Sensed T/C Temp vs. Actual Flow Temp

If the thermocouple is in thermal equilibrium with its surroundings and an independent - calibrated - temperature sensor reading differs by 800°C, you're doing something very wrong.

Has your expected temperature been checked?

Even inside a flame there are large thermal gradients. Downstream from a flame one would expect to find mich lower temperatures because heat is being lost to the environment. In addition, a thermocouple doesn't indicate the temperature of the surrounding gas; it indicates its own temperature. Important distinction becauase the thermal mass of the thermo0cuple introduces a measuring time constant. After three time constants, that problem goes away and you must examine the steady state heat loss to the environment. Conventional wisdom is that 10 thermocouple diameters must separate the measurement location from ambient temperature. More, however, is better. In addition, you may want to determine whether the flow of the gas you are measuring is turbulent or laminar.The moral of the story is always use the smallest thermocoupe and the longest leads you can get away with.

If you are trying to measure a flue gas temperature in excess of about 400ºC with a bare thermocouple you will always get a low value. The thermal system you are trying to measure is complex and not in a steady state. The hot gas is transferring heat to its surroundings (the lining of the hot gas flue) The lining of the hot gas flue is loosing heat to the outside world so the hot face temperature of the flue must be less than the hot gas temperature. When you immerse a thermocouple in this system it gains heat and therefore temperature until it reaches equilibrium with its surroundings Equilibrium in this case does not mean everything being at the same temperature but everything being at appropriate temperatures for the heat transfer between the gas and the thermocouple as one pair and the thermocouple and the surface of the flue lining as a second pair and the heat transfer through the lining to the outside world as a third pair all being equal. As the heat transfer to the thermocouple and the hot gas flue lining is governed largely by convection this is proportional to to the temperature difference raised to the power 1.33 or there abouts. However the heat transfer between the thermocouple and the flue gas wall is proportional to the absolute temperature difference raised to the 4 power and this has a much stronger influence.

The usual tool to use for measuring hot gas temperatures in cool surroundings like the reversal chamber of a steam boiler is a suction pyrometer. This consists of multiple concentric radiation shields surrounding a bare bead thermocouple. A small pump draws the hot gasses over the thermocouple and on to any flue gas analysis equipment. These instruments are supplied with calibration curves. Calibration is not a DIY operation.

Have a look at this link http://www.chec.kt.dtu.dk/Research/Experimental Facilities within CHEC/Suction Pyrometer.aspx

I hope this helps.

Gasman is correct. However, if you want to minimize thermocouple error, try a bare (usheathed) small thermocouple. These small themocouples (~50 micron diameter) will minimize radiation and conduction errors. The down side is that they are fragile. However, they are manageable for test purposes. You may find that accuracy is sufficient for your application. If it isn't, Google "suction pyrometer". There are lots of good references.