Table of Contents
Advertisement
This can be represented mathematically as follows:-
CALC_CTG = CTGi + (CTGi+1 – CTGi) × (T – CTi) / (CTi+1 – CTi)
The actual gain value used is 1 + CALC_CTGO x 10-6 and is multiplied by the uncompensated value MVV.
The offset correction is then applied. Using the same temperature index "i" as found for the GAIN index above.
The Offset value is extrapolated between the same two temperature points.
This can be represented mathematically as follows:-
) × (T – CT
CALC_CTO = CTO
+ (CTO
– CTO
) / (CT
– CT
)
i
i+1
i
i
i+1
i
-4
The actual offset value used is CALC_CTO x 10
and is subtracted from the above gain adjusted value.
The output from the temperature compensation (CMVV) is then calculated as
-6
-4
CMVV = MVV x (1 + CALC_CTGO x 10
) - (CALC_CTO x 10
)
The Temperature Measurement
The temperature sensor used is a Dallas (MAXIM) DS18S20 Digital Thermometer using the "1-Wire" bus technology.
This gives a temperature measurement accuracy of +/-0.5 Degree C over the temperature range –10 to +85 degree C
and +/-2.0 Degree C over temp range –55 to +125 Degree C. The resolution of the measurement is 0.0625 Degree C.
The temperature is sampled and the TEMP variable updated every 5 seconds.
How to Set Up a Temperature Compensation
There are a number of ways of obtaining a temperature compensation curve.
The best possible compensation for a given piece of physical hardware can only be achieved by performing
experiments on that particular unit (DCell/DSC and associated strain gauges), to characterise the measurement
output at a variety of different, stable temperatures in the required operating range.
The basic choice of methods depends on trading off ideal accuracy against the complexity of the calibration
procedure.
Method 1
Apply a simple linear drift correction (i.e. for known constant gain and offset changes per degree), by specifying
zero correction at the calibration temperature, and appropriately adjusted correction values at extreme
temperatures above and below this.
This can be used when the measurement or sensor has known temperature coefficients.
Method 2
Where the temperature characteristics of the measurement are known, but not linear, a similar scheme to Method 1
can be used, with a multi-point table defining an approximation to the known, ideal temperature curves of offset
and gain variations.
NOTE: Both of the above methods are based on 'known' characteristics, which could come from datasheets but
these methods would not compensate for the DCell/DSC.
Method 3
Do a series of measurements at different temperatures and install the appropriate correction values to give exactly
correct results at those same temperatures –i.e. calculate ideal gain and offset corrections at the tested
temperatures.
(This method is the most common). There is a 'wizard' available in Instrument Explorer which will enable this
method to be easily completed by calculating the gain and offset corrections for you.
Method 4
Use a set of test results to plan a 'best correction' curve (not necessarily perfect at test temperatures, but slightly
better overall).
39
Micron Meters
DCell & DSC User Manual
Advertisement
Table of Contents
