Calculator Library / Design of Experiments
Choose a two-level design, enter your factors and levels, and get a randomized, printable run sheet with the resolution, the aliasing, and the line time. Copy it into Excel or Minitab.
Tool
Run sheet = design × randomized order × blocks
Choose a two-level design, name the factors and set their low and high levels, and the tool builds a randomized run sheet you can print and fill in, or copy into Excel or Minitab. Load the example to see the weld experiment from the DOE Toolbox.
Everything runs in your browser. Nothing is sent to a server, and entries are not saved between visits. Levels may be numbers or words; a center point needs numeric low and high levels.
Plan
Aliasing
Run the setups in the order shown, within each block. Reset every factor between setups, write down the setting you actually read, and note anything unusual. Do not change the order to save time.
Instructions
Pick a two-level design, name the factors and enter their low and high levels, and the generator builds a randomized run sheet. It supports full factorial designs and the standard half, quarter, eighth, and sixteenth fractions, with replicates, center points, and two blocks. It reports the resolution and the aliasing, estimates the line time, and lets you print the sheet or copy it as CSV for Excel or Minitab.
It is the companion to the Run step of the DOE Toolbox, and it uses the same standard generators as Minitab’s default two-level designs.
| Element | How it is made | Why |
|---|---|---|
| Base design | A full factorial in the first factors, with −1 and +1 for low and high | Every combination of the base factors |
| Extra factors in a fraction | Each is a product of base columns (for example F = ABCDE) | Runs part of the full design while keeping the columns balanced |
| Resolution | The length of the shortest word in the defining relation | Tells you which effects are mixed up |
| Blocks | Runs are split by the sign of a high-order interaction, chosen so that no main effect or two-factor interaction is confounded with blocks | Separates a day or batch difference from the factor effects |
| Center points | Runs with every factor at its midpoint, spread between the blocks | Estimate pure error and test for curvature |
| Randomization | A seeded shuffle within each block | Spreads drift across all settings; the seed makes the order reproducible |
The pre-loaded setup is the weld experiment from the DOE Toolbox: six factors, a half fraction with 32 setups (resolution VI), two housings welded at each setup, and two blocks for two line days. The generator confirms 32 setups, 64 housings, and 8.5 hours of line time at 10 minutes per setup change and 3 minutes per housing. All main effects and two-factor interactions are clear of each other, and the blocks are confounded with the A × B × C interaction.
Change the design to a quarter fraction of six factors and the resolution drops to IV with 16 runs: main effects are still clear, but two-factor interactions become mixed up in pairs. Change the design to a half fraction of three factors and the resolution drops to III, which the tool flags as a design for screening only.
Yes. Running the setups in the randomized order spreads drift, such as tool wear, temperature, and operator fatigue, across all the factor settings so it does not look like a factor effect. Running in the standard order lines drift up with the factors that change slowly.
The same seed always produces the same order, so you can regenerate the sheet later or give someone else the same order. Click New random order to try another seed, and keep the seed with your records.
A run with every factor set at the midpoint of its low and high levels. Center points estimate pure error and test whether the response is curved. They need numeric levels.
Blocks must be confounded with an effect you are willing to lose. The tool looks for an interaction of three or more factors whose aliases also have at least three factors, so no main effect or two-factor interaction is lost.
Yes. Copy it as CSV or download the file and paste or open it in Minitab or Excel. Use the actual-value columns for the factors, and enter the results in the result columns. Minitab can also create the same design with Stat > DOE > Factorial > Create Factorial Design.
No. The generator runs entirely in your browser, nothing is sent to a server, and your entries are not saved when you leave the page.