• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 48 subscribers
  • Views 2641 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

OPA2376 as differential amp for current sensing

Degu
Prodigy 90 points

Other Parts Discussed in Thread: OPA2376, OPA376, TIPD175, ADS1110

Hi!

I've used the OPA2376 in a small test product. A currently unused feature is to measure a uncritical given USB current (0 - 3A) by using a shunt resistor of 10mOhm.

The goal was to have an OP output from 0 - 256mV which goes into an ADC with an internal gain of 8 to meet the 2048mV ref voltage -> 0 - 3,2A represents 0 - 2048mV

I've build up 100 PCBs, and I've observed @ 40 - 50% of these some issues with the output voltage of the OPA2376.

I used a gain of ~8,6 with 0603 1% resistors and a buffer + small low pass filter @ the output.

Is there any issue with the low gain or low voltage output of this IC? The OPA is also powered by "USB_5V_IN" voltage (+100nF bypass).

Many thanks in advance and best regards!

  • 0
    TI__Mastermind 25875 points

    Hello Degu,

    This is similar to a TI Precision Design that we have.  You may find some if it helpful.  Here is the link:  http://www.ti.com/tool/tipd175

    Note that the layout is important...make the input traces as short and balanced as possible.  Be sure to Kelvin-connect the shunt resistor.

    Overall I don't see any functional issues with the schematic, though I should note that you cannot measure 0 current.  In this scenario the minimum current that you can measure is ~620mA, which represents the minimum output swing voltage of the OPA376 (50mV) divided by the gain (8.06) divided by the shunt resistance (10mohms).  

    Concerning the gain, if possible I recommend placing all the gain up front in the signal chain and using a different ADC.  Here is an article that explains the concept in more detail:

    https://e2e.ti.com/blogs_/archives/b/thesignal/archive/2013/01/21/put-gain-up-front-waxing-philosophical

    In addition, I recommend using precision resistors (0.1%) for R117-R120 because they add gain error.  This is discussed in TIPD175.

    Honestly I don't know what you mean by "...some issues with the output voltage...".  I can only guess that at load currents <620mA you see little or no change in the output voltage.  If that's the case then you already have the answer.  If not, please provide more information (e.g. measured values of inputs and output at the pins of the device...oscilloscope pictures are the best).

  • 0 in reply to Pete%20Semig
    Prodigy 90 points

    Hi Pete and thanks' for your fast reply!

    Unfortunately I haven't currently one of this PCB's there where show the "issue" otherwise I had posted some oszi screenshots. As I wrote on ~50%, maybe only 40% (I haven't count it) of the test PCBs I observed that the measured current (terminal output from I²C) from the connected ADC (ADS1110) is very high, even there is no big current flow through the shunt resistor -> I can remember that the compensate factor in my firmware was very high to match the "0 current error". At some PCB's there was measured by ~6000 - 10000 of max resolution of 15Bit (by max 2048mVpp@ADC input) => 500mv - 625mV error at the OPA's output by ~0A current.

    In the first impression I thought about unstable OP circuit (oscillating), but I can't believe it, that's why I posted this circuit here to be sure that everything is ok. Is it possible, regarding the minimum output swing voltage of ~50mV, that some OPA's generate this offset at the output if there're is no big current through the shunt resistor (including OP's manufacturing variances, gain error because of 1% resistors etc.)?

    Because of the output swing behavior, you wrote 50mV, this is at RL of 2KOhm's, can you explain why you choose this value? After the first diff amp, a buffer is placed (second amp in the OPA case), so normally this should be high impedance for the first stage?

    The other thing is, why was it working by the other ~50% of the test PCB's? I was able to measure exactly the given current through the shunt resistor from 0 to ~3A in 1mA step's.

    Many thank's and best regards

    Degu

  • 0 in reply to Degu
    TI__Mastermind 25875 points
    Hello Degu,

    Have you tried swapping the amplifier from a known good system to one that is experiencing issues? This will tell us if the issue follows the amplifier or the system.

    Concerning troubleshooting, my advice is to use an oscilloscope to look at the inputs of IC29A (right at the pin 1 of R117 and Pin 1 of R119) and the output of IC29B (pin 7) on a board that is failing PCB. I recommend removing R116 to ensure no stability issues. This will isolate the amplifier so we can narrow down the source of the issue. If we see that the difference amplifier is working correctly, then the issue must be somewhere else in the system. Please post the oscilloscope pictures when you get them.

    Concerning the output swing, I generally use worst case scenario. For IC29A it will likely swing to within 10mV of the rails because the impedance seen by the output is much greater than 2kohms.

    Please note that the output swing I mentioned (50mV) is not exactly an 'offset voltage'. The limitation is due to the output stage of the device. The term 'offset voltage' is generally used with respect to the input of the device and is initially due to the mismatch in the input stage transistors. For more information, please watch the TI Precision Labs videos on Vos & Ib and I/O Limitations. They can be found at www.ti.com/precisionlabs

    I honestly can't answer why half of the boards don't work until we find the root cause of the issue. The first step is to isolate the amplifier on one of the failing boards as mentioned above.