• State TI Thinks Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 2 answers
  • Subscribers 49 subscribers
  • Views 230 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

INA168: Do you have the maximum value of Total output error?

Kazuya Nakai59
Guru 20130 points
Part Number: INA168

 Hello guys,

 One of my customers is considering using INA168 for their new products.

 They want to know what the maximum value of total output error at all temperature range is to assume the worst case.

 But it can't be found in the device datasheet (SBOS122E).

 Could you please tell me the maimum value?

 Your reply would be much appreciated.

 Best regards,

 Kazuya.

  • 0
    TI__Mastermind 18780 points

    Kazuya-san,

    this is a hard question to answer objectively, because the error of the device relies on many variables throughout the system, and will change at different setpoints due to the offset error and gain error of the device. The distribution of the gain resistor will also factor in to this. In the data sheet, we specify the total output error over temperature is 2.5%, but this is at Vsense of 100mV:

    But the real error of the device looks more like this over the full scale range, given the chosen parameters:

  • 0 in reply to Carolus Andrews
    Guru 20130 points

     Hello Carolus,

     Thank you very much for your reply.

     Could I ask you a question about this error analisys tool?

     In the result you attached, the sense register is 25m ohm.

     So the difference voltage between IN+ and IN- is 100mV when the current flows to the resister is 4A.

     It seems that the total error at 4A and Ta=25C in the result is about 2.2% . 

     INA168 datasheet says that the maximum total output error at Vsense=100mV, VIN+=12V, Vs=5V, TA=25C is +/- 2%.

     Could you please tell me why the error analisys tool result is more than +/-2%?

     Thank you very much and best regards,

     Kazuya Nakai.

      

  • 0 in reply to Kazuya Nakai59
    TI__Mastermind 18780 points

    Kazuya-san,

    I have checked with our team to confirm, but here is what I believe is the cause of the discrepancy. The tool populated in the product folder is used across several different devices to quickly calculate the RSS approximate error curve for the specific use case defined. It calculates error for each device using the following formula (found by clicking the ? button in the tool):

    As you can see, this is a root-mean-square representation of the error of the device, but it does NOT take into account any hard limits set by TI, such as the worst case output error at room temperature. So in this case, so the calculator has the ability to show >2% here, because it is simply calculating error based on datasheet values, and is not programmed for each individual device to show that we would reject the device in test for failing to exhibit <2% error in this condition. In reality, the worst case error that would/could be seen in this condition would be the 2% listed in the datasheet, although typically it is lower than this value.

    I will let you know once I confirm this to be the case. 

  • 0 in reply to Carolus Andrews
    TI__Mastermind 18780 points

    Kazuya-san,

    I was partially correct here. For our current output devices, to fully be able to state gain error, the value of tolerance of output gain resistor needs to be known. To be able to utilize the tool on ti.com for these devices, we arbitrarily set the gain error to 2% at room temperature as a conservative estimate of gain error. This also establishes a hard limit on the curve of gain error, so the curve itself will converge on this value, as it does for output voltage parts. With this in mind, as I stated prior, the calculator simply performs a calculation, and is not programmed to show datasheet testing limitations, so there is potential that the calculator will show >2% error for the 100mV sense condition.

    As I said, the tool is meant for approximation only, and calculations should be performed based on datasheet criterion. To do this, I would begin with the total output error listed, and then also account for potential additions of the shunt resistor tolerance, PSRR and CMRR, and output gain resistor tolerance. This should give you an overall system error for room temperature condition. Let me know if this makes sense.