How precise can measurements of temperature get?

  • 2 Replies

0 Members and 1 Guest are viewing this topic.


Offline thedoc

  • Forum Admin
  • Administrator
  • Sr. Member
  • *****
  • 513
    • View Profile
David Skytte asked the Naked Scientists:
Great show. I've been an avid listener for several years and I have a question for you. How precise can measurements of temperature get (like in decimals places)? And does this precision vary depending on where you are on the scale?

What do you think?
« Last Edit: 08/08/2015 11:50:01 by _system »


Offline evan_au

  • Neilep Level Member
  • ******
  • 4316
    • View Profile
Re: How precise can measurements of temperature get?
« Reply #1 on: 08/08/2015 12:49:39 »
The lowest temperature recorded are measured in picoKelvins; this claim is for 10-7K. However, this is an order-of-magnitude measurement, which may only be accurate to 10%.

For measuring temperatures near room temperature (around 300K), a measurement accurate to picoKelvins would require an accuracy of 1 part in 109, which would require an extremely accurate measurement!

Since the temperatures we measure are derived from atoms bumping into the thermometer, and every atom has a different kinetic energy, the temperature will vary dramatically on small timescales. So we must talk about average temperature being derived from the energy of many atoms, measured over a long period of time.

The US National Institute of Standards and Technology (NIST) performs precision measurements of many parameters, including temperature, using a variety of methods to cover different temperature ranges. They claim precision of 28uK/K (ie 2.8x10-5) for some methods traceable back to quantum effects in the Josephson Junction.

PS: I've probably confused the terms accuracy & precision - if someone could clarify how this applies to measuring temperature?


Offline alancalverd

  • Global Moderator
  • Neilep Level Member
  • *****
  • 4911
  • life is too short to drink instant coffee
    • View Profile
Re: How precise can measurements of temperature get?
« Reply #2 on: 09/08/2015 00:19:58 »
I have built quite simple systems for measuring temperature changes of less than 1 microkelvin in the ambient region, but they were not referenced to any absolute value.
helping to stem the tide of ignorance