How conservatively are resistore rated?

[JimRayden]

I mean the wattage rating? Can I exceed it a bit? I mean, something like using 0.54W into a 0.5W resistor.


---------
Jimbo

[davebungo]

In this particular case I would have thought that you will be OK but...

the wattage rating of resistors is not an absolute limit; it is usually specified as the limit for an ambient temperature of 70 deg C.  If the ambient temperature exceeds this then the actual wattage must be reduced according to a "derating curve".  I'm not an expert so I would do a search for "component derating" and "reliability" and read a few datasheets if you're that interested.

Personally, I would run it at half the rated wattage to be sure - reliability will be increased in the long term.  It's not always the done thing in consumer equipment though as cost per watt is a big issue if you're talking about amplifiers and so on.  As usual it is always about trade-offs.

[pjwhite]

A good rule of thumb for specifying resistors is to double the expected wattage.  If you know you're going to be dissipating 0.54 watts, go with a 1 watt, or even a 2 watt resistor.  You'll be generating the same amount of heat either way, but with a physically larger resistor, the heat will be spread out more and not so concentrated and likely to burn something.
With some sort of active cooling you might be able to use a 1/2 watt resistor without having your circuit board turn brown, but you're not likely to have a fan or water cooling in a stompbox.

[Paul Perry (Frostwave)]

A resistor is happier up in the air, rather than jammed against the board. I would say, putting it in the air would increase the rating 30%.
A good rule is, if you touch a resistor & you can still see the mark on your finger the next day, you need a bigger one.
Personally, I'd be happy to run .54W into a .5W resistor.

[R.G.]

In general, resistor wattage is the amount of power it's dissipating when the surface temperature is 100C. That's the boiling point of water, the sizzling point of skin. This says nothing about how long the device will dissipate that without dying.

It is good practice for longevity to derate resistors by at least 50% - that is, never dissipate more than 1/8W in a 1/4W resistor. That keeps the surface temperature down a lot.

Life of all electronic components is inversely related to temperature. For most things, a temperature increase of 10C doubles the random failure rate.

You absolutely can push the power rating. For a given surface area and conditions, the temperature varies with about the third power of the power being dissipated, so it goes up *fast* as you get over rated power. And as temperature goes up, lifetime comes down. Eventually you reach a temperature where the remaining life is in seconds. How much life do you want your circuit to have? Some military applications demand one and only one operation. Some consumer applications demand all-day/every-day operations for decades.  What are *you* circuit's application needs?

Pulse power conditions can be withstood to varying degrees by different technologies. This is one where carbon comp is a clear winner. It can eat pulses of two to five times rated power better than film types. Wire wound is also good for pulses.

[JimRayden]

A resistor is happier up in the air, rather than jammed against the board. I would say, putting it in the air would increase the rating 30%.
A good rule is, if you touch a resistor & you can still see the mark on your finger the next day, you need a bigger one.
Personally, I'd be happy to run .54W into a .5W resistor.

Actually I did have it floating in the air, it's a point-to-point tube amp.

Thanks everyone for the grrreat replies. Actually I was guessing the 50% thing myself, but I bought the resistors before calculating wattage through it. Oh well I just might have to go and rebuy 'em.


A quick idea - wouldn't wrapping a fat wire around the resistor work as a heat sink?

-----------
Jimbo

[petemoore]

Add DiY heat sinks to the raised resistor
 Soldering a sheet of copper to one or both ends might make an antenna.