Les Thomaidis shows and explains the calibration rack at the Agilent Melbourne Standards Lab that is used to "calibrate the calibrators". This is where other calibration labs send their multimeter and oscilloscope calibrators to get calibrated.
Forum Topic: http://www.eevblog.com/forum/blog/eevblog-422-how-to-calibrate-a-calibrator'>http://www.eevblog.com/forum/blog/eevblog-422-how-to-calibrate-a-calibrator
EEVblog Main Web Site:
http://www.eevblog.com
EEVblog Amazon Store:
http://astore.amazon.com/eevblogstore-20
Donations:
http://www.eevblog.com/donations/
Projects:
http://www.eevblog.com/projects/
Electronics Info Wiki:
http://www.eevblog.com/wiki/
Forum Topic: http://www.eevblog.com/forum/blog/eevblog-422-how-to-calibrate-a-calibrator'>http://www.eevblog.com/forum/blog/eevblog-422-how-to-calibrate-a-calibrator
EEVblog Main Web Site:
http://www.eevblog.com
EEVblog Amazon Store:
http://astore.amazon.com/eevblogstore-20
Donations:
http://www.eevblog.com/donations/
Projects:
http://www.eevblog.com/projects/
Electronics Info Wiki:
http://www.eevblog.com/wiki/
Got it? Yes, All right, take it away. Les He's going to tell us all about the rack which calibrates the calibrators. Yeah, well. basically you've summed it up.
Um, a need. AG There was a need by agent customers to be A to calibrate a calibrator y be it a fluke like this one here. be a Dayron like the older model down there. uh wave Teck whatever Uh, they wanted.
So we developed a rack. um that's uses a Dayron 4950 which is called a Calibrator Calibrator Mhm. Um I believe Pete's talked to you about the Gold 3458? Yes. so we have one of those frequency counter single Channel power meter, scope, etc etc.
and we use this equipment with a couple of current shunts as well or under Metcal. Uh, control Metcal control. Can you explain that? Okay, Metcal is a program developed by Fluke. Yep that allows you to write script to be able to control electronic equipment and you normally have a user interface as you may see on on that display there that you can operate the program and we've written it in such a way that when there's a cable change or an instruction to be done, rather than having a written instruction, you have a pictorial as you can see over there pictorial instruction which excellent picture is worth a th000 words words Yep This is how you hook it up, how you hook it up and you hook it, you hit the Go button and hopefully everything passes.
And how do how do you ensure traceable calibration of the Dayron calibrator? Okay, for example, how does that primary standard get? Well, our two primary standards say for low voltage Mhm or low frequencies is a Gold 3458 and the Dayron 4950 Y Gold 3458 goes to our Loveland Hub in the US right and they're traceable V Nist got it. The Dayron goes back to Fluke and they're traceable by either by Nist or Yucus depending upon where we stand it. So we have two lines of traceability all Landing up at a all Landing up at the one place. So it's common for Um I Guess you guys all to work together as part of one big, happy, traceably calibrated calibration.
Community Otherwise it in in isolation, it doesn't really work. No, it doesn't. No, you all rely on each other. The whole system breaks down really exactly right? You got it.
So you we all rely upon each other to ensure that our uncertainties are within reason. And what sort of uh, uncertainty does your calibrator, your gold Standard calibrator need to have over and above when you're calibrating something like this fluke calibrator for example. okay, well La T uncertainty ratio. The uncertainty of the measurements that we perform on that has to be lower than the uncertainties that we receive from those just a mathematical fact because what happens is if uh, for example, uh uh, fluke in in the US gives us an uncertainty of say one microvolt per volt Mhm on DC Well, we can't have an uncertainty of that.
It's got to be worse than that. Yes and the same times. Usually there's the Um, not necessarily not. NE It all depends upon the um. Do you know how we do uncertainty? Tell us how much time have we got? Well, basically, really, really short. Okay, yeah, there are two types of analysis that you need to do. Type A is type B's right. Type A is measurement of your device on the test.
Mhm. Um, We normally do it five times right. And thus those five measurements that you get. You get a drift rate over those five measurements.
and that's your Uh. Esdm as we call it right. Esdm you combine for estimated deviation of the standard mean words for that effect. Got it exp.
Sorry. Experimental. Experimental. Thank you.
Deviation: Yes, Now, because we've used, say, this device now. This device can either have an uncertainties if it was calibrated by another accredited lab on that measurement, or if it doesn't and you if it does, it doesn't doesn't really matter. If you want to use the tightest specification, use the uncertainties. If you want to use a bord, you use the uncertainties that the manufacturer says it's got right.
