Loading…
Choose R1 and R2 for a target output from values that are actually stocked, and see the worst-case output window at the resistor and supply tolerance limits before the divider reaches a schematic.
Direct answer
Set the ratio first: V_out/V_in = R2/(R1+R2), so R1/R2 = V_in/V_out − 1. For 5 V to 3.3 V that ratio is 0.515, and the pair the search returns is 13.7 kΩ and 26.7 kΩ, which gives 3.304 V at nominal values. The same pair moves between 3.282 V and 3.327 V at the ±1% resistor corners, a ±0.81% worst case. That fits a ±1% window with about 0.19 points to spare, and misses a ±0.5% one. The search ranks stocked pairs by that corner, not by the nominal ratio.
Calculator
Enter the input and output voltages, the output window you have to hold, the supply variation and the resistor class. The search returns the best stocked pair, its worst-case output at the resistor and supply limits, the divider current, and the next best alternatives.
Model and validity range. Output from the ideal divider V_out = V_in R2/(R1+R2), evaluated over the full IEC 60063 range, E3 through E192, and scored at the four resistor-tolerance corners plus the stated supply tolerance. Resistor classes are the stocked ±0.1%, ±0.5%, ±1% and ±5% bands; the grid step falls from 115% on E3 to 21.2% on E12, 2.41% on E96 and 1.21% on E192. Loading is reported separately: the selection is made at no load and the source impedance is R1||R2.
Requirements
Components
E24 suits ±5% carbon film, E48 ±2%, E96 ±1% metal film and E192 ±0.1–0.5% precision parts; E3, E6 and E12 are the coarsest, cheapest through-hole grids.
Selected divider pair
Ideal R1/R2 = 0.515; 810 stocked E96 pairs scored at the tolerance corners.
R1 13.7 kΩ, R2 26.7 kΩ, output 3.304 volts, worst case 3.282 to 3.327 volts.
From the input to the output tap.
From the output tap to ground.
Target 3.300 V at nominal values.
Resistor corners at ±1.0%.
0.62 mW dissipated in the pair.
Allowed window ±2.0%.
Next best pairs
Same grid and current window, ordered by worst-case error. Every row is a pair you can order.
| Pair | R1 (top) | R2 (bottom) | Output V | Worst case V | Error | Current µA |
|---|---|---|---|---|---|---|
| 1 | 13.7 kΩ | 26.7 kΩ | 3.304 | 3.282–3.327 | ±0.81% | 123.8 |
| 2 | 5.49 kΩ | 10.7 kΩ | 3.305 | 3.282–3.327 | ±0.81% | 308.8 |
| 3 | 54.9 kΩ | 107 kΩ | 3.305 | 3.282–3.327 | ±0.81% | 30.9 |
| 4 | 5.90 kΩ | 11.5 kΩ | 3.305 | 3.282–3.327 | ±0.82% | 287.4 |
| 5 | 59.0 kΩ | 115 kΩ | 3.305 | 3.282–3.327 | ±0.82% | 28.7 |
| 6 | 52.3 kΩ | 102 kΩ | 3.305 | 3.283–3.328 | ±0.84% | 32.4 |
| 7 | 5.23 kΩ | 10.2 kΩ | 3.305 | 3.283–3.328 | ±0.84% | 324.0 |
| 8 | 16.2 kΩ | 31.6 kΩ | 3.305 | 3.283–3.328 | ±0.84% | 104.6 |
Loading moves the ratio
How the pair is chosen
Free, no account. An engineer reviews these inputs before the fixed fee is issued.
Results are engineering estimates, not design verification. Confirm them against the fabrication drawing, a TDR coupon or a 3-D extraction before release. Read the tool terms.
A wrong number, a rejected in-range value or a missing model? Send it to the engineer.
Every result above traces to an equation below. Read the equations and their validity limits.
Equations
Each block gives the equation the calculator evaluates, every symbol with its unit, and the range over which the model holds, with what it cannot see stated alongside. The button in the corner copies the LaTeX source.
The tap sits at the fraction of the input set by the bottom resistor over the total resistance.
R1 and R2 form a series path, so the same current flows through both and the voltage divides in proportion to resistance. The bottom resistor is the one the output is measured across, so R2 is the numerator.
The equation is linear in the input and independent of the absolute resistance. A divider therefore has no scale of its own: 1 kΩ and 2 kΩ give the same ratio as 100 kΩ and 200 kΩ, and the current has to be chosen separately.
Model limits
This is the unloaded divider. A load in parallel with R2 lowers the tap, and the error grows as the load approaches R2 in magnitude.
The target ratio fixes the ratio of the two resistors, and nothing else about them.
Rearranging the divider equation gives the ratio directly. For 5 V to 3.3 V, 5/3.3 − 1 = 0.515, so the top resistor is about half the bottom one.
The ratio is the only constraint the voltages impose. The absolute values follow from the divider current, the source impedance the load needs, or the power the board can spare.
Model limits
Only valid for V_out below V_in. At or above the input the top resistor would have to be zero or negative, which is not a divider.
