Power Resistor

[fuseshut]

I got a basic question regarding power resistors. I know when specifying a power resistor, the power rating shall be twice the expected dissipated power. For example, I am using 3000 ohm, 10W resistor in a circuit application in which 120VAC will be across it. It is a nice power resistor from ohmite, 20 series. If you do the simple calcs I am only drawing 40mA, which gives me 4.8Watts. now it is a 10W part and it is still getting hot to the touch. I even lowered the AC voltage down to 100VAC , which gives me 3.68W and is still getting hot to the touch. It is not a heatshinkable resistor, it's one of the tubular, ceramic axial leaded resistors. I ran it for 10minutes across 120VAC, then 30minutes just to make sure the resistor wouldn't burn up. And it did not. Measured it and still read 3000 ohms on the meter. Current was steady also! In the end application the power resistor will be sitting inside a enclosure full of oil. So the oil will keep it cool. I'm just curious on why they get so hot, even at 3.65W (10W rated part), and there is no heatsink required (so the datasheet states).
Thanks

 

[BrianR]

Comkid, you're right, the low emissivity of gold makes it a poor radiator but aren't those gold resistors meant to be bolted to heatsinks. The white ceramic ones have a high emissivity, typically 0.9 so they radiate well.

It's not always true that black is a better radiator either as surface texture and properties are more important than colour. It varies with the waveband of interest but at lower temperatures (<100C) convection is probably the dominant heat remover and surface coatings can actually insulate. I have used the black pen trick too but don't know what it's emissivity is.

 

[Comcokid]

You're right. The gold ones are meant to conduct heat to what you've bolted them to. But you know how it is - you're trying something real quick, so you take that 25 Watt Dale and hang-it-loose in a 6 watt breadboard application because it's in the parts cabinet. If it's hotter than you want, make it black. The gold does look pretty, but I've seen some in surplus military equipment that were actually produced as black.

If you could see in IR, you could tell if "Black" in visual light was still "Black" in IR. (Jordie of TNG could tell.) Anyway, Sharpies seem to work good. Now, if I spray painted it black, I would be concerned that the paint itself with it's thickness was a thermal insulation barrier.

 

[BrianR]

This is quite beyond the depth of this topic but I was in the position to take some actual numbers and they might be of interest to others.

A piece of steel at 330C radiated 33mW/cm2/sr in 9-12u band and 18mW/cm2/sr in the 3.5-4.5u band.

Metal with laquer paint at 40C radiated 3.73mW/cm2/sr in 9-12u band and 0.13mW/cm2/sr in the 3.5-4.5u band.

Totally useless info.....

 

[felixc]

Aside from warning of a potential problem, the finger test isn't a good one for resistors, even if you can read the manufacturer logo on your finger. 50 Celsius will feel hot on a finger, especially after a minute. Meat cooks at 70 Celsius. But these are nothing compared to the operating temperature of a power resistor. Get an IR thermometer and make real measurements. Burning hot to your finger may not be a problem at all to your product.

 

[Warpspeed]

50 Celsius is not actually that hot. The thermostatic mixing valve on my single bath tap is set to 45C, and the whole family by pouplar vote decare it perfect. Here in Oz, the maximum staturory hot water supply temperature for schools and kindergartens is 50C so the kiddies cannot scald themselves.

If immersion in 50C hot water is not going to hurt a five year old, your gnarled toughened much abused engineers thumb sure isn't going to burst into flames at 50C.

I bet the tea/coffee you drink is hotter than 50C.

 

[IRstuff]

European standards are set for 48ºC surface temperature for a 10 min contact exposure and 43ºC for an 8 hr exposure.

The implication is that you can indeed get burned at 50ºC, it's just that most people will get the message before any damage occurs.

TTFN

 

[itsmoked]

It all gets back to the rule of thumb: [purple]Every 10 degrees doubles the chemical reaction rate.[/purple]

That is: The solder oxidizes faster, the board ages faster and the resistor ages faster. Too fast? Depends of course.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[felixc]

Just to show that our body isn't a reliable temperature measuring instrument. I measured my "hot but drinkable" tea temperature at 54 Celsius with an IR gun. Sipping very small quantities to feel the heat that very quickly dissipates into the mouth, exciting the tastebuds to make them transmit better the senses. Within two seconds the IR gun in my mouth showed 41 Celsius.

I also remember making system longevity tests in a 55 Celsius chamber. Keeping my hands in contact with the stainless steel casings was out of question.

Too hot is relative. The CPU chip on my main product has an operating case temperature of 55 Celsius. Perfectly within specs; customers are even happy to know that the temp is only 55 Celsius. I would not keep a finger on it for long even if the presence of the finger drains heat and makes the temperature go lower.

Hey Keith, would people living in northern countries live
longer than people living in the south? Stats might prove it right.