Oh, a little can not remember now.
Yesterday morning, the timer is initialized changed, ready to write interrupt service routine when my predecessor amassing once awoke - debugging process to transfer a piece. How should I say, I say a piece of the tune as if suddenly understand, but the actual writing process, working together on the stupid.
So the afternoon start debugging ADC, the beginning there are two problems, one is the return value ret is not the number of ADC conversion value is stable after jumping 0 and the other is to convert (* 3.3 / 4096).
After the discovery of ADC initialization process is not stepping in to enter * 3.3 / 4096 that function inside, in order to verify that an ADC alone so this function into the main function of, ret this issue is resolved.
0 jump this thing looked older generation with an oscilloscope output waveform MG811, in theory, should be stable, but actually more like a sawtooth waveform. So my ADC function without any problems, the main cause of the jump 0 analog output of the sensor.
So take me to the senior leadership we ask this question, there are several possible causes, recorded as follows: MG811 is 6V power supply, give me this circuit is 5V, the power may be enough, there are current.
I also think of temperature compensated output MG811 is me hanging, and today asked two of Taobao sellers, concluded that the 5V power supply and temperature compensation vacant no big impact, but the 5V power supply may result in inaccurate measurements.
In addition, the information I found online MG811 does not matter where the analog voltage output and CO2 concentration, I also mentioned before, the formula I use myself regression analysis using the WPS out. But the leadership said it would not do, ah, will continue to find relevant information today.