ISO-17025 accredited TEST lab performing internal calibrations (Yes: Here is how)

AllTheThings

Involved In Discussions
This is kind of an extension of a previous thread:
https://elsmar.c*m/elsmarqualityforum/threads/internal-calibrations-part-of-an-iso-17025-accredited-testing-laboratory-automotive.78443/

The question arose: Does a 17025 accredited TEST lab need an additional CALIBRATION accreditation to calibrate its own equipment being used internally?

The answer from the ABs is: No.
The long answer from the ABs is: No, but you better abide by all various requirements laid out in 17025 for calibration labs.

I thought I would catalog the process as we are able to complete it. So far, I have talked with two ABs, and we have selected one with which to pursue accreditation. I won't be using their names, but they are the usual suspects, and were largely comparable in approach and cost. We are targeting external audits by end of year, so I will update this thread as that process moves forward.

As of June 2021:
* The AB is aware that we want to do our own calibrations. This does NOT affect our scope, but may affect our choice of assessor, as they need competency to evaluate our competency :)
* Right now, this is constrained to environmental chamber transducers (Temperature, Humidity, Pressure), with calibrations being conducted with externally accredited calibrated 'masters', such as calibration-grade humidity sensors, PRTs, etc.
* We have procedures for uncertainty, calibration records-keeping, and calibration traceability (i.e. pick your NMI that is party to the right MRA, etc.), and the other few requirements in the standard.
* We will be training over the next couple months, first by 'calibrating' chambers that are currently in accredited calibration, and ensuring we are consistent with those accredited calibrations. This includes practicing records-keeping, etc.
* I have yet to decide to what degree we do proficiency studies with an external provider, or if proficiency against traceable 'masters' and physical constants (i.e. boiling water, ice bath) is adequate for the relatively relaxed uncertainties we need for this application.

Anyway, that is the story so far. More to come.
 

dwperron

Trusted Information Resource
The things you have learned from your AB are consistent with our previous discussions.
It is nice to get confirmation from the AB as to what they will be expecting, isn't it?
 

BradM

Leader
Admin
I like your approach and the details you provided. There is some error you can pick up from boiling water and ice bath. If you're doing near room temperature measurements, what about buying a high precision (but relatively inexpensive) thermistor and use that to verify the probes?
 

AllTheThings

Involved In Discussions
I like your approach and the details you provided. There is some error you can pick up from boiling water and ice bath. If you're doing near room temperature measurements, what about buying a high precision (but relatively inexpensive) thermistor and use that to verify the probes?

Brad, we have some 3 and 4 wire RTDs and a huge stack of thermistors around for random measurement needs. Unfortunately, our temperature ranges are -70C to 200C right now, and possibly up to 650C. Do you mean setting aside thermistors/RTDs or other gauges as an in-house reference to verify or study against our traceable 'masters'? I like that if it can take the place of external competency studies, and serve as a 'check' for shipping damage.
 

BradM

Leader
Admin
Brad, we have some 3 and 4 wire RTDs and a huge stack of thermistors around for random measurement needs. Unfortunately, our temperature ranges are -70C to 200C right now, and possibly up to 650C. Do you mean setting aside thermistors/RTDs or other gauges as an in-house reference to verify or study against our traceable 'masters'? I like that if it can take the place of external competency studies, and serve as a 'check' for shipping damage.

Oh... I see. My fault. I didn't read your entire temperature range of -70C to 200C. What I was suggesting was using a calibrated thermistor (calibrated by an external source) to do verifications at room temperature as opposed to trying to utilize boiling water and ice baths.
 

gpainter

Quite Involved in Discussions
In my past life we did calibrations of many things. but as i have grown older and with technology. I leave it to the experts
 

AllTheThings

Involved In Discussions
I thought I would update this thread, now that we have been through our initial assessment.

Generally, we did well, with some findings typical of a lab going through the ISO-17025 initial assessment.

