I just had to do a ISO17025 cal on a Compaq PC! Yes that is right! UL told our customer that they had to have a cert for the computer's clock! To top it off our customer had bought NIST's software that corrects the time every 12 hrs, but NIST did not provide a cert so we had too. The UL auditor even told them that if they were to buy a Sat/GPS timebase unit that they would still have to provide a cert for the Sat/GPS! Yes that is exactly what UL told our customer. If you knew what this computer was used for you would even be more blown away. UL told them that the computer clock had to be calibrated with an accuracy of +/-1% @ 7.5 hrs. Yes that's right, do the math. I think I could get that accuracy with a sundial. If my math is correct that would be +/-4.5 minutes! Of course after waiting 7.5hrs the total deviation was .04 sec, which was more than likely caused my my own error. Anyway, I thought I would let you know what the most stupid thing that I have ever calibrated has changed. Have any of you run into this sort of BS from UL or any other auditor?
HEY UL GET A CLUE!!!!!!!!!!!!!!!!
Coup,
Are you gonna be able to look at yourself in the mirror after "calibrating" a computer's clock? I would have told the customer, "Sorry, I have principles. I can't in good conscience certify this. You'll have to find somebody else. " I had a similar situation when a customer wanted their Huntron Tracker "cal'd". Couldn't bring myself to do it. And I slept good that night.
Were you able to keep a straight face when you handed them the Cert?
Quote from: Duckbutta on 01-16-2009 -- 22:16:08
Coup,
Are you gonna be able to look at yourself in the mirror after "calibrating" a computer's clock? I would have told the customer, "Sorry, I have principles. I can't in good conscience certify this. You'll have to find somebody else. " I had a similar situation when a customer wanted their Huntron Tracker "cal'd". Couldn't bring myself to do it. And I slept good that night.
Were you able to keep a straight face when you handed them the Cert?
Believe it or not , I actually found a procedure for calibrating a computer clock. As I said before NIST actaullly makes software to maintain computer clock accuracy. I do think it is BS just like people having to calibrate tape measures. Any measuring instrument can be calibrated, but some of those measuring instruments are not worth the time. If your measurement is that critical maybe they need to be using something more accurate.
The customer that needed the computer calibrated thought that it was a BS cal also, but they had to have it done in order to comply with UL standards.
My point to this post is to point out that more and more auditors are ignorant fools that lack technical knowledge.
Because of some of our customers we have to have a UL audit also. After ranting about this situation to my boss he told me that this past UL audit the UL auditor tried to say that a Fluke 5520A is not accurate enough to cal a Fluke 87! WTF!!!! My boss said that our QA manager had to show her on paper that it was more than accurate enough!
UL must be getting their auditors off the short bus.
Those that can, do
Those that can't, teach
Those that can't teach (fill in the block)
Mike
Accreditation has become the biggest fraud perpetrated on the industry since Wavetek. It's a veritable Alphabet soup. Kinda like all the sanctioning bodies that have watered down Boxing. My experience with these audits is that if the check clears, you're in. They'll come in and make you put on a nice dog and pony show, point out a few things that could use improvement, maybe even make you come up with a Corrective Action Plan. All designed to divert you from that nagging feeling you get when you know you're being robbed. They could at least drop the pretense of legitimacy and wear handkerchiefs over their faces when you give them the loot.
I did not believe that myself on the Fule 87 but they are right. It does not meet 4 to 1 accuracy ratio on the capacitance function.
Quote from: USMCPMEL on 01-18-2009 -- 15:30:20
I did not believe that myself on the Fule 87 but they are right. It does not meet 4 to 1 accuracy ratio on the capacitance function.
Actually, she was saying that it didn't meet spec on DCV. We have actually discussed here on the forums in the past that Capacitance on the 5520A was not very accurate.
Let me come down on both sides of the fence on this.
