Yes, and it is the least glamorous habit that saves the most time. Every fabric you sew needs a combination of needle type, needle size, thread, stitch length and tension. You work that out once with test seams, then forget it, then work it out again next year.
What to record. Fabric and weight. Needle type and size. Thread. Stitch length and width. Tension number. Presser foot pressure if your machine adjusts it. Differential feed if it was an overlocker.
Where to record it. A notebook beside the machine is enough. Better still, staple the test scrap itself to the page — the stitching tells you more than the numbers do, and it settles arguments about what “slightly puckered” meant.
Why it matters more than it sounds. Tension is not absolute. The same number behaves differently with a different thread, and machines drift a little after a service. A written record with a physical sample lets you tell the difference between the machine changing and your memory changing.
Also note what failed. “80/12 universal skipped on this jersey” is as useful as the setting that worked, and stops you repeating it.
Mark your default. Write down the tension number your machine sits at for ordinary cotton, so you always have a known point to return to after an experiment.
And date the entries. When a machine starts behaving differently, the dated record is how you notice it is the machine and not the fabric.
Questions people actually ask
Is a photo enough?
For the dial positions, yes. But photograph the test seam as well as the dial, or you lose the reason you chose that setting.
My machine sets tension automatically.
Still record needle, thread and stitch length. Those are the variables you control, and they are the ones that recur.
Last reviewed 2026-08-11