1
\$\begingroup\$

I'm trying to heat up a steel wire of d=0.05mm with a specific current. The length of the wire is about 5cm. By measuring the voltage over the wire, how can I find the temperature?

\$\endgroup\$
10
  • \$\begingroup\$ Use a temperature sensor is my opinion else justify why you can't. Being a cheapskate can sometimes be a justification but not always. \$\endgroup\$
    – Andy aka
    Oct 24, 2016 at 12:19
  • \$\begingroup\$ @Andy aka: How do you want to use a sensor to measure temperature of a wire having only 50µm diameter? \$\endgroup\$
    – Curd
    Oct 24, 2016 at 12:26
  • \$\begingroup\$ Which temperatures are we talking about? \$\endgroup\$
    – Janka
    Oct 24, 2016 at 12:28
  • \$\begingroup\$ @Curd maybe an optical pyrometer. \$\endgroup\$
    – Andy aka
    Oct 24, 2016 at 12:31
  • \$\begingroup\$ About 500 deg celsius. \$\endgroup\$ Oct 24, 2016 at 12:47

3 Answers 3

2
\$\begingroup\$

I think the only problem is to find the resistance-temperature dependency of the wire.

I wouldn't rely on published coefficients or tables unless you are very sure they refer to exactly the same steel alloy of your wire (I can imagine that you don't know it exactly).

I'd measure the resistance-temperature relationship by myself as follows (see picture):

Build a device using a hot air gun to create an air stream (red) of known temperature.
You can measure the temperature of the air stream by a conventional electrical sensor.
The temperature of the air can be controlled by the power feed to the hot air gun.
You probably have to take some care that the air is mixed enough after heating so temperature is homogeneous.

Use the air to heat a sample of the same type of wire (brown; maybe of larger length to get a more accurate result) and measure the resistance of that wire in the interesting temperature range.
Of course the resistance of the leads connected to the sample wire must be negligible (e.g. by using thick copper wire) and the connection must be heat resistant at the interesting temperatures and resistant against oxidizing; at least for short time. I.e. you can't use soldering; I suggest crimping or welding.

enter image description here

Once you have determined the resistance at several different temperatures in the interesting range you can fit coefficients of an appropriate model function (the most simple model would be a linear regression with only two parameters: slope and intercept) and use it backwards (inverse function) to determine the temperature when resistance is given (measured).

\$\endgroup\$
1
  • \$\begingroup\$ Perhaps it's obvious, but given the way that you have the sample wire coiled, it will have non-negligible inductance. DC measurements are safe, but something like a pulse-driven ADC will get complicated. \$\endgroup\$
    – Reinderien
    Nov 23, 2020 at 22:42
1
\$\begingroup\$

The wire will have a certain resistance and that resistance will change with temperature to a degree, and with time as the wire corrodes and oxidized.

If you know the relationship between temperature and resistance you can get an idea of the average temperature by measuring the current with a constant voltage, or by measuring the voltage at a constant current. If you measure voltage at points along the wire you could get an idea of the temperature distribution.

You can look up temperature coefficients of various steel alloys and you can measure the wire you have at a known temperature (preferably close to the operating point) or you could calibrate the wire in an oven over the expected operating range, depending on what kind of accuracy you expect.

Unless the wire is operating at a low temperature and/or a vacuum it won't be very stable because of oxidation. This works much better with a platinum wire or a tungsten wire.

\$\endgroup\$
2
  • \$\begingroup\$ How do I find the relationship between R and temperature? \$\endgroup\$ Oct 24, 2016 at 12:48
  • 1
    \$\begingroup\$ I gave you several ways above- look up the tempco of the steel alloy you are using (generic, could be wildly inaccurate because it changes with processing etc.), single point measurement plus the value you looked up as a deviation, and calibrate the whole thing in an oven (preferably with Kelvin connections). \$\endgroup\$ Oct 24, 2016 at 12:51
0
\$\begingroup\$

How does a blacksmith know the colour of a piece he is working on? Most blacksmiths will cite experience but it's based on Planck's law and black body radiation.

Here's a simple chart you can use but you need to match any chart with the emissivity of the material being "measured". This is for steel but you should be able to find one for iron (it's probably very similar): -

enter image description here

And here's some more information on black-body radiation: -

enter image description here

So, in principle you can get a fair degree of accuracy without resorting to trying to calculate the temperature based on the voltage across the wire (given the many problems this method has). In other words, use your eyes like a smithy does!

Or, use a pyrometer: -

enter image description here
(source: fluke.com)

\$\endgroup\$
3
  • 1
    \$\begingroup\$ I think you'll find the colours on the chart under the 800 degree mark are the tempering colours for steel (surface colour change) with those above being radiant. Nice chart though. \$\endgroup\$ Oct 24, 2016 at 13:01
  • \$\begingroup\$ @JImDearden yes they won't be visible! \$\endgroup\$
    – Andy aka
    Oct 24, 2016 at 13:09
  • 1
    \$\begingroup\$ I think such an optical thermometer it's absolutely useless for measuring the temperature of a 50µm (!) thin wire. \$\endgroup\$
    – Curd
    Oct 24, 2016 at 13:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.