INA228: Large offset current found in current measurement

Hello Mitch, I did sent out the message but probably due to glitch, it seems like doesn't get to send out. The message is about:

Thanks for the reply. However, based on the input bias current vs CM voltage spec graph, it seems that bias current due to CM effect shouldn't be that high when Vcm=32V, it contributes around 0.025nA. But in my case, Vcm=32V, I_offset=0.8mA. I think it showed much higher than the spec graph.

About setup, I am using SMU as the power supply and load, with an 61/2 DMM measuring the current which showing an accurate 10mA, the offset is not coming from the setup itself, it should be on INA228 itself.

Layout-wise, we had followed INA228 layout guideline using Kevin connections. we can share the layout connections of the design:


Regards,
Yee Yong

Hey Yee Yong,

I see that you marked the first reply as resolved, so were you able to figure this out? (maybe it was from the noisy supply etc?) Let me know if you still need further assistance on this.

Hi Mitch,

I had changed a bit on my setup in which I had shorted IN+ and IN-, but we are still able to readback 0.8mA offset from INA228. Thus, I don't think this is due to noisy supply.

Regards,
Yee Yong

And yup, I still need further assistance on this

Hey Yee Yong,

Can you send me a readout of all register values during the unexpected current reading? Also please send me the expected current at that time, as well as the Maximum Expected Current used in calculations and the shunt resistor size (if different from above).

Regards,

Mitch

Unfortunately, I can't provide the register value, as our design INA228 will be read by CPLD, getting the readout of register value when unexpected current reading is not straightforward. But I can provide the address register setting:


For normal operation: 
Measured current by INA228=10.844mA
Expected current=10mA
Maximum expected current=10.478mA
Shunt resistor: 40mOhm, 1% (same resistor)

For Shorted condition:
Measured current by INA228=10.855mA
Expected current=0mA
Maximum expected current=0.478mA
Shunt resistor: 40mOhm, 1% (same resistor)

Hello Yee Yong,

Was there an error in your above message? Was the second case (shorted) supposed to be 0.855mA measured?

For your setup, are you having this issue with multiple boards/parts? Do you have another board or device that you can test?

When you short the inputs, are you shorting across C436 (where you labeled spare)?

In your original question, you mentioned that the offset voltage was decreased when you changed the common mode down to 6V. Can you test this again with the inputs shorted across C436? Ideally, short the inputs and sweep the common mode from 0V to 32V and see if/how the output offset changes.

Regards,

Mitch

Hi Mitch,
Yup, there is error in previous message. Let me correct my previous message:

"Unfortunately, I can't provide the register value, as our design INA228 will be read by CPLD, getting the readout of register value when unexpected current reading is not straightforward. But I can provide the address register setting:


For normal operation: 
Measured current by INA228=10.844mA
Expected current=10mA
Maximum expected current=10.478mA
Shunt resistor: 40mOhm, 1% (same resistor)

For shorted condition: (32V)
Measured current by INA228=0.077mA
Expected current=0mA
Maximum expected current=0.086mA (Typical case should be 0.018mA)
Shunt resistor: 40mOhm, 1% (same resistor)"

1.) For shorted condition the measured current by INA228 is pretty close to the max expected current. 
2.) For setup, I have this issue replicated on three INA228 (same configuration) on the same board.
3.) In shorted input case, YES, I did short across C436.
4.) For shorted condition: (6V), measured current by INA228=0.077mA. Looks like remain similar as 32V case.

Regards,
Yee Yong

Lee,

Mitch is out of office today, but I have let him know about the above and he should respond shortly. 

Hey Yee Yong,

Could you try increasing the conversion time and averages and see if the offset changes? Also, could you check the system GND and the GND pin of the device to make sure they match. It would also be a good idea to check for GND noise as well.

Regards,

Mitch

Hi Mitch, we did increased the conversion time for bosth VSHCT and VBUSCT to the max (4120us), which we have '0xBFC0' for ADC_CONFIG register.
It looks still the same.

For max conversion time: 
Measured current by INA228=10.855mA
Expected current=10mA
Maximum expected current=10.478mA
Shunt resistor: 40mOhm, 1% (same resistor)