INA186: Recommended shunt resistance value

Hello ttd,

This is a perfectly acceptable condition for INA186. It is a low-side and high-side capable CSA. It is specified to accurately sense shunt voltage down to VCM=-0.1V.

For the choice of ADC, you may not need very fast sampling because the INA186 -3dB BW is max 45kHz. The INA186 may not be able to drive certain SAR ADCs if the conversion/acquisition times are too short. ADCs with delta-sigma architecture are easier to drive. It is possible to use the INA186 spice simulation model to simulate how well it can drive SAR ADCs and how to tune a RC charge bucket filter since the INA186 spice model has the closed-loop BW and output impedance modeled. But a slower delta-sigma may be all that is needed. Refer to the training material below:

https://www.ti.com/lit/an/sboa443/sboa443.pdf?ts=1663092386044&ref_url=https%253A%252F%252Fwww.google.com%252F

https://training.ti.com/ti-precision-labs-adcs-final-sar-adc-drive-simulations?context=1139747-1140267-1128375-1139106-1134076

Choosing the right shunt resistor and ADC all comes down to system constraints like accuracy, available supply voltage for INA186/ADC, signal throughput speed, allowed shunt power dissipation, etc. The INA186 should be able to easily and accurately measure a 400:1 load current dynamic range. If I were making this circuit, I would try giving the INA185 the largest Vs supply voltage as this helps broaden accurate dynamic range, so make Vs=5V. I will choose a shunt that make Vout ~= 4.95V at 400mA (max Vout swing to Vs is Vs-40mV=4.96V). If I go with INA186A3, then shunt needs to be (4.95V/100)/400mA = 123.75mΩ. If you round this down to 123mΩ, then Vout (at I=1mA) = 1mA*123mΩ*100V/V = 12.3mV, which is probably too low even though INA186 is specified with zero current output voltage (Vzl) of 10mV over temperature. The reason is that offset error may become too significant at this point and device is not specified for gain error this low.

The algorithm would then call for decreasing INA186 gain to help provide more dynamic range. For INA186A1, the required shunt is (4.955V/20)/400mA = 619mΩ rounded down. Thus at 1mA, Vout = 61.9mV, which positions Vout in a much better point for linearity and to help decrease offset error. Keep in mind that the device's gain error is specified from 100mV to Vs-100mV. It can maintain this accuracy down to Vout = ~50mV, but it is not recommended to go lower as the output stage may not be linear anymore.

Based upon the exercise above, it is clear that you would need the INA186A1 operating with Vs=5V and shunt = 500mΩ ~619mΩ. Since you are needing all the possible dynamic range, I would try to match datasheet conditions for swing and a load resistor to the output that does not pull more than ~200uA of current at full scale. (In datasheet, the swing specs are tested with Vs=1.8V and Rload=10kΩ, thus Iout ~1.8V/10kΩ = 180µA at full scale). This equates to a 25kΩ load resistor for you application. Having a load resistor helps Vout swing closer to ground, but if the load resistance is too small, then Vout cannot swing as close to Vs. Thus it is a balance, refer to Figure 6-5 of datasheet.

Furthermore I would try to see if I could directly measure Vout with a delta-sigma ADC, but be prepared to need/validate using an op amp buffer in between INA186 and ADC.

To understand how to calculate total error, you could use this document.

https://www.ti.com/lit/an/sboa336/sboa336.pdf 

Sincerely,

Peter

Hi Peter,

Thank you for your reply.

I'd like to check the documentation which you suggested.

INA186A1 is best for my application.

By the way, current is from -1mA  to 1mA while target load is standby.

Then does INA186A1 sense from -1mA at Vout = 61.9mV  and 0A at Vout =123.8mV and 1mA  at Vout =185.7mV ?

It means current sense minimum resolution seems to be 1mA for maintaining accuracy.

Or ADC's resolution is high such as 24bit delta sigma , does it cover low voltage range with  maintaining accuracy?

 If switch the appropriate large value of shunt resistance, I think it is better for accuracy. But circuit is little complicated.

However same value like 500mΩ ~619mΩ for -1mA to 1mA and -1mA to 400mA, it's circuit is simple.

Best Regards,

ttd