Analytical Balances - Looking for Variance Source

G

grzesiek

#1
Ok, perhaps it will be easy to you but I need some basic help

I want to prove that in our analysis balances are the biggest source of variation. If it is of course. So really I want to find source of biggest variation.

The analysis is as follows - analyst weights the sample, does the solution by diluting the sample, and analyses it. There is a satndard of known potency weighted and analysed with the sample.

I can have 2 balances (I suspect both to work wrong), 2-3 analysts, 1 (or 2) equipment for measuring.

I would like to design a statistically sound study to find sources of variation. I think ANOVA is the way but ma not sure and also need help with design.

Do you need any more information? Does anything needs to be improved?

Please feel free to move the topic if this is not the place where I can get help. English is not my mother language so I have some difficulties in understanding this statistical concepts (I'm also not a statistician)
 
Elsmar Forum Sponsor

Marc

Hunkered Down for the Duration with a Mask on...
Staff member
Admin
#2
I moved this to the "Gage R&R (GR&R) and MSA (Measurement Systems Analysis)" forum. I *thik* it might fit in better here, but I'm open to suggestions. I also retitled the thread. Let me know if the title isn't correct.

I can't help with this myself, and things are slow here since it's a weekend (and Monday is a hoiday in the US), so if you don't get any feedback right away be sure to 'Bump' the thread on Tuesday morning.
 

Hershal

Metrologist-Auditor
Staff member
Super Moderator
#3
The balances may be the main contributor, but to absolutely assure that results are true, I would have the balances calibrated (again if they already are) and a thorough uncertainty study for the calibration, to include essentricity, barometric pressure, local gravity, and so forth. At that point you can state that the value of the substance you weigh is whatever the balance indicates, plus or minus the uncertainty value.

That in turn provides the error contribution of the balance that you can include in your analysis, and so allow you to more easily identify the true source of your variance.

Thanks for the post move, Marc.
 

Marc

Hunkered Down for the Duration with a Mask on...
Staff member
Admin
#4
Thanks for helping, Hershal! I personally appreciate your help now and over the years!
 

BradM

Staff member
Admin
#5
Good post from Hershal. I hope that my post works in tandem with his.

Ok, perhaps it will be easy to you but I need some basic help
Please... never assume anything is easy, and always be willing to ask for help. We all learn something everyday here. :agree1::bigwave:

FWIW... I would classify this as a moderate/challenging question.:yes:

I can have 2 balances (I suspect both to work wrong), 2-3 analysts, 1 (or 2) equipment for measuring.
Ok, let's start simple. Why do you think the balances are the issue? Are they? Are they identical? Are they positioned in identical manner? Is one more affected by draft/vibration and such than the other?

It's easy enough to check. Perform calibrations and repeated measurements on the balances and see if they are the same. Have one person use one set of weights (calibrated with reported uncertainty) and perform readings on the two balances. You can "eyeball" the readings and see if there is a difference.

If you wish, you could pick the appropriate statistical test and compare variance between the two sets of results from the balances and see if they are statistically different.

In the end, I might consider approaching this as minimizing the measurement error, using uncertainty analysis to account for all the different variances.


Here is a thread where I started an uncertainty analysis for a balance:

Uncertainty calculation - Estimation for a Mettler Toledo XP 56 balance at 20 grams

If you click on that thread, at the bottom there will be some more threads about uncertainty.

Hope something here helps.
 

BradM

Staff member
Admin
#6
The balances may be the main contributor, but to absolutely assure that results are true, I would have the balances calibrated (again if they already are) and a thorough uncertainty study for the calibration, to include essentricity, barometric pressure, local gravity, and so forth. At that point you can state that the value of the substance you weigh is whatever the balance indicates, plus or minus the uncertainty value.

That in turn provides the error contribution of the balance that you can include in your analysis, and so allow you to more easily identify the true source of your variance.

