Quinn-Curtis Forums
Quinn-Curtis Forums
Home | Profile | Register | Active Topics | Members | Search | FAQ
Username:
Password:
Save Password
Forgot your Password?

 All Forums
 Tools for Microsoft .Net & .Net Compact Framework
 SPC Control Chart Tools for .Net
 GetCalculatedValueRecord(1) returning wrong value
 New Topic  Reply to Topic
 Printer Friendly
Author Previous Topic Topic Next Topic  

blazermaniac

USA
27 Posts

Posted - 21 Jul 2016 :  18:52:26  Show Profile  Reply with Quote
Given the primary points 1.155 and 1.141 a call to (double)this.ChartData.GetCalculatedValueRecord(1).CalculatedValues.GetValue(CalculatedValueIndex) returns 0.014000000000000012 while I expected 0.014. The control limit is 0.014 which results in a violation. I found that this is an artifact of using doubles (base 2). I have cast the arguments in SetCalculatedValue, case SPC_INDIVIDUAL_ABS_RANGE_CALC as decimal (base 10) to fix the problem.

If you have a better solution please share it with me.

Thanks!

quinncurtis

1551 Posts

Posted - 22 Jul 2016 :  08:59:25  Show Profile  Reply with Quote
Statistically, as far as the process goes, I don't think your change will have any affect on the overall control of the process. SPC Control Limits are not perfection. Control limit violations are only strong (not infallible) hints that the process is out of control. There will always be false positives and false negatives. If your measured variable tests within 0.0000000000000001 percent (the floating point accuracy of a double) of a control limit, then assuming that the process is out of control is a far better assumption to make than that it is in control. That's the way a Quality Control Engineer would look at it, as opposed to a programmer. Look at it another way - if you are a doctor, and you were taught in school that a fever of 106 degrees F will most likely kill your patient, but the patient only has a fever of 105.99999999999999999 F, do you ignore the fever, or treat it as if it were life threatening? Just my opinion though.
Go to Top of Page

blazermaniac

USA
27 Posts

Posted - 22 Jul 2016 :  12:06:07  Show Profile  Reply with Quote
I understand your point but on the chart we are reporting a false negative and when you look at the alarm notes it appears the limit and value are equal. For our implementation we compare the values and if OOC we perform some various actions which include at a minimum sending an email alert and up to putting a lot/tool on hold.

In our case at least the code change is necessary.

Thanks,
Brian
Go to Top of Page
  Previous Topic Topic Next Topic  
 New Topic  Reply to Topic
 Printer Friendly
Jump To:
Quinn-Curtis Forums © 2000-2018 Quinn-Curtis, Inc. Go To Top Of Page
Powered By: Snitz Forums 2000 Version 3.4.07