Anyone ever test NTC vs K type?

[mgenius33]

I was looking at buying a stand alone NTC thermistor temperature probe. Anyone use these, and have you calibrated them against the k type to see which is more accurate?
From reading the NTC appears to be +/- 1.0*C, and the k type +/- 0.5*C
However the NTC responds more quickly.
I don't know for certain, as I've never tested them head to head.
Can anyone share their results?

[mark beiser]

I have K-type thermocouple equipment, mostly Fluke, some NTC thermistor stuff, and platinum based PT100 thermistor instruments.

The PT100 thermistor instruments are the fastest response and have the best accuracy rating, but both the instrument and probes are expensive. The probe replacement cost is quite high.

NTC thermisors are very accurate, but getting precise readings of pipe temperatures can be a chore with some types of probes because of the length of the sensor, and the need to have the entire length of it in firm contact with the pipe.
Response speed can be quite good, or quite slow, depending on the size of the sensor and the probe design.

Type K thermocouples have a much higher +/- accuracy range than thermistors, and readings can be subject to interference from stray voltages, and by the temperature of the instrument.
Some thermocouple reading instruments, like the Fluke 52II, will filter out the interference if it is in the 60hz range, and have better temperature correction than others, but are generally considered inferior.
Cheap thermocouple thermometers and probes can be wildly inaccurate.

I've tested 7 different instruments, from 5 different manufacturers, using 3 different technologies, using 8 different pipe temperature probe designs against each other at the same time under the same conditions.

With all instruments, probes, and the pipe at the same temperature, there is only a 0.6ºF difference between the lowest and the highest reading.
With 100ºF ambient temperature, and a 38º pipe, there was less than 1.5ºF difference between the lowest and the highest reading.
To achieve readings that close, all of the NTC thermistor probes took a long time to get to the final temperature reading, and some of them required additional insulation around the probe.
The PT100 thermistors were basically instantly to the temperature of the pipe.
Both of my Fluke 52II thermometers, with 80PK-8 type K pipe clamps, were <0.5ºF off from the PT100 thermistors.
Another brand of thermocouple type thermometer and pipe clamp didn't do quite as well as the Fluke.

While I realize there are potential issues, in real world use, I've found that a Fluke 50 series thermometer, with Fluke pipe clamp thermocouples, gives the overall best combination of fast, accurate and precise pipe temperatures for the money. The Fluke 80PK-8 and 80PK-10 pipe clamp thermocouples use a spring loaded ribbon type thermocouple design that is very forgiving of less than ideal mounting locations.

I realize that a Fluke 52II and a couple of 80PK-8 pipe clamp probes is NOT inexpensive, being in the $550 range, but the only other instruments I've used that matches or beats it for real world speed and precision for pipe temperatures was >$800 5 years ago for the instrument and 2 PT100 probes. ;)

[itsiceman]

Mark I know the RTD your referencing well. It really is too bad it is no longer made. With a few simple modifications it would have been a more robust product. FWIW If you ever had the need to replace them the newer versions are slower to react than the old version :.02:

m With accuracy of the bare sensor in every style they come in ranges of accuracy so its not cut and dry which type is more accurate. Some include inaccuracy of something like a clamp they are mounted if that is the style your looking at and is figured into the total number which also may include the inaccuracy of the meter as well.
IMO the sensor material itself has very little to do with inaccuracies of a measurement as we need to take them. Most of the inaccuracies will be generated external from the sensor material type.

[mark beiser]

Most of the inaccuracies will be generated external from the sensor material type.

Yeah, you can have the most "accurate" sensor in the world, but if the probe it is part of makes it difficult to get mounted on the pipe securely, or transfers ambient heat to the sensor, the readings will not be "precise", rendering the sensors raw accuracy irrelevant.