Reading Resistor Color Bands: Why the Tolerance Band Actually Matters

How to Read Resistor Color Codes covers decoding every band, tolerance included. This is a closer look at just that last band — what it actually promises, and the specific situations where ignoring it causes a circuit to misbehave in a way that’s hard to track down.

What the tolerance band is actually a promise about

A resistor’s printed value — the number the color code decodes to — is a manufacturing target, not a guarantee. No production process makes every resistor exactly, say, 1,000Ω on the nose; there’s always some small part-to-part variation. The tolerance band is the manufacturer’s stated bound on how far a real part is allowed to drift from that printed target and still be sold as that value. A resistor marked 1,000Ω with a gold (±5%) tolerance band is a promise that the real, physical resistance is somewhere between 950Ω and 1,050Ω — not a promise that it measures exactly 1,000Ω.

Four tolerance bands, four different ranges

The width of that range depends entirely on the tolerance band, and it scales with the resistor’s own value, not by a fixed number of ohms:

• Orange–orange–red–gold decodes to 3,300Ω (3.3kΩ) at ±5% — a true value anywhere from 3,135Ω to 3,465Ω, a 330Ω window.
• Brown–green–black–brown–brown (5-band) decodes to 1,500Ω (1.5kΩ) at ±1% — a much tighter true value between 1,485Ω and 1,515Ω, a 30Ω window despite being a smaller resistor than the one above.
• Red–red–black–brown–brown (5-band) decodes to 2,200Ω (2.2kΩ) at ±1% — between 2,178Ω and 2,222Ω, a 44Ω window.
• Brown–black–red–silver decodes to 1,000Ω at ±10% — a noticeably looser 900Ω to 1,100Ω, a 200Ω window on a resistor smaller than the first example.

The pattern: the tolerance percentage sets the width as a proportion of the resistor’s own value, so a small, tight-tolerance resistor (1.5kΩ at 1%) can have a narrower absolute window than a bigger, loose-tolerance one (3.3kΩ at 5%), even though the second resistor is more than twice the size.

Where a loose tolerance genuinely doesn't matter

For an LED indicator circuit, a resistor sized with a target current in mind is deliberately built with margin — the LED will light visibly across a fairly wide range of actual current, and rounding to a standard part already introduces a bigger shift than most tolerance bands do. A 20% swing on a 390Ω LED resistor changes the current only marginally, and the practical result is an LED that’s very slightly brighter or dimmer than predicted — never a failure, never a circuit that stops working. This is exactly why 4-band, 5% (or even unbanded, 20%) resistors are the standard, inexpensive default for hobby circuits: the tolerance is loose, but nothing about the circuit's function depends on hitting an exact number.

Where it does matter: a worked voltage-divider example

A voltage divider — two resistors in series, with the output taken from the point between them — is the clearest case where tolerance directly changes what a circuit delivers, because the divider’s whole job is producing one specific, predictable voltage. Take two resistors intended as 1,000Ω and 2,000Ω across a 9V supply, each with a ±5% tolerance:

• Nominal case (both resistors exactly on target): total resistance 3,000Ω, current 3mA, and 6V appears across the 2,000Ω resistor.
• Worst case in one direction (the 1,000Ω resistor 5% low at 950Ω, the 2,000Ω resistor 5% high at 2,100Ω): total resistance 3,050Ω, current 2.95mA, and the output rises to 6.2V.
• Worst case in the other direction (1,050Ω and 1,900Ω): total resistance 2,950Ω, current 3.05mA, and the output falls to 5.8V.

So a divider “designed” to output 6V can, using two ordinary ±5% resistors, actually output anywhere from 5.8V to 6.2V depending on exactly where each part landed within its own tolerance window — a spread of nearly half a volt on a 6V target. For a circuit that just needs roughly half the supply voltage, that’s harmless. For a divider feeding a reference voltage that something else depends on precisely, it’s the entire reason precision work reaches for 1% (or tighter) parts instead of 5% ones — swapping in 1% resistors for the same divider would shrink that half-volt spread down to a small fraction of it.

