Co2 concentration drops below 400 ppm - MH-z16 Co2 sensor

I am not using any additional boards for this. I simply cut the wires and connected it directly to arduino using UART interface… Apparently there’s a I2C/UART interface board but I do not have this.

The only issue with this is that the cocentration of co2 keeps dropping below 400 which is not very likely and it goes all the way down to 250-200 ppm range… but however it’s very accurate in responding to an increase in co2…

I’m using the code that came with the library and have tried PWM as well but it returns a similar value after 10 seconds…

/***************************************************
* Infrared CO2 Sensor 0-50000ppm(Wide Range)
* ****************************************************
* The follow example is used to detect CO2 concentration.
  
* @author lg.gang(lg.gang@qq.com)
* @version  V1.0
* @date  2016-6-6
  
* GNU Lesser General Public License.
* See <http://www.gnu.org/licenses/> for details.
* All above must be included in any redistribution
* ****************************************************/ 
#include <SoftwareSerial.h>
SoftwareSerial mySerial(10, 11); // RX, TX
unsigned char hexdata[9] = {0xFF,0x01,0x86,0x00,0x00,0x00,0x00,0x00,0x79}; //Read the gas density command /Don't change the order
void setup() {
  
  Serial.begin(9600);
  while (!Serial) {

  }
  mySerial.begin(9600);

}

void loop() {
   mySerial.write(hexdata,9);
   delay(500);

 for(int i=0,j=0;i<9;i++)
 {
  if (mySerial.available()>0)
  {
     long hi,lo,CO2;
     int ch=mySerial.read();

    if(i==2){     hi=ch;   }   //High concentration
    if(i==3){     lo=ch;   }   //Low concentration
    if(i==8) {
               CO2=hi*256+lo;  //CO2 concentration
      Serial.print("CO2 concentration: ");
      Serial.print(CO2);
      Serial.println("ppm");      
               }

  }   
  
 } 

 }

From [url=https://raw.githubusercontent.com/SeeedDocument/Grove-CO2_Sensor/master/res/MH-Z16_CO2_datasheet_EN.pdf[/url] the data sheet:

  1. Notes for maintenance
    7.1 The sensor should be calibrated regularly. The cycle time is better to be no more than 6 months.

I think this is particularly relevant as you're on the bottom end of its range. Also it does have a stated error of +/-50ppm.