So it depends how you want to go and then by uh, by uh, using the summing law the square root of um, you work out what your uncertainty measurement for that point was. and that's your job. as a metrologist. As a metrologist, to as a certified metrologist, to understand all that, take it all in account.
track the data, analyze it. You must love math, right? andth so that that's basically really simplistic. You can get quite complicated, but that's a simplistic. How would you calibrate the oscope calibrators? for example? Well, that one is that's got a 600 Meg oscilloscope out.
On this particular model, it's called the Sc600 Mhm. Uh, Scope calibrator 600 Me so B Basically uh, with the scope output, you've got to make sure that your levels are correct. Yep, your sine wave is a sine wave. you know pulse width Etc So we still use some sort of DVM Mhm.
Okay for linearity across the whole frequency range. from got it, say 10K up to 600 M You would use power sensor or a cist amount. something that's got a broad broad range band range. Because oscilloscopes don't really require Precision calibr because they're generally you know, a percent on the vertical at most you know, yes.
But the the higher the the more sophisticated the oscilloscope. Like this one down here, the our infiniums, the um, the better they are. So what you use to calibrate those has got to be good as well. especially in terms of something like a noise floor or something like that.
Would you measure the noise? Typically measure the noise floor on a scope I wouldn't say the noise floor, but you would definitely measure on the least on the smallest value voltage per division a normally 1 molt. you'd put 1 M volt in there 1 and a half M volts to see how much s you've got EXA will play around with your got it measurement. Excellent! Thank you very much pleasure.
I'm still uncertain about uncertainties.
I understood about 5% of this video!
Dave is like the Gad Fly of the tech sector…he's always buzzing around nosing into things asking questions and such.
Calibrationception
I thought this was going to be about calibrating Les.
Got to love the meth
nice vid
For those who care: The central calibration agencies (National Institute of Standards and Technology in US or JIST Japanese Institute of Standards and Technology in Japan or similar in other nations) of different nations each have their "golden" standard. For Example the mass unit of Kilogram has 8 cylinders of platinum-iridium metal in different labs across the world. Every so often those cylinders are brought together and have their masses compared with differences recorded. The cylinder in France is the only one that is exactly 1 kilogram; all other ones are very very close but not exactly. For example the cylinder from US NIST may be 0.999999997 of a kilogram. Then each national lab will have a set of secondary standards which are regularly calibrated by comparison to the NIST primary standard. Other cal labs would send their standard weights to be calibrated at NIST either against NIST primary standard (Crazy expansive) or against their secondary standards (very expensive). And then you send your scales or set of weights to these cal labs and they cal your part against their primary or their secondary standards. Each step removes the calibration further from the "world" standards increasing the uncertainty of the measurement. This is where the math comes in; to calculate the statistical, combined uncertainty of your part.
calibrate calibrate calibrate calibrate calibrate
How do you calibrate the calibrating equipment? Then how do you calibrate that?
Why on earth are all these Agilent guys Meteorologists?
Are those devices so fragile that weather plays such an important role?
However, this calibration work gave all those Agilent lads quite some belly.
And aren Agilent those guy who made your pince-multimeter that blew up and almost killed you, … Dave?
Calibration is serious.
How would you calibrate a calibrating calibrator(I have no idea what I just said)
Wouldnt the glass block window, if its an outside window, enable sunlight to warm the equipment casings above controlled ambient ?
They are set to a reference.
So was he working for Algient?
Garrus Vakarian.
But how do you calibrate the calibrator calibrators calibrators calibration calibr…wait, what ? Oh, just send it to someone else to do it.
How do you calibrate the calibrator?
Then how do you calibrate the calibrators calibrator?
Not only do they have to all agree to within some uncertainty level but also ultimately the national labs derive their standards from either experiments which employ a fundamental physical principle (like some atomic resonance) or (less preferred) in a few cases a single international physical standard (like the kilogram mass standard, which is a physical object).
alien transmission at 6:15
DEEP!…But it cleared up some mystery! Thanks for this video!
Amazing- I used to work for Datron Instruments- and it is a surprise to see HP still using our calibrators to calibrate calibrators. Here in Stuart FL, we made the standards to calibrate- the calibrators, which calibrate the calibrators… decades ago. It was too uninteresting for the crew at Norwich to bother. So the place where this *was once* done is Stuart, Florida. The old Datron facility is now a self-storage building. Very cool.
Well done Dave.
When you get down to this kind of accuracy, you need to control the humidity as well as temperature.
Now….. lets go to the place where they calibrate the calibrator calibrator.
From the title I was expecting a bunch of engineers getting plastered in the pub.