K Type vs Pt resistance probe. A question

[aamcle]

I have just set up a Pid with a K Type its controlling to 1C which seems close enough to me.
I have read that the platinum probes are "better" are they?
Would using one improve my control?
In what way are they better?

Thanks. Aamcle

 

[Glot]

The main inaccuracies are in the electronics. A Thermocouple generates voltage dependent on the temperature it is exposed to. These can handle higher temperatures and are cheaper to make. A PTC or NTC ( positive temperature co efficient or negative Temperature co efficient) sensor is a thermistor. It varies resistance with temperature. PTC resistance goes higher with higher temp and NTC the reverse.
For what you are trying to measure and control, either thermistor or either type of thermocouple will suffice. As I said, it is the meter itself that determines the accuracy in this context.

 

[Glot]

Just re reading. I should have added, a copper NTC is good for up to about 150 deg C. Platinum is more accurate but do you really need to measure to an accuracy of 0.001 deg C? Most things you read about sensor accuracy will probably be in a scientific context.

 

[aamcle]

Thanks Gents.


Aamcle

 

[schrodinger]

Hi Aamcle,

It's helpful to dintinguish precision (repeatability, or the smallest meaningful increment possible from a sensor) and accuracy (proximity of mean sensor output to the true value).

Thermocouples can in practice be extremely precise and accurate; as precise as PRTDs, in fact, if not more so. Their main limitation is that they measure temp at the measuring junction by reference to temp at a reference junction. ( A thermocouple is really a circuit with two wires, each of a different metal; if the 2 junctions of these diff metals are at diff temps, a voltage will be generated. In reality t/c circuits get far more complex but this is the principle.) So, your precision and accuracy are ultimately limited by whatever device (usually a cheap thermistor) is measuring temp at the reference jctn, and by the precision of whatever voltmeter is measuring the t/c voltage.

Any temp sensor can be calibrated to damn good accuracy by immersion in freezing or boiling water, which have known temps (accounting for altitude), and noting the offset; in general you'll find that both t/cs and PRTDs only drift noticeably in the zero, not the span. In other words, if your t/c reads 1.5 deg C in pure ice water at sea level, then you're pretty safe just subtracting 1.5 from the reading to get true temp. To do even better you can compare readings in ice water and boiling and get both a scale and offset correction. Even betterer, make measurements across a range of temps using a thermometer you know is accurate, fit a polynomial to the data and off you go.

In my field we use thermocouples to measure temp diffs on the order of 0.001-0.002 deg C. The limiting factor is your reference junction temp, not the thermocouple. If you want both accuracy and precision, then, your best bet is to make a simple t/c circuit with the PRTD (which is usually quite accurate) at the ref junction, and convert the t/c voltage to temp using the appropriate polynomial (easy to find with Google). For 0.1 deg C precision this would require around 0.005 mV precision in your voltmeter.

Is your thermocouple plugged into a standard multimeter, or something more advanced like a datalogger of some sort?

 

[Glot]

If you want to get close to zero , add a heap of ice to the water then when cold, add a heap of table salt. This will draw the temp down close to zero. Boiling water is not 100 deg. The steam directly above the water is. Not the water vapour you see but the invisible bit. You can't see steam. It is a gas. You can heat water far above 100 deg at ambient pressure in a microwave oven without any boiling effect.

 

[nu_brew]

Be careful heating water in a microwave http://www.animations.physics.unsw.edu.au/jw/superheating.htm

 

[OzPaleAle]

A the old exploding instant coffee trick

http://www.youtube.com/watch?v=LpDs7Xm1uLo



EDIT: ahh Nu_Brew beat me to it

Says do not try this at home, first thing I did when I saw it was tried it at home.....

 

[schrodinger]

If you want to get close to zero , add a heap of ice to the water then when cold, add a heap of table salt. This will draw the temp down close to zero. Boiling water is not 100 deg. The steam directly above the water is. Not the water vapour you see but the invisible bit. You can't see steam. It is a gas. You can heat water far above 100 deg at ambient pressure in a microwave oven without any boiling effect.

Why do you say boiling water isn't at the boiling temperature? The mixture of liquid and gaseous water is all at the same temperature during boiling.

True that you can superheat water in a microwave (or by many other means), but then it's in a metastable state -- add anything to nucleate the formation of gas bubbles, and it will drop to 100 deg (and probably 'boil over' pretty quickly!).