Combining resistors doesn't average tolerance away

It’s tempting to assume that using several resistors together somehow smooths out their individual imprecision, the way averaging several independent measurements narrows an error. It doesn’t work that way for tolerance, because every resistor in a batch isn’t randomly independent — in the worst realistic case, several parts from the same reel could all drift the same direction at once, and a careful design has to assume that can happen. Take two 100Ω resistors, each rated ±5%, wired in series. Nominally, that’s 200Ω. If both resistors happen to sit at the high end of their own range (105Ω each), the series total is 210Ω — still exactly ±5% of 200Ω, just now representing a full 10Ω of absolute drift instead of the 5Ω either resistor carried alone. If both sit at the low end (95Ω each), the total is 190Ω, again ±5%. The percentage doesn’t shrink by combining same-tolerance parts — it’s the absolute window (in ohms) that grows, in direct proportion to the added resistance.

That’s the conservative, worst-case way to reason about combined tolerance, and it’s the right assumption for anything where being wrong has a real cost — a precision reference, a timing circuit, anything a mistake in would be inconvenient to debug after the fact. For a simple LED resistor or general current-limiting job, this level of rigor is overkill; the point is knowing which category a given circuit falls into before deciding it doesn’t matter.

Common mistakes with tolerance

• Confusing tolerance with temperature stability. Tolerance describes how far a resistor can be from its printed value at the outset, at a reference temperature — it says nothing about how much that value shifts as the resistor heats up during operation. A separate specification, temperature coefficient, covers that, and precision designs sometimes need to check both.
• Assuming a tighter tolerance is always worth paying for. A 1% resistor typically costs more than a 5% one for no functional benefit in a circuit that never needed the precision in the first place — matching tolerance to what the circuit actually requires, rather than defaulting to the tightest available, is the more useful habit.
• Forgetting that an unbanded resistor still has a tolerance. No fourth band doesn’t mean no variation — it means the loosest standard tolerance, ±20%, still applies by default.

A rule of thumb for choosing tolerance

Ask what the circuit actually needs the resistor's exact value for. If the answer is “to keep current roughly in a safe range” — an LED resistor, a general current-limiting job — an ordinary 5% (or looser) part is normal and sufficient. If the answer is “to set a specific voltage, ratio, or timing that something else depends on” — a voltage divider feeding a reference, a precision timing circuit — that’s the signal to reach for 1% or tighter parts, and to check the combined effect of every tolerance-bearing resistor in that specific part of the circuit, the way the divider example above does.

Why loose-tolerance parts are still the default

If tighter tolerance sounds strictly better, it’s worth asking why 5% and 10% resistors remain the standard stock item rather than everything defaulting to 1%. The answer is that tolerance and manufactured value spacing are linked by design, not by accident: the standard E-series of resistor values (E12, E24, E96, and so on) is spaced so that each series’ typical tolerance band just reaches its neighboring values without excessive overlap. A looser tolerance needs fewer distinct manufactured values to cover the same overall range without gaps, which keeps looser-tolerance parts simpler and cheaper to produce and stock in bulk. Reaching for 1% resistors everywhere wouldn’t just cost more — it would mean carrying eight times as many distinct values in a parts bin (96 per decade instead of 12) for precision that most circuits, LED indicators and simple current-limiting jobs included, never asked for.

Reading the band correctly matters as much as having it

None of this arithmetic is useful if the tolerance band itself gets misread. Gold and silver are the two colors that do double duty across positions — both can appear as a multiplier (gold ×0.1, silver ×0.01) or as a tolerance (gold ±5%, silver ±10%), and telling those roles apart comes down to position: multiplier is the second-to-last band, tolerance is the last. When a resistor has no fourth (or sixth) band at all, the tolerance isn’t undefined — it defaults to a loose ±20%, the widest range on this list, and worth remembering precisely because it’s easy to forget an unbanded resistor has a tolerance at all. Decode carefully with the Resistor Color Code Calculator, and the tolerance percentage it returns is the number to actually reach for — not to skip past on the way to the ohm value.