# SPC interview Questions

## SPC (Statistical process control)

### What is SPC?

SPC stand for Statistical process Control. It is Quality tool which is used to measure and control process variation by using statistics. Statistics here is set of information derived from sample data.

Well SPC is Core tool of QMS, comprehensive tool published by AIAG. From interview point of view, SPC was introduced by William A shewart in 1924.

### What are Uses of SPC?

With SPC we understand process behavior. SPC is used as a tool to develop process, identifying source of variation in process & removing them to sustain better quality. SPC can be use as problem solving tool. By SPC we try to set those manufacturing parameters which give optimum value, by which cost of manufacturing will reduce.

### SPC formulas

So SPC is use to study process behavior. With SPC we develop our process of manufacturing, or not only develop but we strengthen our process by doing SPC. This core tool not only use in automotive sectors, but also use widely in other no

automotive fields, like medical, defense.

Process behavior is understand by shape, spread &Location. For Location we calculate mean. For Spread we calculate Standard deviation (sigma). For Shape we use help of Histogram.

SPC summary in short is explained in below picture:-

#### SPC Definition & Variation

SPC is comprises of three words **Statistical Process ****Control**. Each word has its meaning.

Statistical: – Interpretation of Data [which we have collected] to estimate __process__.

Process: – Convert input into output by __controlling__ man, machine, method & material.

Control: – Keeping __process __variation within spec.

Variation: – Gap/Dissimilarity between the two objects called variation. Variation can cause unwanted rejection, undesired rework, Customer dissatisfaction.

**Cause of Variation:–**

1. Common cause- Which will come & cannot be control/stop.

2. Special Cause- Which come sudden [Black Noise, Assignable cause]

**Total Variability:-**

**Process Behavior estimation on the basis of Variation:-**

**Note:-**Process variation is 6 time of standard deviation.

**Types of Histogram**

- Normal –Bell shape & Normal
- Comb like
- Positive or negative skewed
- Precipice type
- Plateau type
- Bimodal
- Isolated Peak Type

**Normal distribution shape is considered OK while doing process study on the basis of process ****distribution****, in normal distribution **process variation is within Control limit**.
**

### Types of SPC charts

- X-bar & R chart
- X-bar & S chart
- I & MR chart
- P-chart
- np &p chart
- u-chart
- c-chart & u-chart

In variable control chart-Most commonly used chart that come in picture is X(bar) & R chart.

For X(bar) & R chart , we have to collect variable data & have to decide followings:-

- Decide Subgroup Size
~~Decide subgroup Frequency~~[Don’t confuse-It is just time interval]- Decide number of Sub-group

Sub-group in variable control table |

- Calculate average of each subgroup. (X-bar)
- Calculate range of each subgroup. (R=Xmax -Xmin)
- Calculate average of average of each subgroup (X-Double bar)
- Calculate average of Range (R-bar).
- Calculate trial control limit for Range chart. (UCLR, LCLR)
- Calculate control limit for Average chart. (UCLx, LCLx)

Now we have done all calculation for UCL & LCL for X̅ & R.

We decide subgroup size n=5, subgroup number 20.

Calculation for control chart in SPC |

To create X̅ Chart, we will hide all Range row, and will select X̅ data and will insert line chart.

Same procedure will be followed to create Range chart.

Range chart in SPC |

By control Chart we identify any special cause in process, see unnatural pattern in data.

Any special cause could be identify when observation cross limit lines.

### Control chart Interpretation

- One point beyond zone A- Caused by large change in process
- Seven point in a row on one side of center line-Caused by process mean shift
- Seven point in row steadily increasing or decreasing-Caused by Mechanical wear, contamination, Chemical depletion, etc
- Fourteen point in a row alternating up & down-Caused by over adjustment, shift to shift variation, machine to machine variation.
- Two out of three point in row in same zone or beyond- Caused by Major special Cause variation
- Four Out of five point in a row in same zone B and beyond.
- Fifteen point in a row in zone C (Above or below centre line)
- Eight point in a row on both sides of center line with none in Zone-C

### Difference between Specification Limit and Control Limit

Specification limit are product tolerance limit, usually taken from product drawing. Example if diameter of hole in drawing is 8+1/-2 (Min- 6 mm and Max-9mm) this is product specification limit. While control limit is related to process control limit.

**There is no relation between specification limit and control limit.**

Control limit is based upon mean and standard deviation. Control limit will be within process width. Process width is mean+3 x standard deviation. Control limit is limit of process, beyond this the variation will come that will affect product quality somewhere.

Cp and Cpk is calculate to study the process capability and process performance.

### Cp & Pp – Process Capability & Process Performance

Process Capability (Cp & Cpk) indicates the ability of process to meet specification when process operates under common cause.

In practical situation, it is obvious to show variation due to both common and assignable cause. So we have

to analysis process behavior due to combined effect of common and assignable cause. We indicate by Pp, Ppk.

### Process Capability= Cp

Cp is for process capability. Which is study by process graph. Cp always show spread of curve. Process will be capable if curve is smaller to specification limit and equally distributed to both side of mean.

- Calculate process standard deviation (σ=R̅/d2)
- Calculate process Capability (Cp)= (USL-LSL)/6σ

σ= 0.21/2.326 ; (0.090)

Cp=(0.9-0.5)/6*0.090;

Cp=0.754

USL & LSL are maximum limit & minimum limit for spec:- 0.7±0.2

Maximum:- 0.9 & Minimum-0.5

Cpk is for Process performance. It show spread as well as location.

Minimum value between Cpu & CpL is consider as Cpk

Cpu=(USL-X̿)/3σ

CpL= (X̿-LSL)/3σ

Cpu= (0.9-0.718)/3*0.09 ; Cpu=0.689

CpL= (0.718-0.5)/3*0.09; Cpu=0.825

*Here Cpk value will be consider 0.689 (Since is minimum)*

Cp is consider only spread, not the location. While Cpk is consider both spread and location.

Thanks for sharing this, very informative