Is GR&R really practical for a Measurescope?

WEAVER

Starting to get Involved
#1
I always get bad GR&R results for our high power mesurescope which have 1 micron resolution measuring 60 to 70 microns with tolerance of +-6 microns. is a passing gr&r result for this even possible?(not to mention the ndc). How do i select the samples? should it come from different production runs-1 machine, or from different machines? pls help :tg:
 
Elsmar Forum Sponsor

ScottK

Not out of the crisis
Staff member
Super Moderator
#2
Re: is GR&R really practical for Measurescope?

I would say that if you are basing product acceptance on it, then you should seriously consider it.

If you are using it observation and not acceptance then I probably wouldn't.
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
#3
Re: is GR&R really practical for Measurescope?

The samples should represent the full variation that the gage should expect to see for that characteristic.

When you are starting out at 12:1 (tolerance to resolution) you do not have a lot of room for measurement and gage error. Make sure each operator measures the exact same spot, or your within-part error (not the fault of your gage) will wipe out your gage R&R. Also, very small variation presented to the process will assure low PV and poor gage R&R results.
 

WEAVER

Starting to get Involved
#4
Re: is GR&R really practical for Measurescope?

Yes sir we are using it to measure our product...The product tolerance is quite large (relative to the variation of the product) but when I try to conduct GR&R study, the tolerance seems to be small in relation to the resolution of the instrument. Also, would it help if we increase the magnification to make sure that the exact same spot can be measured? Or maybe the gage is not suitable for our measurement system requirements after all.
 
Last edited by a moderator:

WEAVER

Starting to get Involved
#5
Re: is GR&R really practical for Measurescope?

I would say that if you are basing product acceptance on it, then you should seriously consider it.

If you are using it observation and not acceptance then I probably wouldn't.
yes sir, it is used for product acceptance
 

WEAVER

Starting to get Involved
#6
Re: is GR&R really practical for Measurescope?

The samples should represent the full variation that the gage should expect to see for that characteristic.

When you are starting out at 12:1 (tolerance to resolution) you do not have a lot of room for measurement and gage error. Make sure each operator measures the exact same spot, or your within-part error (not the fault of your gage) will wipe out your gage R&R. Also, very small variation presented to the process will assure low PV and poor gage R&R results.
yes this is what always happens--it always seem to require nearly zero variation before it could pass (something that never happens in real life)...my question then would be: how can i improve the gage system so we could comply realistically and truthfully with this requirement (however ambitious it may appear)
 

Bev D

Heretical Statistician
Staff member
Super Moderator
#7
measuring microns with a resolution of a micron is tough.

However, there are ways.

can you post one or two of your analyses (in EXCEL so we can play with the data)
 

Bev D

Heretical Statistician
Staff member
Super Moderator
#9
Your issue is that you have 'chunky data'. Meaning very little resolution (as you know) so there are very few possible values. you have enough measurement error that this chunkiness causes an inflation in the standard deviation. at 60 um +/- 6 um you have a fairly tight tolerance. Also using only 10 parts makes the situation worse. you should use 30 parts measured twice. (I know this is contrary to the AIAG manual but just because something is in a manual doesn't make it right)

This doesn't mean that GR&R is not appropriate* - it can tell you a lot about your process if you know how to look at the data.

See my attached analysis using only the first 2 repeated readings for each operator. I plot this on a Youden Plot. Measurement error is the variation that exists perpendicular to the 45 degree line. (if there were no measurement error all of the values would fall on the 45 degree line as the second reading would equal the first.

You will have some difficulty at the upper and lower specs distinguishing the acceptable parts from the unacceptable but if you stay fairly centered this won't be an issue.

One question I would ask is what kind of 'ball' are you measuring? it is possible that it is perfectly round and that is contributing to your 'measurement error'; actually within part variation masquerading as measurement error...

*well Ok the AIAG automotive requirement is about as useless as - you know what - because too many people take a cookbook/rubber stamp approach to it.
 

Attachments

WEAVER

Starting to get Involved
#10
Your issue is that you have 'chunky data'. Meaning very little resolution (as you know) so there are very few possible values. you have enough measurement error that this chunkiness causes an inflation in the standard deviation. at 60 um +/- 6 um you have a fairly tight tolerance. Also using only 10 parts makes the situation worse. you should use 30 parts measured twice. (I know this is contrary to the AIAG manual but just because something is in a manual doesn't make it right)

This doesn't mean that GR&R is not appropriate* - it can tell you a lot about your process if you know how to look at the data.

See my attached analysis using only the first 2 repeated readings for each operator. I plot this on a Youden Plot. Measurement error is the variation that exists perpendicular to the 45 degree line. (if there were no measurement error all of the values would fall on the 45 degree line as the second reading would equal the first.

You will have some difficulty at the upper and lower specs distinguishing the acceptable parts from the unacceptable but if you stay fairly centered this won't be an issue.

One question I would ask is what kind of 'ball' are you measuring? it is possible that it is perfectly round and that is contributing to your 'measurement error'; actually within part variation masquerading as measurement error...

