Go inside Fluke's electrical metrology lab

-April 23, 2012

In a wooded area behind Boeing's aircraft assembly plant in Everett, WA is the home of test-equipment maker Fluke. Within the Fluke facility is the company's electrical metrology lab, the heart of the company's electrical calibration
 Calibration station for multifunction calibrators

Figure 1 An automated station calibrates multifunction calibrators used as factory-floor check standards. Click here to open a video.

business and the force behind the company's many measurement instruments. Fluke's metrology lab sits atop the company's calibration chain where metrologists and technicians support primary standards, transfer standards, working standards, and calibrators for the company's entire line of electrical measurement products. Fluke's electrical-measurement products get their calibration traceability to a national laboratory through this lab.

The lab houses primary standards for DC voltage through a JJA (Josephson Junction Array) and is one of just a few metrology labs to have one. The lab also contains standards for DC and AC voltage and current and resistance. Transfer standards make the reference values from these primary standards available to calibrations stations used to calibrate Fluke multifunction calibrators and reference multimeters, which in turn calibrate Fluke's measurement instruments in the factory. Fluke's Chief Corporate Metrologist Jeff Gust led me through the lab on April 5, 2012.

The lab, which was under renovation at the time of my visit, has a station that calibrates two multifunction calibrators, for each model type, that serve as check standards for Fluke's multifunction calibrators on the factory floor. The station also calibrates multifunction calibrators for customers who want Fluke Primary Standards Laboratory calibration rather than sending their calibrators to a Fluke service center. Often, those customers want calibration to uncertainties that are tighter than published specifications.

Figure 2 Fluke's electrical metrology lab is one of a few that has a Josephson Junction Array. Click on image to enlarge.
Figure 1 shows a check-standard multifunction calibrator connected to the station. This check standard, one of two in the Everett facility, travels to the factory floor to calibrate the test station that calibrates newly built multifunction calibrators. By using check standards, the working calibrators in the test stations need not come into the lab for calibration. Having two check-standard calibrators ensures that even if one should fail or for some reason go out of tolerance, the other can keep the factory running. Fluke uses the check standards to monitor the factory calibration stations on a weekly basis. Such frequency can reveal if s station's equipment is drifting before it goes too out of tolerance. (See Figure 1 for a short video explanation of the station from Jeff Gust.)

Multifunction calibrators produce DC and AC voltage and current as well as resistance that Fluke and its customers use to calibrate a wide range of measurement equipment. The calibration station itself needs reference parameters that have tighter uncertainties. The metrology lab has equipment that goes to top of the calibration chain. Some of that equipment is, in turn, checked by metrologists at national labs in the US, UK, Canada, and Germany, depending on the parameter.

For DC voltage, Fluke has one of the few metrology labs in the world to house a JJA (Josephson Junction Array). See Figure 2. The JJA produces 10 V with an uncertainty of 6 parts per billion. Metrologists use Kelvin-Varley dividers to produce 1 V, 0.1 V, 0.01 V and so on for calibrating DC voltage transfer standards.
Figure 3 A bank of transfer standards transfers DC voltage from the JJA to other equipment. Click on image to enlarge.
For calibrating voltage standards at higher voltages, metrologists also use the dividers, comparing the outputs of 100 V and 1000 V outputs of those instruments to the 10-V JJA reference voltage. Figure 3 shows a bank of DC voltage transfer standards. Some remain at Fluke's facility while others belong to customers. Fluke may send one of its DC Voltage transfer standard to a customer, who makes measurements on the standard, then sends the results to Fluke for uncertainty analysis.

Fluke manufactures DC transfer standards, AC transfer standards, and combination DC/AC transfer standard used at metrology labs around the world. Calibrating AC voltage and current isn't done directly as it is with DC. Instead, calibration is done through AC thermal converters using temperature measurements (Figure 4). AC calibration begins applying a known DC voltage (usually 1 V) to a DC/AC thermal converter, which produces a millivolt output of the temperature that the converter senses.
Figure 4 Thermal converters link AC voltages to DC voltages for AC/DC transfer standards. Click on image to enlarge.
Then, an AC voltage is applied to the same converter and adjusted until the converter's millivolt output matches those produced by the known DC voltage. The AC voltage from the transfer standard is recorded and that's the AC equivalent of 1 VDC. Thermal converters, some of which were developed by John Fluke himself in the late 1960s, can be used to perform AC voltage calibrations up to 50 MHz. The converters have a long and well-documented performance history, which is why they're still in use and produce calibrations for customers with uncertainties as low as 0.08%.

Fluke also uses its AC/DC transfer standards to calibrate other AC/DC transfer standards. Two such standards, facing sideways in Figure 5, are calibrating a third transfer standard, facing front. As part of the lab renovation, the station used to perform these calibrations will get stepper motors under computer control that will automate the turning of the knobs. The transfer standards calibrated here then become the references for calibrating the station shown in Figure 1. Customers with Fluke AC/DC transfer standards also send their standards to Fluke for calibration.

Figure 5 Two transfer standards (rear) calibrate a third (front). Click on image to enlarge.
Gust explained that the calibration station in Figure 5 will be updated to include a Fluke 8508A reference multimeter. Engineers are also updating computer systems and software. But, new calibration equipment requires validation, enough for lab personnel to acquire enough confidence in the new equipment's performance. Thus, the process takes months to complete. Metrologists will validate all instruments, cables, connectors, and software until they're convinced that a new station is ready for operation. In the meantime, they will keep the older, trusted equipment in use. Even after the new equipment begins operation, the older stations will be ready to come online at any time in case measurements raise doubts.

For resistance, the Fluke electrical metrology lab has several 1 Ω and 10 kΩ reference resistors, all of which reside in temperature-controlled oil baths at 25°C (Figure 6). Because calibration labs have just a few reference resistor values, metrologists need a way to transfer those reference resistors to other values. For that, they use resistance bridges that produce resistance-ratio quantities from the reference value to another value. Using a bridge, metrologists and technicians can calibrate multimeters such as the Fluke 8508A reference DMM and the HP (before Agilent) 3458A DMM in the calibration station in Figure 1.
Figure 6 Reference resistors reside in oil baths to minimize drift from temperature fluctuations. Click on image to enlarge.
Fluke sends the reference resistors to Canada's NRC (National Research Council metrology laboratory for calibration to obtain uncertainties of 0.025 parts per million. In fact, Gust will carry a reference resistor to Ottawa himself because one was recently damaged in transport. "We use the NRC laboratory because it has the measurement capabilities we need," said Gust.

Calibration is meaningless without verification and documentation. "Metrology is all about records and we have calibration records that go back 30 to 40 years," said Gust. "We might have a $2000 piece of equipment that has $50,000 of calibration history to go with it. That kind of history gives a measurement standard its value." As a result, the measurement techniques used to validate new test equipment, which lack years of history, are often the subject of conference papers, both by Fluke engineers and from engineers at national labs. "If you don't have absolute integrity in your measurement science," said Gust, "someone will uncover it." The fact that the calibration station contains an HP 3458A shows how much history Fluke has on the station.

Loading comments...

Write a Comment

To comment please Log In