A pre-control chart is a simple go/no-go method that splits a tolerance into a green middle half, two yellow warning quarters, and red beyond the spec limits, then uses a handful of sampled parts to decide whether to keep running or stop and adjust. It needs no control-limit math and no charting software, which is why it survives on setups and short runs.
Pre-control trades statistical rigor for speed and simplicity. It compares parts to the specification, not to the voice of the process, so it is easy to teach an operator in ten minutes and easy to misuse on a process it was never meant for. This guide covers the zones, the setup and running rules, exactly where pre-control earns its place, and the cases where a real control chart is the honest answer.
How does a pre-control chart divide the tolerance?
Pre-control divides the distance between the two specification limits into four equal quarters. The middle two quarters together form the green zone, the go zone, which covers the central half of the tolerance. The two outer quarters, between the green zone and each specification limit, are the yellow zones, the warning zones. Anything past a specification limit is the red zone, the reject zone. The two lines separating green from yellow are the pre-control (PC) lines, set at one quarter and three quarters of the tolerance.
That last point matters. Because the zones are cut from the specification, pre-control assumes the tolerance is a reasonable match to the process spread. If the process is roughly capable and centered, most parts sit in green and a yellow reading is a genuine early warning. If the process is barely capable, yellows fire constantly for no reason; if it is wildly capable, everything is green and pre-control tells you nothing.
How do you set up a pre-control chart?
Pre-control has two phases: a qualification phase to confirm the setup is good, then a running phase of periodic two-part checks. The whole method is designed to be run by the operator at the machine.
- Draw the zones. Mark the two PC lines at one quarter and three quarters of the tolerance, which defines the green, yellow, and red zones for the characteristic.
- Qualify the setup with five in a row. Measure five consecutive parts. If all five fall in the green zone, the setup is qualified and you may start the running phase.
- Reset on any warning during qualification. If a part lands in yellow or red before you reach five greens, adjust the process and start the count of five over. You are not allowed to start production until five straight land green.
- Run periodic pairs. During production, measure two consecutive parts at a set frequency and apply the running rules below to decide whether to continue or stop.
- Re-qualify after any adjustment. Any time you stop and adjust the process, run the five-in-a-row qualification again before resuming the pair checks.
A common rule of thumb sets the sampling frequency so that you take roughly six pairs between process adjustments. Sample too rarely and a drift slips through; sample too often and you are inspecting more than the method needs.
What are the running rules?
Once qualified, pre-control runs on pairs. Every so often you measure two consecutive parts (call them A and B) and read the pair:
- Both green: keep running. The process is comfortably inside the middle half.
- One green, one yellow: keep running. A single yellow with a green is normal variation, not a signal.
- Two yellows, same side: stop and adjust. Two parts in the same yellow zone say the process has shifted toward that limit.
- Two yellows, opposite sides: stop and investigate. Parts hitting both yellow zones say the spread is too wide for the tolerance, which adjustment will not fix.
- Any red: stop and adjust immediately, and address the out-of-spec part. A red is a reject.
Where does pre-control fit, and where does it fail?
Pre-control shines in a narrow band of situations: short production runs where there is no time to build control limits, setup verification before a run, low-volume or job-shop work, and processes where the tolerance is comfortably wider than the spread. It gives an operator a fast, visual, no-math way to catch a bad setup or an obvious drift.
It fails, sometimes quietly, outside that band. Because pre-control reads the specification instead of the process, it cannot tell you whether the process is stable, cannot separate common-cause from special-cause variation, and cannot measure capability. A pair of two parts is a small sample with weak power to catch a modest shift. On a marginal process it triggers constant false adjustments, and every needless adjustment adds variation, the tampering that Deming warned about. And a very capable process passes pre-control while telling you nothing about the drift you would want to see.
| Question | Pre-control | Control chart |
|---|---|---|
| Limits based on | Specification (tolerance) | The process (control limits from data) |
| Math to run | None; visual zones | Compute and plot control limits |
| Detects process stability | No | Yes |
| Separates common vs special cause | No | Yes |
| Best for | Setups, short runs, low volume | Ongoing production, capability work |
The honest rule: use pre-control to verify a setup and to guard a short run, and move to a real control chart the moment the run is long enough to justify one or the moment you need to know whether the process itself is in control. Pre-control is a smoke detector, not a diagnosis.
A worked example: a turned shaft diameter
Say you are turning a shaft to a diameter of 25.00 mm with a tolerance of plus or minus 0.20 mm. That gives a lower spec limit of 24.80 mm and an upper spec limit of 25.20 mm, a total tolerance of 0.40 mm. Divide it into quarters of 0.10 mm each.
The green zone is the central half, from 24.90 mm to 25.10 mm. The lower yellow zone runs from 24.80 to 24.90, the upper yellow from 25.10 to 25.20, and the pre-control lines sit at 24.90 and 25.10. Red is anything below 24.80 or above 25.20.
You set the tool and measure five parts: 24.98, 25.01, 24.99, 25.02, 25.00. All five are inside the green band, so the setup qualifies and the run starts. An hour in, a pair reads 25.12 and 25.14. Both land in the upper yellow zone, on the same side, so you stop and adjust the tool back toward center rather than waiting for a reject. Had the pair read 24.88 and 25.13, one part in each yellow zone, you would stop and investigate instead, because a spread that reaches both edges of the tolerance is a variation problem that a tool offset will not fix. That is the entire method in one shaft: two numbers, a zone, and a decision, with no control-limit arithmetic anywhere.
How does pre-control relate to SPC and capability?
Pre-control is often described as a lightweight cousin of statistical process control but the two answer different questions. SPC control charts are built from the process's own variation and tell you whether the process is stable and predictable. Pre-control is built from the tolerance and tells you only whether recent parts are near the edge of the print. Neither one measures capability; for that you run a process capability study and compute the indices covered in Cp and Cpk. A sensible progression is pre-control to protect a setup, a control chart to hold ongoing production, and a capability study to certify the process against the spec.
The reference behind the method
Pre-control is a recognized quality technique documented in standard quality references, alongside the control charts it is sometimes compared to.
- Under normality with a centered, marginally capable process (Cp near 1), roughly 86% of parts fall in the green zone and about 14% in the two yellow zones, which is the statistical basis for the pair rules (ASQ, statistical process control).
- Classical Shewhart control charts, the rigorous alternative pre-control is measured against, are standardized in resources such as ASQ's control chart guidance and the ISO 7870 series on control charts.
Read those to see what pre-control gives up: the ability to hear the process, not just check it against the print.
Making pre-control checks actually stick
The weak point of pre-control on a real floor is not the rules; it is whether the checks happen and get recorded. A method this simple lives or dies on discipline: the five-in-a-row qualification skipped after a rushed changeover, the pair check that nobody logged, the yellow-yellow that got waved through because the shift was ending. When setup checks and in-process pairs are captured digitally at the station, the qualification is enforced before the run starts and every pair leaves a timestamped record, so a skipped check is visible instead of invisible. That is the kind of shop-floor discipline Harmony's quality intelligence and paperwork digitization is built to hold, whether a station runs pre-control on a short job or a full control chart on a long one. See it working in a real plant in our CLS case study.