# control limits

Six Sigma – iSixSigma Forums Old Forums General control limits

Viewing 9 posts - 1 through 9 (of 9 total)
• Author
Posts
• #37556

paul cohen
Participant

Can somebody please help i have 30 attribute readings 1 for each day.Because i dont have any sub groups i dont understand how to work out the control limits the readings are as follows

192

239

246

231

218

214

201

242

226

214

196

207

229

192

203

195

221

245

219

211

217

205

209

205

228

221

250

255

205

0
#110841

Ken Feldman
Participant

What exactly are these measurements of?  Are they defectives taken from a constant sample group (use a np chart)?  Are they defects taken from a constant lot (use a c chart)?  You say they are attribute otherwise you could probably use an I/MR chart if they were continuous measures.

0
#110851

[email protected]
Participant

Paul
Why not just take that 30 samples and divide it by your preferred sample size i.e. 5 or 4 etc. You then say take a sample of 5 over 5 days or 4 over days. The statistical tables dont have a sample of 1 so I dont know how you would then solve this problem. However if you the do as i suggested above, then for future charting you will have to pull the 4 or 5 etc samples every 5th or 4th day etc. I am not sure if this is of any help to you.
Johnny

0
#110865

Ken Feldman
Participant

Whoa Johnny, bad advice.  You don’t create rational subgroups by artificially grouping single data points.  They are called rational subgroups because there was a reason for collecting them that way.  Assuming they were actually individual data points, you use an I/MR chart with the table factor for n=2 since as you point out, there is no factor for 1.  The thinking being that in order to get the Moving Range you take the difference between two consecutive values.  Let’s keep it simple and not ask why the table has higher values as well.  Plus, Paul said the data was attribute values so we don’t really know what they are.  I asked him whether they might be defectives or defects.  That would allow us to use np and c charts.

0
#110866

Whitehurst
Participant

Paul, you say you have attribute data (Go/no-go, pass/fail) but then you provide numbers.  This looks like variables data.  Please tell us: what the measurements are of, over what time period the data was taken, and how many samples were taken at a time.  We can then provide the calculations.
Thanks,

0
#110868

Sorour
Participant

The readings are of air pressure which is recorded once a day over a thirty day period.one reading is recorded each day.
Paul

0
#110869

Ken Feldman
Participant

0
#110899

paul cohen
Participant

my e-mail is [email protected]

0
#110902

Johnny Guilherme
Participant

You are right. Going back to the original posting I noticed Paul mentioned attribute type data and not variable data. I saw the numbers he posted and assumed variable data. Obviously then using np or c type charts would be the answer. Thanks for pointing this out to myself. I am still on a learning curve, having done some quality engineering and having been exposed to a lot of the tools but lacking application.
regards
Johnny

0
Viewing 9 posts - 1 through 9 (of 9 total)

The forum ‘General’ is closed to new topics and replies.