Reading resistor and capacitor markings

Colour bands, three-digit codes, the R and V conventions, and tolerance letters — how to read what is actually printed on a part.

Two schemes, one idea

Nearly every passive component marking — colour bands on a resistor, three digits on a ceramic capacitor, the code on a surface-mount part — encodes the same structure: significant figures, then a multiplier, then a tolerance. Once you see that shape, the individual schemes stop being things to memorise and become variations on one idea.

The reason they exist is space. There is no room to print "0.000000047 farads" on a part the size of a grain of rice, and printing "47 nF" still needs a unit symbol that photocopies badly and rubs off. A positional code solves both.

Resistor colour bands

The colour sequence runs black, brown, red, orange, yellow, green, blue, violet, grey, white for 0 through 9. That order is the visible spectrum with black and brown bolted on the front and grey and white on the end, which is why it feels arbitrary at the edges and natural in the middle.

A four-band resistor is two significant figures, a multiplier, and a tolerance. Brown-black-red-gold is 1, 0, ×100 — so 1,000 Ω at ±5%. A five-band part gives three significant figures for tighter tolerances, and a six-band part adds a temperature coefficient.

The practical difficulty is not decoding but orientation: which end do you start from? Two habits solve it. The tolerance band is usually gold, silver or brown and sits slightly apart from the group. And if you read it the wrong way round you generally get a value that is not astandard value — a strong hint you have it backwards. Thecolour code calculator shows the bands so you can compare against the part in your hand.

Small surface-mount resistors use printed digits instead, in the same significant-figures-then-multiplier form: 103 is 10 kΩ. Precision parts use the EIA-96 scheme, a two-digit code plus a letter, which genuinely does require a lookup table.

Capacitor three-digit codes

Ceramic capacitors use the same idea with the value in picofarads: two significant figures, then the number of zeros.

  • 104 — 10 with four zeros = 100,000 pF = 100 nF = 0.1 µF
  • 103 — 10 nF
  • 473 — 47 nF
  • 220 — 22 with no zeros = 22 pF

That last one is the trap, and it is worth dwelling on. 220 is 22 pF, not 220 pF. A genuine 220 pF part is marked 221. Reading it as the literal number is a factor-of-ten error, and it is by far the commonest misreading of the code.

Two extensions round it out. A third digit of 8 or 9 means multiply by 0.01 or 0.1 rather than adding zeros — so 229 is 2.2 pF. And R marks a decimal point: 4R7 is 4.7 pF. The capacitor code calculator handles all of these.

The letter conventions

A letter after the value is normally tolerance: J is ±5%,K ±10%, M ±20%. So 104K is a 100 nF part at ±10%.

But letters also stand in for decimal points, in a convention that shows up across the whole field. 4R7 means 4.7 Ω; 1K5means 1.5 kΩ; 5V1 on aZener diode means 5.1 V;3V3 means 3.3 V. The unit letter takes the decimal point's place.

This is not decoration. A printed decimal point is a single dot that disappears in a photocopy, survives poorly on a curved surface, and is easy to mistake for dirt. A letter cannot be lost that way. The convention exists because losing a decimal point costs a factor of ten.

The dielectric code matters more than the tolerance

On ceramic capacitors you will also see a three-character code such asX7R, C0G (often written NP0) orY5V. This describes the dielectric, and it frequently matters more than the stated tolerance.

C0G/NP0 is a class 1 dielectric: stable with temperature, stable with voltage, low loss. Use it where the value must actually be the value — timing, filters, oscillators.

X7R and Y5V are class 2. They pack far more capacitance into the same package, and they pay for it. X7R holds within ±15% over its temperature range; Y5V can lose most of its capacitance at the extremes. Worse, class 2 parts lose capacitance under DC bias — a 10 µF X5R rated 6.3 V can behave as 5 µF or less when actually run at 5 V.

So a "10 µF ±10%" part can genuinely be delivering half its marked value in circuit, entirely within specification, because the tolerance and the bias derating are separate things. For decoupling this rarely matters. For anything where the value is part of a calculation — atime constant, a filter corner — it matters a great deal.

Why the values are the values

None of these schemes explain why you can buy 4.7 kΩ and 10 kΩ but not 4.8 kΩ. That comes from IEC 60063, which defines the E-series: values spaced by a constant ratio so that the percentage gap between neighbours stays the same across the whole range.

E12 divides each decade into twelve steps roughly 20% apart, which pairs with 10% tolerance parts; E24 gives 24 steps for 5% parts; E96 gives 96 for 1%. The tolerance and the series are meant to match — there is no point specifying an E96 value if you are fitting a 5% resistor, because the tolerance band is wider than the gap to the neighbouring value.

One oddity worth knowing when you compare a table against the formula: the lower series are historical lists rather than generated ones, and several entries were rounded to friendlier numbers. E24 contains 2.7 where the formula gives 2.61, and 8.2 where it gives 8.25. E48 and above follow the mathematics properly.