What determines resistors power ratings if not physical size?

[dentaku]

I'm wondering how the power rating of resistors relates to their size.

I got some 100K 500mW 1% resistors recently and they're tiny compared to the regular resistors I'm used to which are mostly all 250mW.
If it's not their size then what determines resistors power ratings?

[rob77]

many factors.
material used - some materials can withstand higher temperatures
surface area - the bigger the surface the more heat can be dissipated
thermal resistance of the materials - if the material is better conducting heat, then more heat can be dissipated though the leads, or PCB traces

but as rule of thumb... the bigger the resistor, the higher the power rating is.

[mazurov]

Temperature coefficient of material is also a factor. For example, copper makes a good conductor but poor resistor since its TC is almost 0.4% per degree C. On the other hand, Vishay's WSLP2726 all-metal beauties with cross-section similar to 16ga wire are rated only 5-7W. They are made of nichrome - an alloy with TC 10 times smaller than copper but much higher resistance.

[dentaku]

Temperature coefficient of material is also a factor. For example, copper makes a good conductor but poor resistor since its TC is almost 0.4% per degree C. On the other hand, Vishay's WSLP2726 all-metal beauties with cross-section similar to 16ga wire are rated only 5-7W. They are made of nichrome - an alloy with TC 10 times smaller than copper but much higher resistance.

The ones I got are Vishay metal film resistors. I was amazed at how small they are and that they where 500mW
http://canada.newark.com/vishay-bc-components/sfr16s0001003fr500/resistor-metal-film-100kohm-500mw/dp/05R0211

[T3sl4co1l]

Chip resistors are almost uniformly rated for 0.1W: 0402 to 0805, very roughly.  The limiting factor is thermal diffusion, which has a sqrt(length) proportion: if the aspect ratio is not changed with size, then the total dissipation varies relatively slowly with size.

They also assume a naked chip resistor with thin traces; you can get much more from an 0603 with heavy copper pours around.

TH resistors vary, but temperature of the materials is probably the biggest factor.  Epoxy vs. silicone enamel (or ceramic, or vitreous!), tinned vs. welded connections, etc.  And whether they guarantee tempco or tolerance within the temp range, or what.

Tim