You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
After many months of using the current sensor cards, I have found that the analog conditioning front-end drifts slightly over time...
Recall how the circuitry works:
the LEM sensor outputs a current proportional to the sense current
a burden resistor converts this current to a voltage
the voltage is conditioned and sensed by ADC
the digital ADC code is sent to the AMDC
The issue is that the 0A input voltage into the ADC is a mid-scale value... i.e. VREF/2 or something. This means that the final ADC code the user gets on the AMDC is not zero for 0A current. It is some value based on the circuitry. The issue is that this mid-scale value slowly drifts over period of months such that the offset is not fixed... This caused ~100mA of offset for me, which was problematic.
Note that the GAIN of the sensor and analog circuitry is pretty stable... I found no issues with the gain. Only the offset.
My quick fix is to simply calibrate the offset at boot every time I run the bench... I flow 0A of current and record the mid-scale voltage and just subtract that from the ADC.
TODO
A better solution would be to move to an analog front that does this offset calibration in hardware... see the LV sensor card for an example of how to do this.
The text was updated successfully, but these errors were encountered:
After many months of using the current sensor cards, I have found that the analog conditioning front-end drifts slightly over time...
Recall how the circuitry works:
The issue is that the 0A input voltage into the ADC is a mid-scale value... i.e. VREF/2 or something. This means that the final ADC code the user gets on the AMDC is not zero for 0A current. It is some value based on the circuitry. The issue is that this mid-scale value slowly drifts over period of months such that the offset is not fixed... This caused ~100mA of offset for me, which was problematic.
Note that the GAIN of the sensor and analog circuitry is pretty stable... I found no issues with the gain. Only the offset.
My quick fix is to simply calibrate the offset at boot every time I run the bench... I flow 0A of current and record the mid-scale voltage and just subtract that from the ADC.
TODO
A better solution would be to move to an analog front that does this offset calibration in hardware... see the LV sensor card for an example of how to do this.
The text was updated successfully, but these errors were encountered: