Metrologists gather at Disney World
NCSL International Workshop and Symposium, Disney World, Florida, www.ncsli.org/conference/2008.
Martin Rowe, Senior Technical Editor -- Test & Measurement World, 8/5/2008 9:07:00 AM
August 4, 2008--After two days of tutorials, the annual NCSL International Workshop and Symposium’s technical sessions and exhibits opened on Monday. The day began with the keynote address by Dr. Richard S. Davis, a metrologist with the International Bureau of Weights and Measures (BIPM) keepers of the world standards for basic measurements. In his address, Davis discussed the kilogram, the last standard of the International System of Units (SI) that’s still defined by an artifact and has been in service since 1899.
| Related coverage: Companies display new calibration products at NCSLI conference |
Davis discussed other possible methods for a new kilogram definition. One involves a 1-kg silicon ball whose size can be measured using x-ray crystal density and laser interferometry. Changes in size will indicate changes on mass. The ball must be kept under tightly controlled temperature. Another method was based on electrical techniques. “Right now, the solution is worse than the disease,” he noted. “The mass and electrical methods aren’t as good as comparison to the artifact.” But Davis expects scientists to eventually, perhaps in two years, develop a measurement technique for the kilogram that doesn’t involve comparison to a single standard. Thus, national labs will be able to produce their own kilogram standrds just as they do with other SI measures.
Technical sessions began immediately after the keynote address. In a session on electrical calibration, Dean Jarrett of NIST discussed measurement methods and NIST traceability paths for high-resistance (1 gigohm and higher) measurements. Methods include a resistance bridge, timing of an RC network, and using a teraohmmeter. He showed a measurement system developed by NIST for the US Air Force where the USAF primary standards lab can use a teraohmmeter to transfer resistance from NIST to the primary standards lab. “The Air Force wanted an automated method for transferring high resistance from NIST,” he said.” From the USAF primary standards lab, metrologists can use other methods to transfer high-resistance to labs worldwide.
Next, George Jones of NIST discussed power loading of 1-ohm resistance standards. He found that reference resistors were producing different values depending on the excitation current (50 mA or 100 mA). He noticed that a cold start (applying current immediately before measuring) produced different results than when the coil resistor was allowed to heat before measuring. The problem was self-heating, and it also existed differently depending on the size of an oil bath. A large (5-ft2) bath produced smaller differences than a 1-ft2 bath. The smaller tank had more cooling and thus produced lower measurements than the large tank.
In the final presentation of session, Ken Kochar of the National Research Council of Canada (NRC) discussed how metrologists implemented a method for calibrating ratio transformers and inductive voltage dividers. The technique, first presented by A.M. Thompson in 1983, describes a simplified version of a switch developed by Thompson. The switch won’t change parasitic capacitance when connected to a load. A change in capacitance introduces an error because it changes the load on the measuring circuit.
The afternoon saw a second set of papers on electrical measurements and a session on metrology lab construction, among others. Paul Roberts from Fluke explained how phase noise affects the calibration of RF spectrum analyzers. “You must measure phase noise when you calibrate a spectrum analyzer, but there’s no such thing as a noise-free oscillator,” he said. Thus, any signal source you use to calibrate a spectrum analyzer will add noise to the measurement. Most calibrations require two signal sources to cover common frequencies, one for frequencies up to 4 GHz and another for higher frequencies. Roberts went onto explain how the different components of a spectrum analyzer contribute to overall noise as you move away from a carrier. Because of digital modulation, noise from one channel can affect the performance of an adjacent channel. Thus, you usually need a bandpass filter on the output of your signal source to minimize phase noise. A calibration report on a spectrum analyzer must convey displayed average noise level.
Following Roberts, Dr. Li Pi Su from the US Army Primary Standards Lab reported on a comparison of vector network analyzer (VNA) measurements conducted at the three primary standards laboratories of the US Department of Defense (the US Air Force Primary Standards Lab,the US Navy Primary Standards Lab, and the US Army Primary Standards Laboratory) and two industry labs (Agilent Technologies’ Santa Rosa Metrology Services and the Anritsu Standards Lab).
The method involved calibrating four Agilent 8493C attenuators at frequencies from 1 GHz to 26.5 GHz, then using a VNA at each lab to measure S-parameters on the attenuators. The measurements showed consistency in the equipment and procedures at each lab because all showed the attenuators to be within manufacturer’s specifications.
In a session on metrology laboratory facilities, Steven Stahley at Cummins discussed the contruction of a new metrology lab in Columbus, IN. He cited issues with location, lab planning and unforeseen circumstances in his presentation. “Start with a high-level view” he told a full room of about 120 people. He reminded them to consider the needs of lab personnel as well as the materials moving through the lab. He explained that lighting is more important than you think, and that you must give each person some work space for personal items such as family photos.
Stahley stressed the need for careful planning. For example, after the Columbus lab was built, he discovered that some machinery used for dimensional measurements was too high for the ceiling. Because the ceiling was made of concrete and a manufacturing facility was above, Stahley had to have a pit dug in the floor for the equipment to fit. He also found a doorway that was too small for equipment to fit through. Finally, Stahley expressed the need to consider unforeseen disaster. "Always build a lab that’s above the highest water level over the last 100 years,” he said. Soon after the lab was complete, a flood destroyed it. A river overflowed because 12-in. of rain fell in a 3-hr period.
Talkback
Related Content
Related Content
Sponsored Links
TMW Resource Center
Browse Resources by Type:
- Boundary-Scan for PCB Interconnect Testing
Corelis, Inc. | White Paper
VIEW NOW
Agilent Technologies | Web Event
VIEW NOW
Protek, Inc. | Data Sheet
VIEW NOW
View More
