be between 0 and 15v and V3 should be the common mode of a 0 to 20 v differential power supply that floats at 1/2 of the positive rail.  In other words V3 should always be 15V  (this common mode is provided by running both sides of the diff-power supply through a RC filter and then through a 10k voltage divider).  If this is confusing then just take a look at my TINA schematic 2376.INA826 Absolute Value Difference.TSC.

The output goes into a INA225 Vbus measuring ADC with a 10k pull-down resistor.

Thanks again

Matt

Matt,

Using a 51k gain setting resistor, you should expect a gain for the INA to be approximately 1.968. I took a look at your TINA simulation and I am assuming your expected input voltage into the IN (-) pin on the INA to look like Figure 1, which has a range of approximately 0 to 15V.

Figure 1: IN (-) into INA826

I noticed in your TINA simulation, the output of U1 is not wired to anything. I also noticed your non-inverting input pin on U1 OPA209 is wired to ground. Assuming that the output of U1 connects to IN (+) on the INA826, this could cause the IN (+) to deviate from your expected steady 15V.

Please let me know if the circuit you have built/tested also has the non-inverting pin of that op-amp wired to ground. Thank you.

Hi Micheal,

Thanks for your comments, but I'm afraid that in by haste to post a response I hadn't noticed that I had been messing with the saved model.  U1 doesn't actually exist.  The outputs of R1 and R7 go straight to the non-inverting input of the INA826.  I have since discovered some further issues that may be my problem.  The load is actually a relatively complex inductive load driven by an MOSFET H-bridge.  It turns out that if the bridge driver experiences any sort of fault condition, it turns off all four gates of the bridge.  The inductive energy has to go somewhere and I thought it would just conduct back through the body diodes of the bridge causing a voltage change only equal to the forward voltage of these diodes (less than a volt) which should not be enough to fry the INA826.  This may be fine for the ground connection, but the high side is connected to a Buck regulator that may not be able to shunt away the energy.  I'm thinking of adding a TVS diode or zener or both across the regulator outputs to handle high side voltage injection from the load.  I'm still working on a realistic model for the regulator that shows this phenomenon.

Thanks,

Matt

Matt,

I agree with your plan of action and would always recommend protecting your devices and other circuitry from an inductive spike in voltage by using a zener diode or alternative method of sinking current on the voltage supply. Thank you for communicating your findings.