[SOLVED] How to use single point calibration formula to adjust span of MQ135?

[samic45mit1]

Dear All,

I am working on MQ135 gas sensor and interfacing it with micro controller. I read lot of paper for the calibration of MQ135 gas sensor. I want to use single point calibration formula to
adjust span of MQ135. This is possible? I am using formula found in a application note (please see the attachment).

Thankyou

 

[FvM]

What are you exactly asking for?

Why shouldn't a "single point calibration" be possible? Of course it requires a previous zero calibration, so mathematically it's rather a two point (or zero + span) calibration.

 

[samic45mit1]

Thanks for replay,

Actually I working on gas detection system. I read the data sheet of MQ 135, In its response curve (as given in data sheet.Please see the attachment) which is non linear. I am confused that mathematically it is correct or not. In my project we are using two detector units. We keep two detector unit together and use one calibrated unit to calibrate other unit which having the error. So this procedure is correct or not and this is possible with single point calibration???.

 

[FvM]

The single level caliration model presumes that either the measurement characteristic is linear, the sensor/instrument is used in a small (almost linear) range of the characteristic, or the sensor characteristic follows a known non-linear function that can be reversed in signal processing.

MQ135 is surely non-linear, any attempt to use it over a wider concentration range would require a respective linearization function in the signal processing, or preferably an empirical multi level calibration.

I don't exactly understand what the "calibrated unit" in your calibration scheme is, but calibration by comparing two MQ135 sensors will further reduce the rather poor accuracy.

I would suggest to use a primary standard for calibration, a gas or vapour mixture of known concentration. You'll need a test gas for calibration anyway.

 

[samic45mit1]

My aim to compare two MQ135 is to reduce number of calibrations require in particular time interval and also to reduce
maintenance cost of instrument. You told about the "linearization function in the signal processing". So please also tell
about this.

Thank you

 

[FvM]

I already told my opinion about using MQ135 as calibration reference.

Can you give application details like measured substance, concentration range, expected accuracy?

For a linearization function, you can refer to the datasheet curves that show sensor conductance versus concentration as a straight line in log-log scale. This corresponds to a function Y = X^a which can be modelled in software.

 

[samic45mit1]

My project is on WSN based pollution monitoring and I want to calibrate MQ135 without any connection using micro controller and WSN. In this first I want to find error by placing both unit on one place and then send the correction factor to other unit using WSN . I measuring CO by MQ135. Average accuracy is sufficient for my project.

Actually I work on analytic instrument like pH meter which use two point calibration to correct error. Same thing I want to develop using WSN for MQ135

Thanks

 

[FvM]

Reviewing the MQ-135 datasheet again, I see a problem with the rather high CO2 sensitivity. Ambient level of CO2 is about 400 ppm, considerable higher in low ventilated rooms. There are other sensors with more specific response to CO, e.g. MQ-7.

 

[samic45mit1]

Thanks

Please tell , my approach for calibration using wsn (as I told above) is correct?

:???:

 

[FvM]

A SnO2 semiconductor gas sensor characteristic isn't like a pH electrode that has only small variations between sensor exemplars. It has large sensitivity variations, e.g. R0 (sensor resistance for a reference concentration) is said in the datasheet to vary by a factor of about 7. It's not clear how much the slope varies.

As far as I understand, you want to calibrate the sensors with an arbitrary gas concentration (may be zero) and expect sufficient measurement accuracy of the calibrated sensor. But it's quite difficult to estimate the real calibration error.

On the other hand, if you assume that the sensor characteristic in the datasheet applies to all sensor exmplars, the calibration can be reduced to determination of the individual R0 value. But this is more or less a guess and should be verified by testing a number of sensors with different gas concentrations.

 

[samic45mit1]

Thanks for replays and giving your time.

 

[samic45mit1]

I am using MQ135 and MQ07 for the air quality monitoring. Please tell it is possible to use Automatic Baseline Calibration algo for these sensor.