ISO 7870-1-2014 pdf free download

ISO 7870-1-2014 pdf free download.Control charts — Part 1:General guidelines Cartes de con trôle — Partie 1: Lignes direct rices générales.
5.6.4 Subgroup selection
Subgroups are samples of collected items obtained from the process in a defined manner. Data from characteristics of these items are determined from which statistics, such as a number ofnonconformities or an average or a range, can be computed and plotted on the control chart.
Rational samples or subgroups should be selected in a manner that makes each subgroup as homogeneous as the process will permit. Within a rational subgroup, variation is presumed to be due only to random causes. These causes are sources of variation inherent in a process over time. Rational subgroups are selected to enable the detection of any special causes of variation between subgroups. Short-term variability is measured using the variability within a series of reasonably homogeneous subgroups and determines the position of the control limits on the control chart, while long-term variability is usually evaluated in terms of changes between subgroups. Time order is often a good basis for forming subgroups because It allows for detection of special causes that can occur over time. However, other bases, such as the need to study operator-to-operator variability, machine-to-machine variability, or supplier-to-supplier variability can suggest that subgroups be defined across operator, machine, or supplier instead of over time.
The rational subgroup should be subject to all usual sources of random cause variation if it is to have meaningful value. For example, a series of repeat readings on a piece of material set in a testing instrument will fail to include the contribution of locating the material in the instrument or of obtaining the sample. If these aspects were inherent in a usual testing environment, the repeat readings would give an unrealistic, low estimate of inherent measurement variability. Thus, almost any actual measurement from the process would appear to be out of control’.
5.6.5 Subgroup size
Subgroup size should be selected so as to balance the ability to detect small shifts in the process and the risk of not detecting special causes. A larger subgroup size, although more costly, will provide a more precise assessment of the process, therefore allowing for a more efficient monitoring. However, if the sample is too large, special causes have more opportunities to occur within the collection period of the sample, causing increased within-sample variation; hence, control limits might be unduly widened and many special causes can occur without detection.
When dealing with attributes data, the subgroup size needed to detect changes in the process proportion will ordinarily be very much larger than the subgroup size using variables data, since attributes data carry much less information than variables data.
In some situations, it is impractical, or it does not make sense, to form subgroups, but rather to collect information on individual units, so that, essentially, subgroup size is equal to one. This is the case when testing is destructive or sampling is costly, or when repeated measurements on the process (continuous or batch processes) differ only because of instrument or analysis error.ISO 7870-1-2014 pdf free download.