*well Ok the AIAG automotive requirement is about as useless as - you know what - because too many people take a cookbook/rubber stamp approach to it.
thanks so much!!! i'll try to answer your question in a bit (as soon as my nose stops bleeding) :D
 
Thread starter Similar threads Forum Replies Date
D Is PMCF really a continuous activity per Annex XIV,Part B? EU Medical Device Regulations 5
N IPC-A-630 - Is this free or do i really need to pay for it? Manufacturing and Related Processes 4
A Touch current in single fault conditions test and earth leakage current in normal conditions test, are they really different tests? IEC 60601 - Medical Electrical Equipment Safety Standards Series 8
D As a newcomer to QMS, I really appreciate how ISO is set up. ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 3
E Is it possible to start an ISO 9001 QMS from scratch (Really) ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 13
K GDPR - Is it really necessary for the DPO(s) to be knowledgeable to Data Privacy Law? IEC 27001 - Information Security Management Systems (ISMS) 3
W Is the RPN (risk priority number) in the PFMEA really a RPN without the detectability ISO 14971 - Medical Device Risk Management 4
H CQA Exam - Do I really have to go five hours without a drink? Professional Certifications and Degrees 4
DMLqms Medical Device Expiry Date or Manufactured date - really? Other Medical Device and Orthopedic Related Topics 1
I Is risk acceptability really needed if all risks must be reduced as far as possible? ISO 14971 - Medical Device Risk Management 6
Q Really do they add value (Vision, Mission, Values)? ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 1
Q Do we really need a traditional Quality Manual? ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 6
J ISO 17025 testing laboratory - Receiving lab supplies - What really needs a C of A? ISO 17025 related Discussions 2
Q Is Medical Device 510(k) exempt or not really exempt? 21 CFR Part 820 - US FDA Quality System Regulations (QSR) 2
Marc What Airlines Really Charge (In 1 Simple Table) Travel - Hotels, Motels, Planes and Trains 6
H Is 5S really that good? Lean in Manufacturing and Service Industries 34
P What is a QMS (Quality Management System), really? ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 18
Chennaiite Do Attribute Control Charts really help? Statistical Analysis Tools, Techniques and SPC 9
K Customer Audit - Just a positive rant really General Auditing Discussions 2
M "Value", A product that really meant something to you Coffee Break and Water Cooler Discussions 15
Jim Wynne A Lame Phishing Email--Are we Really This Gullible? After Work and Weekend Discussion Topics 16
J Does a Lower Thermostat Setting really save Energy? Coffee Break and Water Cooler Discussions 11
AnaMariaVR2 What Do Flight Attendants Really Think of You? Travel - Hotels, Motels, Planes and Trains 0
N Does Debt Consolidation Really Help Your Credit Score? Coffee Break and Water Cooler Discussions 7
L CE Mark - but not really. What are consequences? CE Marking (Conformité Européene) / CB Scheme 3
T The best way to pull some meaningful stats from a really big project? Using Minitab Software 3
L Do I really need to be AS9100 certified? AS9100, IAQG, NADCAP and Aerospace related Standards and Requirements 12
B Is Immediate Action really necessary before a Corrective Action? Nonconformance and Corrective Action 7
B Is accreditation to ISO 17025 really necessary? ISO 17025 related Discussions 15
V Really Urgent: Error in Minitab Using Minitab Software 8
R What does the Programmable Electrical Medical System really mean? IEC 60601 - Medical Electrical Equipment Safety Standards Series 6
J Are these things really required in a document control procedure? Document Control Systems, Procedures, Forms and Templates 17
B What Have They Done to AS9101? Was this really necessary? AS9100, IAQG, NADCAP and Aerospace related Standards and Requirements 23
Q Does 21 CFR 820 really revise every year? 21 CFR Part 820 - US FDA Quality System Regulations (QSR) 4
Chennaiite Value Engineering - Does Value Engineering really add value? Preventive Action and Continuous Improvement 3
S Qantas Airbus incident - What really happened? AS9100, IAQG, NADCAP and Aerospace related Standards and Requirements 3
L When really deserve to compress the Control Limit as an Improvement? Statistical Analysis Tools, Techniques and SPC 9
Wes Bucey A really DUMB Nigerian phishing letter! "FBI" (USA) Email promising dire consequences Coffee Break and Water Cooler Discussions 9
H Less Really is More When it Comes to Response Scales The Reading Room 9
P Is the author of ?How to Become a Billionaire?? really wealthy? Coffee Break and Water Cooler Discussions 6
M Is PQ really needed after relocation of a Visual CMM General Measurement Device and Calibration Topics 2
A Would anyone really fall for this scam? Coffee Break and Water Cooler Discussions 32
P ISO 9001 for a REALLY small business ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 31
I Core Processes - Are Core Processes really needed in the quality manual Quality Management System (QMS) Manuals 4
A During tough economic times, does 'ISO' really help Coffee Break and Water Cooler Discussions 27
Q Continuous Improvement Form - Really needed? IATF 16949 - Automotive Quality Systems Standard 3
S Do Key Performance Indicators (KPIs) really make a difference in the QMS? ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 9
C Employee Certifications based on ISO standards: What does it really mean? Training - Internal, External, Online and Distance Learning 11
J 5S and Moving Forward - Is 5S really working here? Lean in Manufacturing and Service Industries 9
BradM Blue man group is really going to be blue World News 1

Similar threads

Top Bottom