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.

INA186: INA186 - Measure Negative Supply

Part Number: INA186
Other Parts Discussed in Thread: INA282, INA286


I want to use the INA186 to measure  a +/-12V boost converter with one INA186 on each rail. Note that the INA186 supply voltages are provided separately by a 3.3V regulator. For reference, I've attached

a screen cap of the converter and where the current is to be monitored.

The output of the converter that will be measured  is +/-12V. To connect the sensing resistor on the negative supply, does one simply connect the sensing resistor the opposite as the +12V supply?  Seems logical because the common mode output of the converter is at system ground, so either way the negative current is referenced to the same point as the positive current.

Bad assumption, or will it work? BTW - it's amazing there are no reference designs on this seemingly basic application as many designs use negative supplies that need current monitoring.


  • Maybe I solved this with more digging through TI datasheets. I finally found one that provides a design example with a negative supply. Please see below.

    If I connect the GND (instead of -10V) to GND, my range should be -14V to 80V, which is good for the application (-12V side). Only two questions remain.

    1 - I expect highest currents to peak around 1000mA. Using this device with a gain of 50, the highest Vo would be 1000mA * 50 * 0.1 = 5. Since my ADC has a 3V maximum input, I will instead use a 50mOhm resistor to keep it to 2.5V maximum. Since on the negative supply I never expect the voltage to be above GND, can I simply connect the -12V output to the "-IN" sense pin(resistor) and the "+IN" to the load? (current flows back to the supply pos to neg) This way I will always get a positive output at Vo, starting at 0V.

    2 - As for reference - can I simply connect the REF1 and REF2 pins to GND, as this will now be a unidirectional operation for the negative supply?

    Thanks for your help and feedback.

  • Hello groger57,

    I am going through your questions right now and will respond shortly.



  • Hey groger57,

    The INA186 will not be able to properly function on a -12V (and GND pin = system ground).

    The INA282 will be able to survive sensing on a -12V return rail as you point out.

    To address your first post, you could sense either the +12V rail or the -12V rail. If all the current is to return though the -12V rail, then yes it would be the same current sensed. Whether to sense both or just one will depend upon your system requirements. The advantage to sensing high side (either +12V and -12V) is that you could know if there was some short-circuit event on the +12V or -12V rail, while on sensing on the low-side (system ground) you would not be able to see this.



  • Hi Peter,

    Yes, I believe the INA282 would be the right choice. But please clarify - I am unfamiliar with the diagram as made in your simulator tool.Are you showing it as a negative supply connected up like a "normal" positive supply? it appears so!

    The source, if a negative supply WRT the INA286 supply and ground should have the NEG terminal with (conventional) current flowing back since it's a negative terminal. The Cosel MGFW400512 device itself is marked as "-Vout". Also  - your resistor Rsense shows the + terminal to the supply side.

    1 - If the INA282 is to be used to sense the output current flowing back as a biderectional sensor your output appears correct as (1.0 * 0.05 * 50) = 2.5V  with a 5V supply and Vref 1 and 2 at ground. I'm still a bit perplexed about the polarity shown!

    2 - Can you clarify that the sense resistor is connected as shown in my diagram posted, WRT the converter?


  • Hey groger57,

    You are correct that in my simulation the load polarity with respect to the voltage source is not reflective of what is happening in real life (load really will go other direction, thus device inputs need to be switched). In the simulation the device is still sensing a positive differential voltage and it will thus amplify it w.r.t. to system ground. The math still works out the same as the simulation will force the parameters given. I apologize for the unnecessary confusion.

    How you have shown in your scope shot will still work.

    Hope this remedies the evaluation.



  • Peter,

    No confusion at all. Please refer to the screen cap I posted, R38, terminals 1 & 2.

    To connect should R38 terminal 1 go to the POS or NEG input of the INA282?(I am thinking NEG)

    The misunderstanding is on my side, not really understanding how the device works in terms of bidirectional and positive/negative supplies.

    When connected this way and ref1 and ref2 tied to ground, and because this is a unidirectional application Vout will always start at 0 and go to some maximum value (close to 5V if current was 2A) Yes?

    Is this all correct?

    Thanks again...

  • Hey groger57,

    All that is needed to answer your question is understanding the gain equation of the device (and all other current sense amplifiers), which is:

    Vout = Vdiff*Gain + Vref

    Vdiff = (V_IN+) - (V_IN-)

    where Vout and Vref are with respect to device ground and Vdiff is the differential input/shunt voltage.

    So if REF pins are grounded (Vref=0V), then you want Vdiff > 0. To make this happen in your screen cap, the IN+ pin needs to connect to the load side of R38, while IN- pin connects to the VOUT- side. This is assuming current is flowing from load side and into the VOUT- pin of the regulator. I can't see in the picture what the 1 and 2 terminal are.

    Hope this makes sense.