THE PRO ACCREDITATION SIDE
It can be a useful tool (if they would ever get it straightened out) for understanding if a lab truly has the ability to legitimately cal something. The whole issue of "4:1" I think is at its heart. It's aim (if you presume it has honest aims) is to assure that competent people do cals in an appropriate lab environment using adequate standards and correct methods. These are shortcoming all too common in the industry. Anyone who blindly believes a 5520A is good enough for everything out there (at a handheld meter level) may have have a competence problem. Ensuring labs are staffed by competent metrology professionals helps overcome that problem. I've had work done by one too many cheeseball labs over the years, and with 30+ years in the business, I'm fed up with low paying cheeseball idiot labs that make it hard for real metrology people to get a decent paying job. ISO17025 (in its own way) has helped.
THE ANTI ACCREDITATION SIDE
It makes honest people honest. But a lab that knows how to shuffle papers really good (a socalled "cheeseball" lab) can pass accreditation but still do crap work. Every ISO17025 accreditation auditor had darn well better be an EXPERT at what ever calibration discipline they audit. The unfortunate truth is that they're not. ISO17025 has many good points, but it can get really anal about a lot of meaningless stuff.
Let me ask a quick dumb question about calibrating the computer clock. I thought anything cal'd had to somehow be a part of NIST traceability in the scope of accreditation. I'm trying to understand how the clock on a computer could be part of the scope. The only thing I can think of is having a National Instruments card installed;l in which case, I would think it has it's own internal clock.
Time to get back to work.
USMC PMEL and Hawaii 596, I think are missing the point. Definitely USMC. Accuracy and Uncertainty are two different animals, though not mutually exclusive. My point is that a simple "Test ACCURACY Ratio" of 4:1 is adequate for the vast majority of measurements. This ratio gives you enough margin for error to not be overly concerned about your uncertainties (Type As), which are numerous and EXTREMELY time consuming to quantify properly. If you know what you're doing, these uncertainties are minimized and trying to determine what they truly are is only a "feel good" measure that isn't adding any value, kinda like airport security. Everyone feels better that there is "security" in place, but are we truly safer? Same principle applies here. Don't misunderstand me, there is a place for Uncertainty Analysis, particularly nowadays when the T. I. s approach the accuracy of the standard. But for the VAST MAJORITY of work, an old school 4:1 will suffice.
Quote from: Hawaii596 on 01-19-2009 -- 08:19:11
Let me come down on both sides of the fence on this.
THE PRO ACCREDITATION SIDE
It can be a useful tool (if they would ever get it straightened out) for understanding if a lab truly has the ability to legitimately cal something. The whole issue of "4:1" I think is at its heart. It's aim (if you presume it has honest aims) is to assure that competent people do cals in an appropriate lab environment using adequate standards and correct methods. These are shortcoming all too common in the industry. Anyone who blindly believes a 5520A is good enough for everything out there (at a handheld meter level) may have have a competence problem. Ensuring labs are staffed by competent metrology professionals helps overcome that problem. I've had work done by one too many cheeseball labs over the years, and with 30+ years in the business, I'm fed up with low paying cheeseball idiot labs that make it hard for real metrology people to get a decent paying job. ISO17025 (in its own way) has helped.
THE ANTI ACCREDITATION SIDE
It makes honest people honest. But a lab that knows how to shuffle papers really good (a socalled "cheeseball" lab) can pass accreditation but still do crap work. Every ISO17025 accreditation auditor had darn well better be an EXPERT at what ever calibration discipline they audit. The unfortunate truth is that they're not. ISO17025 has many good points, but it can get really anal about a lot of meaningless stuff.
Let me ask a quick dumb question about calibrating the computer clock. I thought anything cal'd had to somehow be a part of NIST traceability in the scope of accreditation. I'm trying to understand how the clock on a computer could be part of the scope. The only thing I can think of is having a National Instruments card installed;l in which case, I would think it has it's own internal clock.
Time to get back to work.
No your measurement uncert., which in this case is time, has to be on your scope. So it can be a 17025 cal due to our scope including time. He have a lot of timers for Pharm and other companies.