Thanks for the post move, Marc.
Exactly. We have a lab, tightly controlled for temperature and humidity. At each end of the lab is a balance used for pipettes. We get different results on one end than the other end of the lab.;)

My hunch is vibration and operator error. An uncertainty analysis is in the works for them.:yes:
 

Ajit Basrur

Staff member
Admin
#7
Ok, perhaps it will be easy to you but I need some basic help

I want to prove that in our analysis balances are the biggest source of variation. If it is of course. So really I want to find source of biggest variation.

The analysis is as follows - analyst weights the sample, does the solution by diluting the sample, and analyses it. There is a satndard of known potency weighted and analysed with the sample.

I can have 2 balances (I suspect both to work wrong), 2-3 analysts, 1 (or 2) equipment for measuring.

I would like to design a statistically sound study to find sources of variation. I think ANOVA is the way but ma not sure and also need help with design.

Do you need any more information? Does anything needs to be improved?

Please feel free to move the topic if this is not the place where I can get help. English is not my mother language so I have some difficulties in understanding this statistical concepts (I'm also not a statistician)
Welcome to the Cove, grzesiek :bigwave:

Some weeks back when I was googling, I had come across Quantifying Uncertainty in Analytical Measurement and hope will be useful to you; its heavy on statistics though !

Hope to have your active participation :)
 
G

grzesiek

#8
Thank you for all the replies, I even didn't expect such a massive support :)

I will try to respond in chronological order

"Let me know if the title isn't correct."
well it can be so there's no problem althought suspected sources of variation are balances (and operators) and I would like to break the variances of all measurment process into bits :)

would have the balances calibrated (again if they already are) and a thorough uncertainty study for the calibration, to include essentricity, barometric pressure, local gravity, and so forth.
balances are calibrated each day, we use 1mg, 10mg, 100mg and ifI remember 500 or 1000mg weights, but what is measured is something like 10.24 mg, that's the number I get, 2 decimals. How do I perform the study of pressure or local gravity effect on uncertainity?

Why do you think the balances are the issue? Are they? Are they identical? Are they positioned in identical manner? Is one more affected by draft/vibration and such than the other?
we tend to see results not stabilizing or stabilizing for too long. I really can't tell that's why I want to perform study to find out. they are standing next to each other. They are affected the same way.

In the end, I might consider approaching this as minimizing the measurement error, using uncertainty analysis to account for all the different variances.
ok, but what about finding if they are a problem or not?

Exactly. We have a lab, tightly controlled for temperature and humidity. At each end of the lab is a balance used for pipettes. We get different results on one end than the other end of the lab.
temperature and humidity can be a problem but remember I weight samples and standards in the same time, one after another

"Some weeks back when I was googling, I had come across Quantifying Uncertainty in Analytical Measurement and hope will be useful to you; its heavy on statistics though !"
- will read through it again :)
 
Last edited by a moderator:

BradM

Staff member
Admin
#9
well it can be so there's no problem althought suspected sources of variation are balances (and operators) and I would like to break the variances of all measurment process into bits :)
Yes! And the uncertainty analysis can help you with that. Start with the balances; calculate them. Next, perform some studies with the two operators on the balances. Finally, make sure you have estimated all the components to the uncertainty. Then, you have an amount (in mg or ml; whichever is more relevant for your studies) that the measurement process is contributing to your study.

balances are calibrated each day, we use 1mg, 10mg, 100mg and ifI remember 500 or 1000mg weights, but what is measured is something like 10.24 mg, that's the number I get, 2 decimals. How do I perform the study of pressure or local gravity effect on uncertainty?
Hmm... is your balances only used for your study? Maybe the calibration test points need to be down lower near your use range. Can these balances produce the required accuracy you are looking for?

