How to calculate standard time
A time study turns stopwatch readings into a time you can plan with. Three steps, element by element:
Observed time (OT) = average of the valid readings
Normal time (NT) = OT × rating
Standard time (ST) = NT × allowance factor, with factor = 1 ÷ (1 − allowances)
| Term | What it is |
|---|---|
| Element | a part of the operation with a clear start and end, like “pick and place” or “screw” |
| Rating | the observed operator’s pace against normal (100%) |
| Allowances | personal needs, fatigue and unavoidable delays, as a % of the working day |
| Allowance factor | with 15%, 1 ÷ 0.85 = 1.176 |
Mind the allowance convention. This calculator (and TaktZen) treats allowances as a share of the working day: factor = 1 ÷ (1 − p). Some textbooks use factor = 1 + p, treating them as a share of normal time. At 15% the first gives 1.176 and the second 1.15. Stick with the convention your plant already uses.
Output per hour at standard time is 3,600 ÷ ST (in seconds).
How many cycles to time
A few readings give a shaky average. The number of cycles needed depends on how much the readings vary, on the confidence and on the error you accept:
n = ((z ÷ e) × √(n·Σx² − (Σx)²) ÷ Σx)²
where z comes from the confidence (1.645 for 90%, 1.96 for 95%, 2.576 for 99%), e is the relative error (0.05 for ±5%), n is the number of readings taken and x the readings. This is Barnes’ statistical formula, the same one TaktZen uses. Compute it for each element and keep the largest.
Some textbooks use the range method instead: N = (z × R ÷ (e × d2 × x̄))², with R = highest reading − lowest reading and d2 from a table by the number of readings (3.078 for 10). Both usually land close together, and the calculator shows both.
Worked example
The calculator’s values are 10 cycles of an assembly operation with 3 elements, a 100% rating, 15% allowances, 95% confidence and ±5% error.
1. Readings out of range. For “screw”, the 10 readings average 12.56 s with a standard deviation of 1.50 s. The mean ± 2 SD band runs from 9.55 s to 15.57 s, so the 16.5 s reading in cycle 5 is out (a dropped screw, say). It is left out, and the element keeps 9 readings.
2. Observed, normal and standard time.
| Element | Readings | OT | NT (× 100%) | ST (× 1.176) |
|---|---|---|---|---|
| Pick and place | 10 | 4.40 s | 4.40 s | 5.18 s |
| Screw | 9 | 12.12 s | 12.12 s | 14.26 s |
| Inspect and pack | 10 | 6.19 s | 6.19 s | 7.28 s |
| Cycle | 22.71 s | 22.71 s | 26.72 s |
3. Output per hour. 3,600 ÷ 26.72 = 134.7 units per hour.
4. Cycles needed. For “pick and place”: Σx = 44.0, Σx² = 194.4 and n = 10, so n·Σx² − (Σx)² = 1,944 − 1,936 = 8, and n = ((1.96 ÷ 0.05) × √8 ÷ 44)² = (39.2 × 2.83 ÷ 44)² = 6.35, so 7 cycles. The other elements need 4 and 6. The largest is 7, and every element has at least 9 valid readings: enough cycles.
Had the 16.5 s reading stayed in, “screw” alone would call for 20 cycles. One abnormal reading moves the math a lot, which is why it pays to note what happened in each cycle.
And if the operator had been faster than normal, rated at 110%, the normal time of “pick and place” would be 4.40 × 1.10 = 4.84 s, and its standard time 4.84 × 1.176 = 5.69 s.
How to read the result
- Standard time is the number that goes into planning: use it for line balancing, to compare against takt time and for production capacity.
- Cycles needed above cycles taken means take more readings before publishing the time. If one element drives the number up, check that its method is well defined.
- Readings out of range flag something abnormal in that cycle. Leave them out only when you know why. If they repeat, they are not exceptions.
- A large deviation on one element usually points to an unstable method, fuzzy start and end points, or an element that is too long.
Common mistakes
- Timing the whole cycle as one reading. Without elements you cannot tell where the variation comes from.
- Skipping the rating. Timing your fastest operator at 100% produces a standard nobody else can meet.
- Counting allowances twice. If the coffee break is already out of the available time, it does not go into the allowances again.
- Throwing out high readings because they are inconvenient. A delay that happens every day is a real loss and should show up in the analysis.
- Using standard time as the OEE ideal cycle. Standard time includes allowances. For OEE performance, use the ideal cycle of the machine or station.