Implementing ISO 17025 - Where Do I Start?

D

D_Wood

Where Do I Start?

When I first joined this site back in October, it was only speculation that our test lab would seek accreditation. I learned today that I am now in charge of "refining" our test lab and get it accredited to 17025 through A2LA. There is a lot of work to be done before coming close to accreditation.

I am new to this standard, and would think my best bet is to find a consultant that may be able to guide me through this process. I have not been given a time frame as of yet, but I know the sooner the better.

Our company manufactures medical devices, and already has an ISO approved quality system. I figure some of the requirements for 17025 are being met now. However, I understand our test lab is sort of its own entity, and needs its own quality system to be compliant with 17025. Am I understanding this right?

Are there consulting firms that will walk me through the process until our lab is ready for its audit, and until we obtain accreditation? Or should I go ahead and impliment what changes I know need to be made, and then seek consultation?

Thank you for your patience. I hope I am not boring you with my question.

Dan
 
D

D.Scott

Hi Dan - You are going to be bombed with offers on the consultant thing so I won't even try to give you some names.

I think you should start by getting a copy of the standard, read and understand it, then do a gap analysis on where you are and where you need to be. Yes, the Lab needs a quality system, but most of the requirements are already covered by your existing system. You can make a statement in the Lab Manual that brings the corporate system into the lab then worry about only the additional stuff and deviations.

Once you see where you are and what else you need, you can work with a consultant on HOW you will get there. There will probably be a number of areas where what you thought was covered needs to be polished up a bit and a consultant could help you there as well.

As long as you have no timeline, why not get your feet wet on your own, ask as many questions here as you can and see where you are then. I think you will benefit greatly by the experience. You'll sure be getting a better understanding of your own lab and the standard.

Good luck and don't be afraid to ask for help.

Dave
 
D

D_Wood

Dave,

I sure appreciate your reply (and e-mail), and the encouragement. I have been looking over the standard, and starting to highlight the areas where I feel we may need the most work. I would like to do as much of this as I can on my own before seeking the outside help. Knowing I can ask questions on this site makes me feel comfortable.

Being a moderator myself for an outdoor website, I can really appreciate what you guys have done to provide this site, and keep it going. I look forward to becoming a regular here.

Thanks again,

Dan
 
K

Ken K

Hi Dan,

Been there, doing that. Our final assessment is scheduled in two weeks. Been at it for almost nine months so I can understand what your going through...or about to.

Dave was a tremendous help when I first asked questions about 17025. He won't steer you wrong. Lot's of other folks will also answer any questions you might have. They will help you every step of the way.

With that being said, read everything you can get your hands on about 17025. Start with the spec itself. A2LA has an auditor checklist you should download to know what they are looking for.
Although we didn't use them they have some really good information on their websight.

Do a search on the internet and you'll find some manuals or parts of them. You'll see how everyone else interprets the standard.
Get to know it and then do your gap analysis. Then when you think you have an understanding of what needs to be done, hire a consultant. We didn't and it's been a bumpy road. The assessor assigned to you cannot consult. And there is a fine line concerning a conflict of interest resting on his/her shoulders. You can ask but don't expect an answer.

Finally, dot your i's and cover all the "shalls" and be ready to correct the non-conformances. Our assessor told me one thing that has stuck with me...I don't care how you get there, just get there.

Good luck Dan. When your finished you'll realize what a long, strange trip it's been. But very satisfying.
 
R

Ryan Wilde

Dan,

The easiest route would be to hire a good consultant, but there are a few things to watch for before you sign a contract. Basically, your consultant will have to have a fairly strong background in whatever disciplines that your testing lab uses.

I'll explain a bit more. ISO 17025 is divided into two sections: Section 4 is, in effect, the sections of ISO9001:1994 that pertain to a calibration and testing lab. Section 5 is the technical section. The assessor will look at the section 4 portions very quickly (like 1/2 day or so if your system is in good shape). The technical section is another story.

The bulk of the assessment is the technical section. Your assessor will usually have a minimum of 10 years of experience in the types of tests that your lab performs, and quite often I have found that the assessors are also on the standards writing committees for the types of tests that they assess. The competence of the technicians will be sampled in-depth, as will the adequacy of your procedures. The uncertainties of your specific measurements will be carefully scrutinized. This is the difference between ISO 17025 and QS/ISO9000 audits, the fact that the assessor will know the tests that you perform and what he/she is looking for as far as weaknesses, the proper way to perform the tests, what to expect for uncertainty, etc.

Therefore, since the bulk of the assessment is technical, your consultant will have to be technically competent in order to make your assessment relatively painless. A quality system cosultant will not help on this, you need a technical person for the meat and potatoes of ISO 17025. There are quite a few good consultants out there, and each will have strengths in particular disciplines. Most you will find will probably be strong in calibration activities, such as electrical and/or dimensional. I don't knowo what you actually test, but you should be able to find a consultant that did it for years, and may in fact be an assessor for that discipline. Shop around, and find someone that understands exactly what you do before you write a check.

Ryan
 
D

D_Wood

Ryan,
Thanks for the input. I have really been looking over section 5 now. Our testing is mainly to show compliance to EN 60601-1, which is electrically operated medical devices, and IEC 60601-2-38, electrically operated hospital beds. It is a two-edged sword, but I would be more comfortable with an auditor who is familiar with these two standards than one that is not.

