NZSM Online

Get TurboNote+ desktop sticky notes

Interclue makes your browsing smarter, faster, more informative

SciTech Daily Review

Webcentre Ltd: Web solutions, Smart software, Quality graphics

Quick Dips

Resistor Network

Measuring temperature to a thousandth of a degree or more may seem academic to most people but for others, such painstaking accuracy can mean the difference between life and death. An Industrial Research scientist's breakthrough will help ensure that where temperature precision is crucial, testing equipment won't let you down.

As we all know from different bathroom scales, our measuring equipment does not always tell the truth -- on one set of scales you could weigh several kilos more than on another. How can we be sure a measuring device is accurate?

The temperature scale is defined internationally, and the Measurement Standards Laboratory at Industrial Research Ltd in Lower Hutt maintains this standard scale in New Zealand. Organisations such as oil-testing labs, and other research bodies that measure temperature to a few thousandths of a degree, take their measuring equipment there to be checked and calibrated.

Temperatures are defined by using fixed points of reference, such as the known freezing (solid state) points of pure metals. Platinum resistance thermometers calibrated at these fixed points are then used to measure temperatures between them and fill in the gaps.

Industrial Research use devices called resistance bridges to work and calibrate these resistance thermometers. Temperature is measured in terms of electrical resistance which the resistance bridges display.

While the bridges can detect resistance changes of one part in 10 million, the bridge itself must be calibrated to ensure the readings are correct.

"Up until now we haven't been able to calibrate resistance bridges because to do so by conventional methods is prohibitively expensive," IRL temperature standards scientist Rod White says.

There are perhaps three or four AC-voltage divider instruments in the world which have greater accuracy than a good resistance bridge, and are purpose-built to calibrate them.

"One would take about two years to build, cost about $500,000, and there are probably less than a dozen people who would know how to make them," says White.

Although the standards team has the expertise to construct one, it is not economically feasible.

Up until now, the best the Measurement Standards Laboratory and many other National Standards Laboratories have been able to do is simple one or two-point checks on the bridges using calibrated resistors. Another option, costing less at around $160,000 and commercially manufactured, is to use a Hamon resistor, whose basic design was proposed in the mid-50s.

Hamon resistors use a number of equal-valued resistors connected in series and in parallel to gain accurate resistance ratios. They have mainly been used to establish resistance standards at decade intervals (e.g. 1, 10, 100, 1,000). Again, however, the gaps between these intervals are wide, leaving large grey areas and concerns as to whether a resistance bridge is performing accurately between these points.

White has come up with an invention based on the Hamon resistor which can be used to calibrate the bridges. It is a lot less complicated and could retail for around $4,000.

It occurred to him that there were ways to test within the grey area. He varied the values of the resistors in the network, and connected them both in series and parallel, and combinations of the two. With unequal resistor values, a great variety of resistances can be generated from a very few components.

Exhausting all possible combinations, four resistors can produce 35 interrelated resistances, which cover the range of the capability of the bridge. With appropriately chosen resistors, the 35 different readings can also be made to check every part of the bridge circuitry, and demonstrate that a bridge is accurate and trustworthy.

"It's been a pleasant surprise to find something so simple could work so well and cheaply."

During the testing stage, White examined over 30 resistance bridges with the new network and found that one in five was giving faulty readings.

"We didn't expect it to be this accurate. It's 10-100 times better than we first thought it would be, better than any commercial AC or DC bridge, and it has the potential to improve the temperature standards of measurement labs all over the world."