Tool/software:
Hi Team,
I've come across an ADC read error which i'm not sure how to solve. Essentually, reading any voltage,not GND or 3.3, produces a significant error. For example, when reading half scale 1.65 voltage, I get an output of 1.633 voltage which is around 1% error, but is significant for digital power applications.
Please see the methodology below:
Specs
Vref Selection - Internal 3V3:
ADC input selection - A1
Vref Switches: POS1 Down
Input voltage measured: 1.650 V measured with Fluke Multimeter at HSECDOCK Connector pin 11
ADC measurement by C2000:
ADC Register input - 0x07EB = 1.63308105469 V
1) For the future, if the default configuration of the SDK examples is using the internal 3.3V pins, and the default configuration of the EVM switches is external 3.0V into the VREF device, there should be some clear documentation in the ADC example software that the switches must be changed.
From Rev E1 to Rev A, we've made changes to the Vref circuity, could you please clarify why these changes were made and what benefit they have to the board design?
I also did a test with external Vref on the SOM:
ADC output from C2K: 0x948 = 1.77465820313 V which is 7.5% error
Hi Nishka,
Can you help answers the below queries :
I will take the below suggestion for the next revisions of software.
1) For the future, if the default configuration of the SDK examples is using the internal 3.3V pins, and the default configuration of the EVM switches is external 3.0V into the VREF device, there should be some clear documentation in the ADC example software that the switches must be changed.
I will get back to you on this.
From Rev E1 to Rev A, we've made changes to the Vref circuity, could you please clarify why these changes were made and what benefit they have to the board design?
BR,
Nilesh
From Rev E1 to Rev A, we've made changes to the Vref circuity, could you please clarify why these changes were made and what benefit they have to the board design?
I had a look at the schematics for both E1 & A revision. The reason for change in VREF circuitry is for avoiding measurement errors due to high impedance (1K) between 4.7uF & OPAMP output (Through Switch).
Revision A Revision E1
I just realized that you are using the E1 version, to fix the issue i would suggest you to replace the 1K resistor (R53 & R57) with a 0 ohm.
BR,
Nilesh