How effectively are RF absorption levels policed?

But while there are internationally recognised safety limits for RF exposure from wireless devices based on the body’s specific absorption rate (SAR), the standards for actually measuring SAR have lagged far behind.

There are no standards for measuring SAR from a transmitter that is held on the lap or kept in a pocket, or for multiples of such transmitters operating near the body. And there are none at all for devices that operate between 3-6GHz, such as WiFi and WiMAX phones.

The first single, globally accepted standard for measuring SAR from a wireless product was only finalised last year. This is IEC62209-1, a harmonisation of the various co-existing CENELEC (European Committee for Electrotechnical Standardisation), IEC (International Technical Commission) and IEEE (Institute of Electrical and Electronics Engineers) standards covering exposure to the head at frequencies up to 3GHz.

Since then, the IEC and IEEE have started addressing these measurement ‘gaps’ by defining extensions and additions to IEC62209 (known as IEEE1528 in the US) to cover multiple, body-mounted transmitters and devices that operate in the 3-6GHz range.

The IEC is taking on the body-worn devices under the heading IEC 62209-2, which will cover WiFi laptops, police radios used on the lapel, and various types of distributed computing devices.

“All current mobile phone and radio measurement standards address exposure to the head, but if you’re using a device on your lapel, the exposure geometry is really quite different,” says Dr Philip Chadwick, chair of the CENELEC 106X committee that writes European standards for human exposure to EM fields. He is also technical director of research consultancy MCL, which develops the head phantoms used in phone testing. “At the moment, it’s not tested because the view is if it’s safe against the head, it’s safe against the body. This assumption is probably correct but we should still look at it,” he adds.

The IEEE is in charge of the extension to 6GHz, under the 1528 banner. The aim is to have a standard by 2009.

The measurement challenge from 3-6GHz is that the body absorbs these higher frequency signals very effectively at its surface and so there is a very steep gradient to the electric field.

“In order to make an accurate measurement, you need to be at the surface of the SAM head phantom, however the available field probes haven’t been able to get near enough,” explains Motorola’s Mark Douglas, who is chairing this particular working group. “The probes need to be very much smaller so you can get close up to the surface without displacing too much of the tissue-simulating liquid in the phantom. We are pushing manufacturers to develop suitable probes and that’s going very well.”
The liquid used inside the phantom also needs a rethink because the usual recipes of water, sugar and salt that are representative of the dielectric parameters of human tissues at 900MHz, do not seem to track to these higher frequencies. “On one hand you want the ingredients to give the properties of human tissues but you also want them to be safe to use and cheap,” explains Douglas.

Dr Camelia Gabriel at MCL has recently developed a possible prototype liquid using water and a form of oxidised mineral oil, which Douglas says looks promising.

All these measurement standards involve putting a real wireless device against a physical phantom and measuring the electric field. But there is great interest in using computer models of the phone and human body instead and the IEEE is working on this as an additional element to the 6GHz extension.

The first part is IEEE1528.1 – a general standard for compliance assessment using finite difference time domain (FDTD) analysis for simulating spatial-peak SAR anywhere on the human body from wireless communications devices at 30MHz to 6GHz. 1528.1 will describe FDTD concepts, allowed extensions, anatomical models, techniques, code validation procedures, uncertainties, and limitations.

A second part, 1528.2, will describe the specific use of FDTD when the exposure is from vehicle-mounted antennas for applications like police car radio. A version called 1528.3 will be for mobile phones – assuming all goes to plan.

“We started this project almost a year ago to see if it would be possible and it is not yet clear whether you can do without any measurements,” says Wolfgang Kainz, who works at the FDA in the US and is chairing the IEEE working group.

Computational standards are a welcome development in compliance-assessment of applications like police car radios, where field measurement is extremely difficult and all manufacturers currently use their own particular computational techniques. For mobile phones, the need is less obvious but the big phone companies are very keen to move to computer models in order to decrease compliance time and cut test equipment costs, according to Kainz.

“A computer-based method would be much easier for them because phones today have so many accessories to test: clips, ear-pieces, headsets and different batteries and so on,” he explains. “Instead of physically measuring all the different variations, once you have a correct computer model of the phone, you just hook it up to the different attachments and re-run the simulation without doing hours of measurements.”

That is the theory. In practice, it may be difficult to ensure that phone models are a true representation.

As an example, Kainz mentions a research project that was comparing SAR measurements between a computer model and a real phone. There was a curiously big discrepancy in the SAR readings. Finally, the researchers discovered that in the real phone, a metallic clip attached to the display was not grounded to the PCB, whereas in the computer model, it was. As Kainz says: “A mobile phone has so many parts and we don’t know yet how detailed the model has to be. Do we need sub-millimetre resolution, for example? We need to do a lot of comparisons between measurements and simulations to find out what’s required.”

It may not be feasible to have a pure computational standard without measurements, says Kainz: “You may need at the very least spot-checking of measurements on a real phone.”

This idea of a purely computational approach is ambitious but also slightly controversial – certainly for phones where it is easy to do physical measurements. Dr Chadwick sees benefits for lower frequency applications such as anti-theft RFID systems. “With these you are trying to protect against current flow in the central nervous system which you can’t measure, all you can do is calculate it in some other way,” he says.

However, he thinks it may be difficult to standardise and future-proof computerised human anatomical models. “Today, the state of the art body model is at 2mm resolution. In a year or so, the resolution will be 1mm and then 0.5 and so on,” he explains. There are many different groups around the world who have developed anatomical models, all based on different techniques, and so they give different answers to simulations, according to Chadwick.

“As with the physical measurement standards, there’s a need to define a computational system that is always conservative, so whatever you run, you always gets a bigger exposure level than the real one. It doesn’t have to be accurate, just conservative and stable,” he states.