Can there be a mix of good and bad parts for a variable GR&R?

QE1993

Involved In Discussions
#1
I know there is a general "rule" that a certain percentage of parts have to be non-conforming for attribute studies, which makes sense.

Can there be a mix of good and bad for variable studies, or do you have to have all good parts?

Can variable studies be done on ALL good or ALL bad parts?

Thanks in advance.
 
Last edited:
Elsmar Forum Sponsor

Miner

Forum Moderator
Staff member
Admin
#2
re: Can there be a mix of good and bad parts for a variable GR&R?

It depends on the purpose for which you will use the gage.
  1. Inspect product to tolerance - It doesn't really matter how you select your parts. The part variation is not part of the analysis.
  2. Measure product for SPC or capability studies - Product must be selected randomly to fully represent the typical process variability. Deliberate selection of product will bias the results.

For more in-depth discussion see my MSA blogs:Part 5a R&R and Part 4 How to Sample Parts.
 

Ron Rompen

Trusted Information Resource
#3
re: Can there be a mix of good and bad parts for a variable GR&R?

For a GR&R study, you should select parts which represent the range of your process. With that being said, you should also include known 'bad' parts, so that you can validate that your process controls will, indeed, detect them.
 

QE1993

Involved In Discussions
#4
re: Can there be a mix of good and bad parts for a variable GR&R?

So, are you saying I should deliberately select parts that span the process control limits or will randomly selecting production parts ensure (for the most part) that I am representing the appropriate range?
 

Ron Rompen

Trusted Information Resource
#6
A good question, Golfman - in some cases you can 'force' the process to make parts which deviate from the nominal (i.e. machining). In other cases, you may have to artificially manipulate parts to create the non-nominal value (secondary processes which are not part of your normal process).

The intent is to ensure that the gauge you are using is capable of detecting and discriminating between parts that are both good and bad.
 

Miner

Forum Moderator
Staff member
Admin
#7
For a GR&R study, you should select parts which represent the range of your process. With that being said, you should also include known 'bad' parts, so that you can validate that your process controls will, indeed, detect them.
This is only true for attribute studies or safe where the gage will solely be used for inspection to a tolerance. If the gage will be used for SPC or capability studies, this practice will bias the results.
 

stevegyro

Involved In Discussions
#8
What Mr. Rompen says[emoji846].

GRR summaries depend on known process variation, or at least the desire to understand what your (most important) process behave like.

Except for “making money”, what is your core competency? What measurement system do you have in place to validate that your ‘competency’ is CONSISTENT part after part, day after day, ... year after year?

Great, you likely have that ‘measurement system’. But how do you check that?

Now to the point:
Use Minitab help and ‘Assistant’ to guide your GRR, which is a process in itself.

The GRR will compare variance for each part’ measurement results, x (5 or 10 pieces)

It will also compare (variances of measurement) ‘within’ each person, for each part, and across all parts.

You SHOULD choose parts representing a range meeting or exceeding your process variation.

The results cannot be guessed or predicted, which is the ‘power of the GRR test’ when a range of test pieces/events is used.

If all test pieces are identical, you are wasting your time, and the company you work for’ time.

The NDC no of distinct categories is a requirement, bearing the need for variation in your test samples.

Use MiniTab’ (or whatever Stat package you prefer).
If you think Stat software is (too)costly, then one of two things is true:
1. You could be missing out on all the ‘tutorials’, as well as ‘live’ tech support from your software vendor; or
2. You are not using a software designed specifically for Stats and GRR’s.

I have seen many companies use an excel template for GRRs, to ‘cheap out’ in my opinion.
Where did that ‘template’ come from?
Is it protected? Is it validated?

Not anything wrong with Excel, just saying it is a number cruncher by design and intent. A Stat package is a specific tool, with ability to guide your people, and tell you if a test is “weak” or “strong “

