EGT for mixture setting

[wmcgonegal at rr.etc.ncr.doe.ca]

This message is intended for those who are interested in
thermocouple information.

>I have seen this done using a thermos full of iced water. The
>trouble is   the ice melts.

>Is it practical to, put the cold junction into a lot of
>insulation, eg shove it into a giant block of polystyrene foam ,
>or in a thermos bottle(?)

>How about placing the cold end of the thermistor in the coolant
>stream,..... It won't give extremely accurate results,..... if you 
are
>just looking for peak EGT, this should work.

Every connection in the TC (thermocouple) system between dissimilar 
metals acts as a TC.  Cold junction compensation is used
to offset the effects of joining your TC to a measuring device
(which has different metals in it).  Using a cold junction allows
you to read the voltage directly and get the temperatures from a TC
reference table.  0C (zero Celsius) is usually used as the
reference junction temperature, and in TC reference tables the
TC voltage at 0C is 0 volts.

For junction compensation of a TC you do not need an actual cold
junction at 0C.  You need to measure the temperature at the point
where the TC probe is attached to your conditioning circuit.  Both
of the TC wires must be at the same temperature at the junction
where they join into the conditioning circuit.  If you know the
temperature of the junction and you have a TC voltage/temperature
look up table you can properly correct your TC voltage reading.

To correct for the junction, look up what the voltage for your TC
type is for the junction temperature.  Add this voltage to the
voltage reading from the TC leads at the junction.  Then look up
the final voltage in the TC reference table to get the actual TC
temperature.  Of course it is easiest to get a computer to do
all this looking up and calculating.  If your junction is always 
at the same temperature (like room temperature) you could always 
use the same offset voltage (depending on what accuracy you want).

A circuit which adds in the proper compensation voltage based on
the junction temperature can also be used.  Using an ice bath and
a second TC at 0C properly connected in with the first TC does the
voltage addition for you if you prefer not to have a compact, easy
to use, and portable system.

If you are just looking for temperature peaks, take your two leads
coming from your TC and attach a digital voltmeter in the
millivolt range.  When you get your highest meter reading you have
reached your highest temperature.

Curves can be substituted for the tables.  Tables and curves are
available from TC sources such as Omega Engineering.  There is a
very good description of TC principles in the Omega Engineering
temperature catalog.

Some suitable semiconductor devices for measuring the junction 
temperature are:
AD590 - Analog Devices part that outputs 1 uA per degree Kelvin
(very easy to use, require a resistor and voltage supply to get a
voltage proportional to temperature).  This device is good to
150C, so it can be used to measure other things as well.
LM335 - National Semiconductor part similar to AD590

>If you accept more error, extension grade T-Couple wire is
cheaper.

Extension grade has similar characteristics to TC grade over a
limited temperature range.  The error outside this range may be
quite large.

Will McGonegal
Mobile Sources Emissions Division
Environment Canada

wmcgonegal at rr.etc.ncr.doe.ca