The resistor spec nobody checks until it burns

A resistor has three specifications that matter: its value, how close to that value it is guaranteed to be, and how much heat it can shed before it fails. The colour bands tell you the first two. They say nothing at all about the third — and the third is the one that destroys parts.

What the bands actually encode

Reading them is mechanical. The first two or three bands are significant digits, the next is a power-of-ten multiplier, and the last is tolerance. Yellow-violet-red-gold is 4, then 7, then ×100, then ±5% — a 4.7 kΩ part. The resistor colour code calculator does the decoding and, more usefully, tells you when the answer is not a value anyone manufactures.

That last check matters more than it sounds. Colours fade, brown and red look identical under a desk lamp, and the commonest mistake is reading the resistor from the wrong end. If your decoded value belongs to no standard series, you have almost certainly read it backwards rather than found an exotic part.

Tolerance is a promise, not a measurement

A ±5% 4.7 kΩ resistor measuring 4.55 kΩ is not faulty. It is in specification, and the manufacturer never claimed otherwise. Tolerance is a guarantee about the worst case, not a description of the typical part.

This has a practical consequence for anything ratiometric. Two ±5% resistors in a divider can be 10% apart from each other in the worst case, so a divider built from ordinary parts is not a precision reference. Where the ratio matters, buy ±1% parts — and where it matters a lot, buy a matched pair from the same reel.

Why the value you calculated does not exist

Resistors are only made in the E-series of preferred values, which are spaced so that consecutive values are roughly one tolerance band apart. E12 gives twelve values per decade, E24 twenty-four, and precision parts go to E96 and beyond. The practical effect is that your calculated 500 Ω does not exist: you buy 470 Ω or 560 Ω in E12, or 510 Ω if you stock E24.

Which way you round is a design decision rather than an arithmetic one. For a current-limiting job, round up — more resistance means less current, which costs a little brightness rather than the part’s life. For a pull-up, round either way. If nothing close enough exists, the series and parallel resistance calculator will make the value up from parts you already own.

Power is where resistors actually fail

A resistor turns electrical energy into heat, and its rating is the heat it can shed in free air at room temperature before it drifts, discolours and eventually goes open circuit. The rating is a function of the part’s physical size, and nothing in the colour code refers to it.

The dissipation is P = I²R, or equivalently V²/R across the part. Both forms are in the power calculator, and either is worth a moment before you fit anything carrying real current. The number that surprises people is how small a quarter-watt is: 25 mA through 400 Ω is already a quarter of a watt, and 25 mA is not much current.

Do not design to the rating either. A resistor run at its stated wattage is running hot enough to bake the board around it and shift its own value, and the derating curves in the standards pull the permissible dissipation down further as ambient temperature rises. A working rule is to pick a part rated at twice what you calculate, and more inside a sealed enclosure.

The LED case, end to end

An LED is the clearest worked example, because the resistor is doing an obvious job. Take a red LED with a 2.0 V forward voltage, running at 20 mA from a 12 V supply. The resistor has to absorb 12 − 2.0 = 10 V, so it must be 10 ÷ 0.020 = 500 Ω — which, as above, is not a part. Round up to 560 Ω and the real current becomes 17.9 mA, close enough that no eye could tell.

Now the rating. P = 0.0179² × 560 = 179 mW. That is well past a 1/8 W part and uncomfortably close to a 1/4 W one, so the honest answer is 1/2 W. A single indicator LED on a 12 V rail needs a resistor four times the rating most people reach for, and the LED resistor calculator works the whole chain out including the wattage.

It is also worth noticing where the energy went. Only 2 V of the 12 reached the LED, so five-sixths of the power became heat in a component whose entire purpose is to waste it. That is why LED strips are wired as long series chains: more LEDs sharing one supply means less surplus voltage to throw away.

Two ways your meter lies

Measuring a resistor while it is still soldered in reads low, because your meter sees it in parallel with every other path around it. Desolder one leg before comparing a measurement with the marking, or you will condemn a good part.

The second is subtler: a resistor that has been overheated often measures fine but has drifted, and carbon film parts drift upward with age and thermal cycling. If a circuit worked and now does not, a scorched-looking resistor is worth replacing whatever the meter says. Where you need to confirm a part in situ, put a known voltage across it and check the current against Ohm’s law rather than trusting either the bands or a cold reading.