With respect to in-house calibrations specifically, the assessor was great to work with, and went through the clauses carefully with respect to our in-house calibrations. A few notes to share:

  • We got a finding on environmental factors in uncertainty calculations. We assumed that we didn't have to put null results in, but the assessor wanted to see those. As an example: Our lab is maintained at 22C +/- 3 degrees. Our temperature DAQ maintains its declared tolerances between 18 and 35 C. Since we always use the DAQ in the lab, there will never be an effect on the DAQ from the temperature of the lab environment. However, the assessor wanted the uncertainty budget documentation to include the zero result from lab temperature. I think I agree with the assessor that this is best practice, if only to demonstrate we considered it. We updated our templates in response.
  • We still need to do ILC/Proficiency with respect to the calibration disciplines we execute on in-house. But not having one of these complete did not gate the assessment. Just having them scheduled (and knowing we need to show outcomes on our surveillance assessment) was sufficient.
  • Right now, only myself and one other person is qualified to conduct calibrations, which is fine, given that we only have 8 environmental chambers and ovens to cover right now. However, the equipment is increasing by 50% over the next year. The early training we've already done to start to bring 2 more people up to speed has been extremely valuable to the staff, even if they don't actually end up conducting calibrations.

Overall, there are three huge outcomes for us:
  1. The benefit of avoiding on-site cal events that are highly disruptive to lab operations and are often lacking in execution compared to our competency with our own equipment.
  2. Staff is now thinking in terms of uncertainty in a much stronger way. This yields a better understanding of decision rules, and what a measurement actually means during a test or during test planning. For instance: If a customer asks for 0.25C degrees as a test condition (don't ask, long story), the technicians are much more equipped to realize this may be an issue, or isn't possible with the device being tested due to thermal mass/drift issues, or may require a chamber with an RTD rather than a thermocouple.
  3. We are also able to declare uncertainties in test reports, as related to chamber conditions or monitoring of conditions during test. This is not typical practice in our industry, and allows customers to better understand test and design margin.
 

Trebor123

Involved In Discussions
I thought I would update this thread, now that we have been through our initial assessment.

Generally, we did well, with some findings typical of a lab going through the ISO-17025 initial assessment.

With respect to in-house calibrations specifically, the assessor was great to work with, and went through the clauses carefully with respect to our in-house calibrations. A few notes to share:

  • We got a finding on environmental factors in uncertainty calculations. We assumed that we didn't have to put null results in, but the assessor wanted to see those. As an example: Our lab is maintained at 22C +/- 3 degrees. Our temperature DAQ maintains its declared tolerances between 18 and 35 C. Since we always use the DAQ in the lab, there will never be an effect on the DAQ from the temperature of the lab environment. However, the assessor wanted the uncertainty budget documentation to include the zero result from lab temperature. I think I agree with the assessor that this is best practice, if only to demonstrate we considered it. We updated our templates in response.
  • We still need to do ILC/Proficiency with respect to the calibration disciplines we execute on in-house. But not having one of these complete did not gate the assessment. Just having them scheduled (and knowing we need to show outcomes on our surveillance assessment) was sufficient.
  • Right now, only myself and one other person is qualified to conduct calibrations, which is fine, given that we only have 8 environmental chambers and ovens to cover right now. However, the equipment is increasing by 50% over the next year. The early training we've already done to start to bring 2 more people up to speed has been extremely valuable to the staff, even if they don't actually end up conducting calibrations.

Overall, there are three huge outcomes for us:
  1. The benefit of avoiding on-site cal events that are highly disruptive to lab operations and are often lacking in execution compared to our competency with our own equipment.
  2. Staff is now thinking in terms of uncertainty in a much stronger way. This yields a better understanding of decision rules, and what a measurement actually means during a test or during test planning. For instance: If a customer asks for 0.25C degrees as a test condition (don't ask, long story), the technicians are much more equipped to realize this may be an issue, or isn't possible with the device being tested due to thermal mass/drift issues, or may require a chamber with an RTD rather than a thermocouple.
  3. We are also able to declare uncertainties in test reports, as related to chamber conditions or monitoring of conditions during test. This is not typical practice in our industry, and allows customers to better understand test and design margin.
Hi could you DM me please and have a chat about 17025, just starting Thanks
 
Top Bottom