I already have identified a big problem. Our lab has no documented "proficiency" rating of our techs. We have their training records, and education records, but nothing to show a performance level. It is going to be diffficult to implemement a program to "rate" our techs without getting them all hot & bothered by it. I can already forsee this causing bitter feelings among technicians, and towards myself & management. If anyone has a suggestion, I'm all ears.
-Dan-

P.S. I am awaiting a reply from Marc or a moderator to tell me it is OK to post a link to another site. I'm still new, and I want to respect this site's policies. I just got into a dispute on the outdoor site with a guy for just this reason. This guy actually started sending SPAM to members using their e-mail addresses located in their profiles. The administrators took it personnaly, and I don't blame them. So, I am a little reluctant to just throw a link down. I know it is not a site that is in competition with this one, but it could still be seen as SPAM by some. So, I'll be patient, and when I get the word, I'd be more than glad to post a link and invite y'all to visit the site.
 
R

Ryan Wilde

D_Wood said:

I already have identified a big problem. Our lab has no documented "proficiency" rating of our techs. We have their training records, and education records, but nothing to show a performance level. It is going to be diffficult to implemement a program to "rate" our techs without getting them all hot & bothered by it. I can already forsee this causing bitter feelings among technicians, and towards myself & management. If anyone has a suggestion, I'm all ears.
Click to expand...

Proficiency is not a rating, it is a pass or fail. I'll give an example:

You have an artifact that was calibrated by a national lab to provide 10.00014 VDC ± 0.5 PPM (pretty dang tight, but it makes this example much easier). Your lab's accredited best measurement uncertainty for VDC measurement is ± 25 PPM. Therefore, on a blind test (the techs shouldn't know the nominal that they are shooting for), the techs should measure between 9.99989 and 10.00039 VDC. Your results from your techs are:

Johnny measured 10.00016, and passes
Janey measured 10.00008, and passes
Jimmy measured 9.99988, and fails

What this means is that Jimmy either needs training (after which you repeat proficiency testing for Jimmy), you change your uncertainties to include how sloppy Jimmy is, you don't allow Jimmy to perform accredited tests, or you sack Jimmy altogether.

Ryan

P.S. No Jimmys were harmed in the above example. Any similarity to any other Jimmy, either living or dead, is purely coincidental.
 
Last edited by a moderator:
D

D_Wood

Ryan, thank you for clearing that up, and for being nice to "Jimmy". ;)

I realize there are several factors in determining the uncertainty of measurement. I am curious as to the information provided me on a certificate from A2LA for our source for equipment calibration. I have the "best uncertainty" numbers for various measuring parameters (ranges) for several types of measurements used by this facility. Am I able to assess a "best uncertainty" number to my measurement ranges based on these numbers alone, or do I have to take into account other factors?

Thanks,

Dan
 
R

Ryan Wilde

D_Wood said:

Ryan, thank you for clearing that up, and for being nice to "Jimmy". ;)

I realize there are several factors in determining the uncertainty of measurement. I am curious as to the information provided me on a certificate from A2LA for our source for equipment calibration. I have the "best uncertainty" numbers for various measuring parameters (ranges) for several types of measurements used by this facility. Am I able to assess a "best uncertainty" number to my measurement ranges based on these numbers alone, or do I have to take into account other factors?

Thanks,

Dan
Click to expand...

That depends on what instrument or gauge you are talking about.

If you are talking about a physical attribute gauge (such as a plug gage) then yes, the stated uncertainty on your certificate is the starting point. Then you would have to determine any drift (such as wear, relaxation of metal, etc.) by comparing previous certificates to figure out the change over a calibration cycle. Then you have to figure out your temperature range at which you use the gauge (a good lab calibrates physical gauges at 20°C), and using the µin/in/° compensation, put in your error that will be caused by environmental effects on both the standard and the test piece (especially if you have dissimilar metals, such as a steel standard and an aluminum test piece).

Mechanical gauges (such as calipers) you would actually use the measurement uncertainty from the calibration certificate, as well as the tolerance of the gauge itself. Again, environmental effects would have to be taken into account.

Electrical gauges (such as a multimeter) are actually easier. The uncertainty on the calibration certificate is basically useless unless you have a very high end multimeter (such as a Datron 1281, HP/Agilent 3458A, Keithley 2000, etc). The uncertainty starting point on a standard handheld multimeter would actually be the tolerance of the meter itself. If the meter is not made by one of the big guys (such as Fluke/Wavetek, HP/Agilent, Keithley, etc.) then you would probably want to take the tolerance as a rectangular distribution (for Fluke, it is actually a normal distribution with k=2.54, for the others it is normal with k=2). Then you would have to figure in lead effects (close to nil on 3-4 digit multimeters). Then there is always environmental, unless you hold to the magic 21-28°C range.

That should be enough for now to keep your head spinning for a day or two...

Ryan

Electrical gauges are actually easier, in that
 
K

Ken K

After three days of intense scrutiny by our assessor, we acheived our goal of ISO17025 accreditation. 3 very minor non-conformances were written which will require corrective action.

I'd like to thank everyone who was kind enough to offer assistance when I needed it. Your advice and help were greatly appreciated by me. Glad you were here.


Yahhhhhhhhhhhhoooooooooooooooooo!!!!!!!!!!!!!!!!!!!!
 
You must log in or register to reply here.
Top Bottom