Are Type K Thermocouples Being Used in Error? (Steven Heale)

The following post has been submitted by Stephen Heale

Some time back – I read that type N thermocouples were used in a brick manufacturing business. They have many positive and stable properties with a good output voltage.
This is an excerpt from the information that I submitted to MFMP back in 2012, comparing type K to type N.

Type K thermocouple:-

Sensitivity 41 µV/ °C, Range -200 °C to +1250 °C.
Causes a deviation in output when the material reaches it’s Curie point (magnetic state change), at 350°C.
Prone to a gradual and cumulative drift in thermal EMF on long exposure at elevated temperatures due to a compositional change caused by oxidation, carburization or neutron irradiation that can produce transmutation in nuclear reactor environments. Manganese and aluminium elements from the negative wire migrate to the positive wire resulting in a down-scale drift due to chemical contamination. This effect is cumulative and irreversible.
Experiences short-term cyclic change (hysteresis) in the temperature range 250 to 650 °C.
Produces random perturbation in thermal EMF between 25 and 225 °C due to magnetic transformations.

Type N thermocouple:-

Sensitivity 39 µV/ °C, Range -270 °C to +1300 °C.
The Curie point is below 0 °C.
Resistant to oxidation with enhanced thermoelectric stability relative to other standard base-metal thermocouple alloys.
Type ‘N’ thermocouples are not merely an improvement but should be considered as the optimum – overcoming all the disadvantages of not only type ‘K’ but any nickel based thermocouple.

A paper presented by Barbara Hudson in relation to the use of type K thermocouples in the clay-brick manufacturing business concluded that they should change to type N.
They were concerned with the problem which occurs between 500 °F to 1020 °F (260 °C to 549 °C) due to short-range ordering, where an erroneous EMF is produced. The paper goes on to discuss the problem caused by hysteresis and cumulative drift when cycled at higher temperatures – which is an area we would expect this research to achieve with further work. (this was in 2012 before LENR was capable of delivering temperatures above 600 °C – hence my comment).

Their ultimate solution was to select a type N thermocouple – (developed by Noel Burley at the Defence Science and Technology Organisation in Australia).

Stephen Heale

  • BillH

    If both types can resolve a temperature to within 0.25 of a Deg C then that should be accurate enough, I’d go with the cheapest. It all depends how accurate you need to be?

    • Alan Smith

      Hi Bill. Even the inexpensive Chinese K-types are accurate to 2-3C at 1000C. Sometimes better than that. Which if you think about it is around 0.2-0.3 %. Good enough for most purposes.

      • Matt Werner

        According to their spec sheet that is true. Temperature cycle them a couple times and then verify their accuracy. Not so good then.

        • Alan Smith

          Calibration is a fetish of mine.A lot and often – this kind of science requires it. The K-types hold up pretty well inside a ceramic thermocouple protection tube. ETF at least 20-30 hours at 800C+.

          But naked at those kind of temperatures- not so good.

          • Steve H

            Hi Alan.
            It begs the question – what is Rossi using?
            He can’t be constantly changing thermocouples otherwise it would not be a commercial product.

          • Alan Smith

            Type ‘S’ are the best- Platinum/Rhodium ones. It may be that he is using them outside the core, of course so that they are running 1-200 degrees cooler. Another possibility is to use a sturdy wire and rather than measuring the current generated by the bi-metal junction as in a thermocouple measure the change in resistance of the wire with changing temperature. From a distant memory I think Tungsten wire may be good for that. Must check.

          • Steve H

            Would that be RTD method?

          • Alan Smith
          • Steve H

            Thanks for the link

          • Bob Cook

            I think that niobium-iridium are better yet. They are good for very high temperatures. However, Ir is not cheap, but it should last a long time.

          • Alan Smith

            Thanks for the tip, Bob, will check them out.

  • Alan Smith

    Very good article and excellent points. The main problem with ‘N’ type is (for most of us) one of cost. A good quality ‘K type thermocouple is around 1/4 the cost of an N-type. So for the moment at least, I am happy to use K-type thermocouples, make plenty of provision for error-checking and throw them away as soon as the become suspect.

    • Matt Werner

      If you are throwing them away four times as often due to failure are they really a bargain?
      Even in IC engine testing I was always concerned about the reliability of K type thermocouples. Most DAQ hardware today is capable of operating with lowered sensitivity devices just as well.
      Anything close to the exhaust were not K types.

  • Andreas Moraitis

    I wonder how different types of thermocouples behave when they get in contact with hydrogen. This is likely to happen in various LENR experiments. Unfortunately, I could not find any literature about this issue.

    • Alan Smith

      @ Andreas. Look up ‘hydrogen embrittlement of K-type and platinum thermocouples’ and you should find some papers.

  • Wishful Thinking Energy

    This is good advice. I switched over to N type thermocouples about a year ago and have found them to be far more stable and long lasting at high temperatures. They are more expensive, but if you check eBay frequently you can find deals.

  • Obvious

    Type K is good enough if kept below 1300°C and oxidized. N is better, no question, but often not needed.

    The biggest problem is that people are often using diameters that are way too small for the heat range, in reducing atmospheres (lots of alumina around helps), running them close to or over their melting points, or putting them into materials that conduct, make electrolytes or have galvanic activity.
    No thermocouple is good for those conditions.

  • Asterix

    Long ago, in this forum, I noted that type K was not a good choice. I preferred type S.

  • bachcole

    So, what does all this mean?

    • Steve H

      It just means that most replicators are using cheaper, throw away – temperature measuring devices. And they are aware of the small errors emanating from type K thermocouples.
      No worries.

  • Alan Smith

    Thanks for the tip, Bob, will check them out.