The two resistor limits put the output between these two ratios, and the supply tolerance widens them further.
delta is the resistor tolerance as a fraction. The minimum ratio uses R2 low and R1 high, because a smaller bottom resistor and a larger top resistor both push the tap down. The maximum uses the opposite pair.
The two bounds are not symmetric around the nominal ratio unless R1 and R2 are equal. A divider that steps 5 V down to 3.3 V has more room on the low side, which means a pair with a small positive nominal error can still have the smallest worst case.
Model limits
It assumes both resistors move independently to their limits. A resistor network in one package, or two parts from the same production lot, can be more correlated than this.
Once the ratio is fixed, the absolute resistance sets the standing current and the power it costs.
The pair is a series circuit across the supply, so the current is the input voltage over the total resistance. That current is also the load the divider places on the source, whether or not anything reads the tap.
Power splits in proportion to resistance: R1 carries I²R1 and R2 carries I²R2. The source impedance is R1 in parallel with R2, which is lowest when both resistors are small and the current is high.
Model limits
The current is the nominal value at the specified input. At the high supply limit it rises proportionally, and the resistor power ratings have to cover that corner.
The grid is logarithmic, and its step is the smallest ratio change two stocked values can make.
An E-series divides each decade into N logarithmic steps, so the ratio between neighbours is 10^(1/N). The values are rounded to the published preferred numbers, but the step is what limits the choice.
E3 gives a 115% step and E192 a 1.21% step; E6, E12, E24, E48 and E96 sit between at 46.8%, 21.2%, 10.07%, 4.91% and 2.41%. A coarse grid can miss a target by more than a tight resistor tolerance can correct, hence the calculator reports the series next to the pair.
Model limits
The step is nominal. Manufacturers sometimes add values outside the series, and distributors sell merged ranges such as E96 + E24; those values are not part of this search.
A load in parallel with R2 lowers the tap, and the error is set by the load current against the divider current.
The load draws current from the tap, so it behaves as a second resistor to ground in parallel with R2. The parallel combination is always smaller than R2, which pulls the output down.
The error is small while the load current is much smaller than the divider current. A load drawing 1% of the divider current moves the tap by about 1% of the tap voltage, so the ten-to-one rule is a useful starting point rather than a precise specification.
Model limits
A load that is not purely resistive, such as an ADC sampling network or a comparator input with bias current, needs its worst-case current at the tap rather than a single resistance.
Selection
The divider equation is one ratio: R1/R2 = V_in/V_out − 1. For 5 V to 3.3 V that is 0.515, and a calculation that stops there hands you 51.5 kΩ and 100 kΩ. Neither is a stocked E96 value, and the pair that is stocked lands somewhere else on the ratio curve. The difference between those two answers is the quantization error, and it is the floor under every tolerance figure that follows.
Adjacent values differ by 115% on E3, 46.8% on E6, 21.2% on E12, 10.07% on E24, 4.91% on E48, 2.41% on E96 and 1.21% on E192. A pair chosen from those grids can be close, but no closer than the grid allows, and a coarse grid can miss the target by more than the resistor tolerance does. The search here enumerates the chosen grid and ranks every reachable pair by its worst case, so the recommendation is a pair that can be ordered rather than a ratio that cannot be built.
Tolerance
Each resistor has a tolerance band, and the output is worst when one sits at its low limit and the other at its high limit. That is two corners, not a root-sum-square: R1 high with R2 low puts the ratio at its minimum, and R1 low with R2 high puts it at its maximum. The calculator evaluates both, then widens them again by the supply tolerance on the input.
For a 5 V to 3.3 V divider on ±1% parts, the resistor corners move the output by roughly ±0.7% around its nominal value. Add a ±5% supply and no resistor pair can hold a ±2% window, because the input alone is already outside it. The tool reports that instead of returning a pair that cannot work, which is the more useful answer when the requirement is wrong rather than the components.
Worst-case addition is deliberate. RSS assumes independent, symmetric errors and describes a yield, not a limit; a production limit has to hold at the corner where both parts are wrong in the same direction. Use RSS when you are predicting how many boards pass, and even then remember that the E-series grid is not a random distribution.
Drive
A divider is a resistor pair with a Thévenin source impedance of R1||R2, and that is the number that decides whether it can drive anything. At 5 V and 100 µA the pair totals 50 kΩ, so a 3.3 V tap has a source impedance near 11 kΩ; a 10 kΩ load in parallel with R2 changes the ratio by about 7%.
Pick the current from the load: at least ten times the worst-case input bias or leakage of whatever reads the tap, and low enough that the standing power is acceptable. 10 µA to 1 mA covers most ADC scaling and comparator thresholds. Below about 10 µA, board leakage, flux residue and humidity become part of the divider, and the tolerance calculation stops describing the circuit.
Total power is V_in²/(R1 + R2). At 5 V and 50 kΩ that is 0.5 mW, which any 0402 part carries; the reason to care is a battery, where a permanent 100 µA drain is 0.9 mAh per day. The power splits in proportion to resistance, so the top resistor dissipates more whenever the tap is below mid-supply.
FAQ
Send the stackup, the interface list and the voltage divider inputs. An engineer re-solves them against the real layout before the fixed fee is issued.