Do you have good calibrations done on the check weights with a reported uncertainty for them? You are down in a pretty low range. Reported uncertainties on test weights become significant.:notme:

As far as the study, if you are interested in performing an uncertainty analysis for your balance, all these components will need to be estimated. Have you been writing down the as found/as left values when you calibrate the balances? From that you can ascertain what kind of variation the balance is producing. If you are wanting to keep your uncertainty analysis simple, you can assume that many of the variances present would be accounted for during your daily checks.

we tend to see results not stabilizing or stabilizing for too long. I really can't tell that's why I want to perform study to find out. they are standing next to each other. They are affected the same way.

ok, but what about finding if they are a problem or not?
Well, you can take the readings from the two balances and perform a statistical comparison. Depending on your confidence interval, you will find significance. What then?:D However, if you perform an uncertainty analysis, you have a specific amount of error for each balance, and you can then determine if the errors are significant or not for your process.

As I mentioned, I know I have variance between my two balances, but I am pretty sure the difference is a significant amount (in micrograms). I have more variance in the pipetting process that will make more of a difference (in milligrams).

Just off the top, I'm not sure how precise you desiring here, but those balances may be contributing up to .05 milligrams (ore more).

Ask the users of the balance if they ever see any swing in the balances. Are these in draft-controlled areas? Are they on granite tables? Do the balances have auto-calibration?
 
G

grzesiek

#10
"nd the uncertainty analysis can help you with that. Start with the balances; calculate them. Next, perform some studies with the two operators on the balances. Finally, make sure you have estimated all the components to the uncertainty. Then, you have an amount (in mg or ml; whichever is more relevant for your studies) that the measurement process is contributing to your study." - ok I understand where you are all going, finally :)

"Hmm... is your balances only used for your study? Maybe the calibration test points need to be down lower near your use range. Can these balances produce the required accuracy you are looking for?" - balances are universally available, anyone can use them. I think so, this doesn't seem good to calibrate 1mg, when I want accuracy of 0.01mg right?

"As far as the study, if you are interested in performing an uncertainty analysis for your balance, all these components will need to be estimated." - I am far more interested to find out what is the main cause of variation in whole measurment process

"From that you can ascertain what kind of variation the balance is producing." - what if I calibrate with a weight, but I weight solid sample that is powder to a little bottle? that's not the same, is it?

"If you are wanting to keep your uncertainty analysis simple, you can assume that many of the variances present would be accounted for during your daily checks." - I want to keep it simple, but as I said I suspect that calibration is preatty good, it's the weighting process that I am concered about

"Are they on granite tables? Do the balances have auto-calibration?"- yes, yes

" Ask the users of the balance if they ever see any swing in the balances. Are these in draft-controlled areas?" - don't understand, what do you call a swing? movement? they're stable. what is draft - controlled area?
 
