Resistor ratio finder
Give it the ratios you need — R1:R2 and R1:R3 together, one plain ratio, or the two attenuations of a switchable divider — and get the standard resistor sets that come closest, ranked by the worst ratio error, from E6 to E192 or from the parts in your drawer.
The largest current any one resistor can carry with the full voltage across it; the level shifter page gives the exact currents.
Enter the voltage to apply the current limit.
Best standard pairs
Resistor sets from the chosen series with the smallest worst-ratio error first, then the lowest total. Click a heading to sort by it, or a row to see that set in the drawing.
The formula
k2 = R1 / R2 k3 = R1 / R3 Error = (R1 / Rx) / k − 1, peak = the larger |error| Divider: k = 1 / A − 1; Rp across R2: R1 / Rp = 1 / A2 − 1 / A1
When to use it
Use it whenever two or three resistors must hold fixed ratios and only the ratios matter: a gain-and-offset network around an ADC, a ladder, a switchable attenuator, a reference scaled in two taps. Matching one ratio is easy; matching two at once with parts from one series is what this tool searches for, trying every R1 in your range and the neighbouring series values for R2 and R3. Resistor tolerance is the next limit: a ratio good to 0.02 % on paper is only as good as the parts you buy, so use 0.1 % resistors, or matched networks, when the paper accuracy is meant to survive.
Pitfalls
- The error is the ratio error of nominal values. Real parts add their tolerance: two 1 % resistors can still move a ratio by 2 %, so a 0.02 % match needs 0.1 % or better parts (or a matched network).
- Absolute values matter once something else is connected: very high values pick up bias-current error and noise, very low ones burn current and load the source. Set the range to decades you can afford; the answer in each decade is the same ratio.
- Each decade of the same set is shown once, kept in the decade nearest the middle of your range. Different mantissas are different answers and stay separate, so 10k/20k/40k and 15k/30k/60k both appear.
- In the switchable divider the search ranks the ratio errors. With a resistor switched across R2, an error in the second ratio matters less than the same error in the first, so check the attenuation errors in the table, not only the peak.
Related note
Reverse engineering a board with no schematicNeed the whole analog front end designed, not just the resistor values?