Best,
Steve Giarratana




Sent from my iPhone using Tapatalk
 
Thread starter Similar threads Forum Replies Date
Q Is it acceptable to mix components from two different lots into an assembly? ISO 13485:2016 - Medical Device Quality Management Systems 1
B Determing Machine Capability in a High Mix, Low Volume Sheet Metal Shop Manufacturing and Related Processes 4
B Mix between two Tablets Batch of the same Product Manufacturing and Related Processes 7
R Mix Multi-Site/Individual Certification - Has this ever happened to anyone? Miscellaneous Environmental Standards and EMS Related Discussions 14
T Packaging Box Mix Joint - Corrective Action help Manufacturing and Related Processes 9
R Paint Procedure on How to Mix Paint and Water wanted Human Factors and Ergonomics in Engineering 2
Howard Atkins Marketing and Quality Mix Misc. Quality Assurance and Business Systems Related Topics 3
Hershal Want to mix it up in political and similar discussions? Coffee Break and Water Cooler Discussions 2
R Control on Setting piece mix-up - Are there any fool proof systems? Quality Tools, Improvement and Analysis 2
G Idea To Prevent Mix-up of Medical gases Preventive Action and Continuous Improvement 7
kedarg6500 Potential Failure Mode - Mix Charge for Hardness - Casting Process FMEA and Control Plans 1
N Painting experts: Coverage area for each litre of paint and thinner mix Manufacturing and Related Processes 5
J Product Mix - Routing Problem - LINGO or EXCEL Excel .xls Spreadsheet Templates and Tools 2
F My boss wants a mix of Procedure and Form - Forklift and Manlift Equipment Document Control Systems, Procedures, Forms and Templates 5
J Product mix problem using excel .xls - Can anybody throw in some light? Excel .xls Spreadsheet Templates and Tools 5
B Mix of leaded & leadfree components - Affected Solder Joints Quality - Vibration Manufacturing and Related Processes 3
Marc Twin Mars rovers - Instrument mix-up - Instruments installed in the wrong rovers World News 2
D Limiting Scope of TS 16949 Certification - Mix of Automotive and Non-Automotive IATF 16949 - Automotive Quality Systems Standard 9
M Mix(ed) up Procedure ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 6
B A mix of plastic and solid colorant - 'Gage' R&R on a process Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
T Short Lead Times - APQP and PPAP - Commercial Product Mix APQP and PPAP 10
E Determining what is good and what is bad can be subjective - when is it a quality issue? ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 9
optomist1 Aircraft Cockpit Automation - Good or Bad Human Factors and Ergonomics in Engineering 71
L How to Introduce the Quality Function to the production floor in a bad environment Misc. Quality Assurance and Business Systems Related Topics 6
V Inference of bad predicted-R-Sq - But very good R-Sq and Adj-R-Sq. Using Minitab Software 2
P Bad communication in construction projects are probably the biggest generators Quality Manager and Management Related Issues 4
A How can a sample can give information about bad products in whole lot? Quality Manager and Management Related Issues 3
D Requiring a "bad" part for attribute MSA studies Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 3
L Correct Sample Size for Attribute Gage R&R (Good / Bad parts) Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 12
D MSA (Measurement System Analysis) on a Bore Measurement - Good or bad result? Six Sigma 9
T Zinc Phosphate Test Panels look bad - Need Help Manufacturing and Related Processes 2
B How do you measure whether a QMS is good or bad? Quality Manager and Management Related Issues 9
G Cpk vs. "Bad" Parts - Unreasonable Diameter Tolerance and a 1.67 Cpk Statistical Analysis Tools, Techniques and SPC 22
H Correcting Bad Process Habits Quality Manager and Management Related Issues 11
B How to Correct Bad Employee Performance-Attendance and/or Tardiness Quality Manager and Management Related Issues 84
Q Employees perceive CAPA as Bad - How to cope with the situation Nonconformance and Corrective Action 15
M Cost of Quality - Is it good or bad and why - Critical Review Research Philosophy, Gurus, Innovation and Evolution 28
I Do we need GOOD and BAD Wire Harness "Dummy" Samples? IATF 16949 - Automotive Quality Systems Standard 3
J Easy Method to Test a Caliper - Good or bad? General Measurement Device and Calibration Topics 10
P VDA 6.3 - Process Audit - You think you had it bad? VDA Standards - Germany's Automotive Standards 35
O How many bad parts needed for Attribute MSA? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 5
F Incentives for blowing the whistle on bad quality? Philosophy, Gurus, Innovation and Evolution 11
Colin Bad news/good news day Coffee Break and Water Cooler Discussions 22
J Bad driver? Blame your genes Coffee Break and Water Cooler Discussions 7
Wes Bucey Update: State of mind during the job hunt (Good habits vs. bad habits) Career and Occupation Discussions 12
Pancho Dvorak keyboard - one cure for RSI and bad typing habits Imported Legacy Blogs 1
H My job sucks so bad Funny Stuff - Jokes and Humour 20
Wes Bucey Agencies and Agents - good guys or bad guys? Career and Occupation Discussions 9
R Sampling - Total probability of accepting bad lot - Producer's risk Inspection, Prints (Drawings), Testing, Sampling and Related Topics 2
bobdoering You know things are gettting bad: US anti-kidnapping expert kidnapped in Mexico Coffee Break and Water Cooler Discussions 2
Similar threads


















































Top Bottom