ADC oneshot calibration max 3.114 (5.1.1 is worse) What is going on please?

[greg-dickson]

I was just experimenting with the ADC one shot read and found that reading a pin basically pulled high will max out at 3.114 volts.
The actual measured voltage at the pin is 3.3v.

Is this normal or is there something wrong.

Code: [Select all] [Expand/Collapse]

1. if((adc->error = adc_oneshot_read(adc_handle, adc->adc_channel , &(adc->raw_result))) == ESP_OK)
2.     {
3.           ESP_LOGI(TAG, "ADC%d Channel[%d] Raw Data: %d",adc->unit_id, adc->adc_channel, adc->raw_result);
4.           if (do_calibration == ESP_OK)
5.           {
6.             adc->error = adc_cali_raw_to_voltage(adc_cali_handle, adc->raw_result, &(adc->millivolts));  
7.             ESP_LOGI(TAG, "ADC%d Channel[%d] Calibrated Data: %d",adc->unit_id, adc->adc_channel, adc->millivolts);
8.           }
9.     }

GeSHi © Codebox Plus Extension

produces
adc_gpio_read: ADC0 Channel[2] Raw Data: 4095
adc_gpio_read: ADC0 Channel[2] Calibrated Data: 3114

Also then doing a resistance calculation on a voltage divider (using the calibrated millivolts) I need to use 3.25 instead of 3.3v as the supply voltage to get reasonable resistance values across the unknown side of the divider.

[greg-dickson]

Oh My go esp-idf5.1.1 is worse
This is on several S3 units.

Code: Select all

 
(14593) truc_adc_gpio_read: ADC0 Channel[2] Raw Data: 4095
(14593) truc_adc_gpio_read: ADC0 Channel[2] Calibrated Data: 3076

What's going on guys?