Thread starter Similar threads Forum Replies Date
R Tolerances for Analytical Balances General Measurement Device and Calibration Topics 3
D Tolerance when verifying an analytical balance General Measurement Device and Calibration Topics 3
C FDA on changing acceptance criteria re: analytical method validation US Food and Drug Administration (FDA) 1
M Analytical Test Instruments Used in V&V Manufacturing and Related Processes 2
K QMS Software Recommendations for an Analytical Laboratory General Measurement Device and Calibration Topics 2
V For Analytical Laboratory Equipment, is PQ = Calibration Criteria? Qualification and Validation (including 21 CFR Part 11) 2
K Validation of an Analytical Method that relies on Multi-Variate Calibration General Measurement Device and Calibration Topics 7
V Pharmaceutical Analytical Method Validation and Gauge R&R Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 10
V Drift and Variation over Time of an Analytical Balance. Qualification and Validation (including 21 CFR Part 11) 2
K Do we need to perform Analytical Method Validation upon Relocation? General Measurement Device and Calibration Topics 5
AnaMariaVR2 Innovation In Assays And Analytical Methods Urgently Needed [article] US Food and Drug Administration (FDA) 0
I Statistical Stability for the PQ of Analytical Equipment Qualification and Validation (including 21 CFR Part 11) 1
A Writing a Procedure for Validation of an Analytical Method General Measurement Device and Calibration Topics 5
J Log Book for Analytical Balance Calibration Verification General Measurement Device and Calibration Topics 4
H Analytical Balance Calibration and Verification General Measurement Device and Calibration Topics 9
Q FDA Bacteriological Analytical Manual (BAM) - Health Care Items Inspection, Prints (Drawings), Testing, Sampling and Related Topics 1
J Flexural Strength Testing - Analytical Laboratory recommendations for Testing Alumina Inspection, Prints (Drawings), Testing, Sampling and Related Topics 1
Q Analytical Tools to increase or improve Analytical Skills Problem Solving, Root Cause Fault and Failure Analysis 1
V Operating Range & Calibration for Analytical Balance General Measurement Device and Calibration Topics 3
E Process Capability Analytical Tools Statistical Analysis Tools, Techniques and SPC 1
F How to determine Minimum Weight of analytical Balance Measurement Uncertainty (MU) 11
B Analytical Lab Environmental Condition (Temperature & Humidity Control Limit) General Measurement Device and Calibration Topics 5
smryan Analytical Scale - Seeking brand and buying advice General Measurement Device and Calibration Topics 7
H Analytical Environmental Laboratory ISO 17025 Registration Scope ISO 17025 related Discussions 3
J Association of Analytical Communities validation study data - How deep should I go? General Measurement Device and Calibration Topics 1
Geoff Cotton NWA (Northwest Analytical, Inc.) SPC software - Your Opinion, Review and Experiences Statistical Analysis Tools, Techniques and SPC 2
C Time Crunch on ASQ's More Analytical Exams Professional Certifications and Degrees 12
O Analytical Method Transfer Approval and Use Document Control Systems, Procedures, Forms and Templates 4
B Validation vs. Confirmation in the context of an Analytical Laboratory General Measurement Device and Calibration Topics 10
C AIAG Attribute MSA - Analytical Method - Page 135 of MSA manual 3rd edition Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 19
R Document for daily check of analytical balance and criteria General Measurement Device and Calibration Topics 2
R MSA Analytical Method Acceptance Criteria and Requirements Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 2
P Pareto Charts - Analytical tools - Form Attached Quality Tools, Improvement and Analysis 5
H Food and Drink in Production Areas - Manufacturer of FDA analytical and diagnostic ISO 13485:2016 - Medical Device Quality Management Systems 5
S Jewelry vs. Normal Laboratory Balances - Accuracy and calibration General Measurement Device and Calibration Topics 2
R Weight Sets for Checking Table Balances Quality Assurance and Compliance Software Tools and Solutions 5
C Conventional Mass vs. Nominal Mass - Certify the in-house balances General Measurement Device and Calibration Topics 9
L Technician Proficiency Testing - Small Scale and Balances Laboratory General Measurement Device and Calibration Topics 7
J Looking for partners Paid Consulting, Training and Services 2
bryan willemot Looking for an Approved vendors list template Document Control Systems, Procedures, Forms and Templates 11
C RA (Regulatory Assurance) Training (FDA) looking for resources Training - Internal, External, Online and Distance Learning 5
M Data Protection and Privacy Policy - looking for a template/example EU Medical Device Regulations 1
M 'Active' device? Software/laptop with attached camera 'looking' at passive metal probe EU Medical Device Regulations 3
K Looking for guidance to write an SOP on Statistical Methodologies? Statistical Analysis Tools, Techniques and SPC 7
blackholequasar Looking into Nadcap for the first time AS9100, IAQG, NADCAP and Aerospace related Standards and Requirements 7
A 21 CFR 820 - Risk Management - Looking for some guidance US Food and Drug Administration (FDA) 3
S We are looking for EU authorised Representative for both Class I and Class IIb devices EU Medical Device Regulations 7
K Looking for a job as Quality Engineer Career and Occupation Discussions 2
B Looking for 10 Internal Audit Online Training Participants ISO 17025 related Discussions 2
M We are looking for a Contract Manufacturer US Food and Drug Administration (FDA) 2

Similar threads

Top Bottom