Applications - www.dataproof.com

Maintaining The Laboratory Volt

A variety of voltage standards are used around the world to maintain a local or national volt. The volt in this case is usually not the output of a particular device; rather it is the statistically derived value from a group of artifact standards. All of the standards in the group have a known offset from the statistical volt and thus any one of them can be used to calibrate other standards.

The following pages describe examples of how to intercompare voltage standard such as saturated cells, solid-state devices or a Josephson Array. Laboratories using artifact standards (saturated cells or solid-state ) should make periodic comparisons with units traceable to their national standard. For a laboratory to have confidence in its own internal volt requires having more than one voltage standard. We recommend a minimum of four standards to maintain a good statistical average at the 1 ppm level. If requirements exceed 1 ppm, then more than four standards will be needed in the group. Alternatively a Josephson array may be used. The process recommended for laboratories to maintain their internal volt with three or more voltage standards follows.

Recommended process
Voltage standards, whether they are saturated cells or solid-state are compared in pairs. The low terminals of the two standards being compared are connected together. The small difference between the two high terminals are then measured by means of a suitable meter. Measurements should be taken both in the forward and reverse directions to effectively eliminate the zero offset of the meter circuit. Only the negative terminals of the pair being measured should be connected together; all other devices are left floating.

All of the standards in a group are compared in pairs in an appropriate order or "measurement design". The design used should follow the procedure recommended by your national laboratory. Most national laboratories recommend one of three designs: ring, favored cell or statistically-balanced. Ring designs are used in some Asian countries and favored cell designs are used in some European countries. In The United States, NIST recommends the use of statistically-balanced designs. Procedures will be described in more detail in the next section (For more information, refer to NBS Technical Note 430).

Improving measurement results
Voltage measurements are extremely sensitive. Handling the leads will increase measurement errors because of the thermal emfs caused by friction and body heat. Saturated cells are also sensitive to changes caused by stray capacitance from your body presence and other electrical noises in the laboratory. The best way to reduce these errors is through the use of a Low Thermal Scanner. A switching system eliminates the need to handle the measurement leads during the measurement process. With automation, data runs can be taken when disturbances are a minimum, usually at night.

Below is a graph showing the readings taken over a five month period. For the first three months the measurements were taken manually with the leads moved by hand. During the last two months, the graph shows the significant improvement in standard deviations when a Low thermal Scanner had been installed and the leads were no longer handled.