I came across something I find odd while hacking a space heater into a high power dummy load: my cheap space heater is rated 1.5 kW 110 VAC, so I expected to see less than 8 ohms which, when heating, would increase to about 8 ohms in steady-state with the fan ON.

Indeed, the resistance of copper increases with temperature - even if the resisting element is not made of copper, I assume most metals (except NTC) do the same in varying degrees.

I'm not familiar with US electrical supply, so I just assumed here the increased initial current is handled - otherwise there might have been some inrush control built-in.

After tearing it apart, which revealed the heating element was as expected directly connected to the supply, I measured the resistance at 50 ohms. Puzzled, I applied 30 V DC across the prongs and the current increased from 0.6 A to more than 3 A in a matter of minutes. Without suprise, the resistance had gone down with heating to about 10 ohms. The fan wasn't on, it's not an universal motor but an AC machine.

Am I forgetting something fundamental or isn't it a bit odd that the resistance decreases with temperature like NTC resistors? If not, what kind of material does that?

  • 2
    \$\begingroup\$ NTC devices are not metals; all metals (at least as far as I know? I can't think of any physical way a metal could have a negative TC) have a positive temperature coefficient, NTC thermistors are semiconductors. \$\endgroup\$
    – Hearth
    Commented Jan 19, 2022 at 21:05
  • \$\begingroup\$ Can you tell what the heating elements are made from, or, if not, post a picture? If they look like coils of metal, they're probably nickel-chromium restance wire (basically pretty close to stainless steel, just a different mix). If they look like solid sticks, they may be ceramic, and then @SpheroPefhany's post is spot-on. \$\endgroup\$
    – TimWescott
    Commented Jan 19, 2022 at 21:22
  • \$\begingroup\$ In the electrical engineering laboratory, it is customary to use carbon filament lamps ... whose resistance decreases very slightly with temperature ... \$\endgroup\$
    – Antonio51
    Commented Jan 19, 2022 at 22:08
  • \$\begingroup\$ something as this ? french.alibaba.com/product-detail/… \$\endgroup\$
    – Antonio51
    Commented Jan 19, 2022 at 22:20
  • \$\begingroup\$ Thanks for your comments. The heater's insides were similar to this, so more like a grid: allaboutcircuits.com/uploads/articles/… \$\endgroup\$ Commented Jan 19, 2022 at 22:58

3 Answers 3


There are few types of heating elements used in space heaters, all with their peculiarities.

  1. The traditional Kantal wire, formed as a spiral/serpentine and supported by mica paper, ceramic body or quartz tube in its place.

It has more or less constant resistance between room temperature and working temperture (600 ... 1000K)

Modern carbon mesh heaters are similar in electrical behavior, but rather expensive.

  1. The traditional tungsten wire in a quartz tubing. These are in essence low-efficiency incandescent lamps.

When cold, they have 3-5 times lower resistance. Their resistance is more or lesss proportional to the temperature, just like in most pure metals. When powered on, the wire heats up quickly so they don't trip the breaker by their inrush current.

  1. "Ceramic" heaters. They look like the one in your comment, a ceramic blob (in general more than one) pressed between pieces of aluminium radiator / conductor.

These are rather complex.

When cold, they have 3x-5x higher resistance.

When powered on, they heat up to like 400K and their resistance gets at minimum (rated power).

When cooling is restricted and the temperature goes up, their resistance sharply increases. This supposedly makes them safer. This also allows some power regulation by adjusting only the fan speed.

I don't really know how the ceramic heating elements implement this complex behavior, but one can imagine a mixture of NTC and PTC substances and/or a phase change that makes the substance non-conductive.

In essence, these ceramic heaters are the worst choice in regard to making a constant resistance dump load out of them.

Get a kanthal-based, convection-driven space heater instead. You will get more or less stable resistance without the need to control a fan.


Silicon carbide and similar IR heating elements are semiconductors and have a negative temperature coefficient. Image from here

enter image description here

  • 14
    \$\begingroup\$ To add a bit of color to Spehro Pefhany's response, the density of free charge carriers in a semiconductor is proportional to its temperature. Higher temperatures "dislodge" more carriers from crystalline bonds and make them available to contribute to current flow. \$\endgroup\$
    – vir
    Commented Jan 19, 2022 at 20:06
  • \$\begingroup\$ Interesting, thanks. So you think this kind of heating element is made of this semiconductor material? allaboutcircuits.com/uploads/articles/… \$\endgroup\$ Commented Jan 19, 2022 at 22:59
  • \$\begingroup\$ TBH, I don't know how those are constructed internally. \$\endgroup\$ Commented Jan 20, 2022 at 3:24
  • 2
    \$\begingroup\$ @vir the density of free charge carriers in a semiconductor increases WAY faster than linear. It is more of an exponential. \$\endgroup\$
    – fraxinus
    Commented Jan 20, 2022 at 10:15
  • 1
    \$\begingroup\$ @uhoh : Yes. That's what in metals results in the generally increasing resistivity the OP observed. In this case, you have those two effects competing with each other, which creates the minimum in the graph shown. \$\endgroup\$ Commented Jan 20, 2022 at 13:41

PTC materials are prone to spot-overheating. Part of the wire gets hot, so the resistance increases, so more power is lost there, so it gets hotter, and the rest of the wire gets colder, and ... the wire burns out.

For a resistor, you can partly handle this by giving the resistor good thermal conductance, so that it heats up evenly. For a long thin bit of wire suspended in air or in a vacuum, that's not possible.

So space heaters (and light globes) are made with Negative Temperature Co-efficient materials. That's just the way it is.

  • 1
    \$\begingroup\$ Is a light globe a different word for incandescent light bulb? The filament is made out of tungsten, which is a metal and suffers a great deal from spot failure. \$\endgroup\$
    – Arsenal
    Commented Jan 21, 2022 at 12:06

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.