OK, i have a k-probe connected to a MAX31855 breakout and this is connected to a ESP8266
i have written the code, tested the code, and let it run for days on the bench... all great!
but now i tried to mount it on the outside of the fireplace - a cast iron box.
and the readings are bizarre! mostly negative, but all wrong...
I put the k-probe tip in free air, it is fine, touch the fire box, wrong readings!
i think the thermocouple is a non-grounded type by measuring the continuity between the wires and the sheath, and get nothing.
any advice?
is there a thermocouple that would work for this situation - or is the MAX31855 the problem?
Wareemba:
but now i tried to mount it on the outside of the fireplace - a cast iron box.
...
I put the k-probe tip in free air, it is fine, touch the fire box, wrong readings!
You have stated the thermocouple is touching the cast iron fire box, which is most likely grounded.
Remember that a thermocouple works by the dissimilar metals creating a voltage difference when heated. That voltage difference is amplified and linearized by the MAX31855.
There are two way the thermocouple can be wired to measure the temperature difference. Either true differential or single ended (where the low side of the thermocouple is connected to ground). I don't know which you have here. Either way, it is likely the high side is being connected (shorted) to ground by the electrical connection to the cast iron fire box.
What happens if you stick the thermocouple through the door and measure the internal fire box air temperature?
Hi, thanks for your reply and good thermocouple explanation!
but i don't think the cast iron firebox is grounded? it is non-electrical...
however, i did an experiment and removed the stainless steel sheath from the thermocouple so it is now just two wires covered with fiberglass sheaths.
the probe is now inside the fire box and reading fairly reliable, so for some reason the metal braided sheath was causing the issue?
problem now is how long the exposed wires will last in such a harsh environment! im going to try and find some better fiberglass type sheathing i guess...
Does the firebox touch the ground or come in to contact with ANY electrical component? If so, then it is grounded.
If you assume a single sided reading on your thermocouple, the low side is connected to power ground. For safety the power ground is likely connected to earth ground at the power supply. The firebox touches earth ground. Therefore when the thermocouple high side touches the firebox, it is now touching and connected to ground. Do you get the same bad result as you list in the OP if you take a piece of copper wire and connect thermocouple + and - (you don't need to remove the thermocouple to test)?
Using thermocouples to for direct contact measurement is tricky as you have just encountered.
well, the fire box sits on a "hearth" made of slate and has no electrical connections.
it has a flue that passes through the roof, but that has nothing electrical in it either...
when i was observing the error reads, the steel braided sheath was probably touching the fire box and so was the probe tip.
my multimeter showed no continuity between any of these things, but i removed the braided sheath regardless... and it seems to have solved the issue to a large extent... we'll have to see how it goes after a few more nights of fire...
if you short the thermocouple it gives an error read, (but i didn't try it with a piece of copper wire)
Just it being a huge chunk of metal, especially in comparison with a tiny thermocouple involving minute electrical charges, is more than enough for your firebox to constitute "ground". Just don't touch it with your thermocouple.
Most likely the thermocouple and firebox are picking up AC signals by capacitive coupling with the house wiring. Try electrically isolating the thermocouple from the firebox.
Or, as mentioned above, it could be a grounding issue, especially if you are collecting data using a laptop powered by house wiring. If the latter, disconnect the laptop from the house wiring and operate by battery.