INA138: Amplifiers forum

Jason Wang1

Part Number: INA138

what will the Total Error % be with V(sense) at 3 mV?

from Online Current sense amplifier error analysis INA138 with 3mOhm, 24Vdc: about 36.1% total error at 1 Amp load current. please explain the possibilities of total error population.
1. 36.1% total error is the maximum?
2. if I have a hundred (100) of INA138 parts tested for the same design, what can I expect? the total errors can be completely random numbers between 0 and 36.1%?
from Figure 4. Total Output Error vs VIN in Datasheet: there is no trace data between 0mV and roughly 5mV of Vin. how to understand this region? for the same example above (Vin=3mV).

0 Ian Williams over 4 years ago

Hi Jason,

The 36.1% error reported by the tool is the expected RSS (root-sum-of-squares) error. This is the most realistic approach to determining worst-case error based on multiple device parameters. You can see more information in this tutorial video.
If you test 100 units, you will see a distribution of total error which will likely look like a Gaussian bell curve with values between 0% and 36.1%. It is possible, but extremely unlikely, that some parts would show > 36.1% error.
No information with Vin < 5mV is provided because of the shape of the error curve. Keep in mind that the offset voltage of the INA138 can be up to 2mV over temperature, which is very large compared to 5mV. That creates large errors, which you can see in the curve dropping off sharply. It's recommended to use Vin > 20mV or so to stay in the flatter part of the curve with better error performance.

Best regards,

Ian Williams
Applications Manager
Current Sensing

0 Jason Wang1 over 4 years ago in reply to Ian Williams

Thank you very much for your explanations.