Huidig land wijzigen? Nederland | Nederlands

detection, standards!

sensNORM President Olaf Riebenstein on the new standardised measurement method for PIR sensors

For many years, manufacturers of presence and motion detectors have been using their own method to determine the fields of detection of their PIR sensors. Thanks to sensNORM, an association of leading manufacturers have now created the basis for a uniform approach while substantially influencing IEC 63180, which was developed in parallel. Olaf Riebenstein, President of sensNORM and a manager at ESYLUX, discusses the background and content.

At ESYLUX, Olaf Riebenstein heads up the Serial Product Support and Technical Editing departments and has been involved in standardisation work at a national and international level since joining the company over 16 years ago. He became President at sensNORM three years ago and has been an executive board member from the outset. At the same time, he is actively involved in the drafting of IEC 63180. As an ICT electronics expert, his origins lie in safety technology.

Mr Riebenstein, aren't standards just a bit boring?

There are definitely people that say standards are boring. They find them dry and struggle to understand the text. But there are also lots of people who love these kinds of things. These include people sitting on standardisation committees. They just love it. Making sure to choose their words carefully. It is important for the wording to be as unambiguous as possible and not to generate any questions. All of which is exciting. And that's why a standard is anything but boring for me. It's fun.

A few years ago, the leading manufacturers of presence and motion detectors got together and have now defined a long-awaited measurement and testing methodology for devices in the shape of sensNORM.
What was the co-operation like?

Well, initially it's a bit strange to suddenly find yourself sitting around a table with your competitors, slowly getting to know each other better. At first, everyone's trying to get a feel for each other. Over time you bond and find common objectives. And ultimately, we got on so well that we now all work together constructively and so far have also managed to meet our objectives.

sensNORM focuses on motion detection using passive infrared technology, or PIR for short. This is the most commonly used technology by far. What are the characteristics of this technology?

PIR technology uses piezoelectric semiconductor crystals. These detect temperature changes in their environment and respond to the body heat of moving people. Another characteristic feature of presence or motion detectors with PIR technology is the lens. It focuses the infrared rays from people who are moving in the field of detection onto the sensors behind the lens. Seen from the outside, the lens structure resembles a honeycomb. These honeycombs divide up a detector's field of detection into several sectors.

How exactly is motion detected?

Motion is only detected once it crosses over the boundary between two such sectors and a difference in heat is produced. The focal length of the lens means the sectors grow as the distance to the detector increases. But as the distance increases, you also need greater motion to cross over a sector boundary. Hence the field of detection is divided up into various sections for needs-based planning. In the outer area, the detector only reliably identifies tangential walking motion. This means motion that runs diagonally in relation to the detector. Further inside, you have a connecting area where the detector also detects radial walking motion running head-on.

METAS, the Swiss metrology institute, is now home to the first independent test laboratory where manufacturers can have their devices tested in accordance with sensNORM. Could you describe the structure and process of the measurement method used there, and explain the content of the standard?

sensNORM has defined and stipulated a measurement method that runs fully automatically. We have defined a test dummy to simulate a real person in the test. It's got legs, a body and a head. The precise individual measurements are specified in the standard.

What material is it made of?

Aluminium. It is coated on the outside with black paint. And inside, various heating plates are attached to these aluminium panels. The dummy's temperature is then adjusted using electronics and temperature sensors and controlled to keep it constant during the measurement. It must always be above the temperature of the test room. The difference in Kelvin is specified exactly and corresponds to the average difference in practice.

And what does the dummy do during the measurement?

Two identically sized dummies are mounted onto two different rails in the test room. The detector being tested is fixed on a rotating platform. The measurement starts by scanning the exterior area, which means the tangential motion at 90° to the detector radius. The first dummy moves on its rail from right to left diagonally in relation to the motion detector, and then back again. The speed, acceleration and deceleration are determined. If the detector is not triggered at the start, the measurement distance between the dummy and detector is gradually reduced. Once something is detected, this walking movement is repeated.

For a positive result, the dummy must be moved three times in total in one direction and three times in the other direction, with the detector being triggered at least twice in each direction. Only then can we say this range is triggered. A similar process is used with radial motion. The only difference is the direction. The second dummy moves on its rail towards the detector until the detector is triggered. It goes through the whole process three times. Then the average is taken of the three values and documented as the measurement result.

And why is the presence or motion detector mounted on a rotating platform?

Depending on what lens geometry you use for the PIR sensors, the range may vary in certain cases in a detector's individual lines of sight. With tangential motion, the ranges don't vary that much, whereas variations are fairly common in the case of radial motion. That's why we rotate the detector 10 degrees after each measurement, both with tangential and radial motion. That's the easiest way to do it.


1. Ceiling-mounted presence detector with 360°-field of detection
2. Test dummy for tangential measurement
3. Test dummy for radial measurement
4. Test arm for presence measurement

For needs-based planning, the field of detection of presence and motion detectors is divided up into several sections. In the outer area (A), the detector detects tangential walking movements running diagonally to it. Further inside, it also detects radial walking movements coming towards it head-on (B). In addition, presence detectors detect micromotion in their presence area (C).

The sensNORM sets out very precisely the requirements for environmental conditions. The room should be big enough, for instance.

It's all about the distances. Where you've designed and built the test room well, as at METAS, the size is irrelevant. The problem only occurs if you have a motion detector that has a range of 40 metres in diameter, for example. For that I'd need a room that is at least 20 metres long. For rooms that don't meet these requirements, we have specifically scaled and recalibrated the test dummy. As a result, we have defined a 50 % dummy and a 20 % dummy. These then allow larger ranges to be mapped in rooms that are actually too small.

sensNORM also includes precise specifications for ambient temperature.

The difference between the temperature of the dummy and its environment must be as realistic and as constant as possible so that we also have reproducible results. I could theoretically use a sports hall. It just needs to be kept at the right temperature with no draughts. The measurement result must not be distorted by external influences. If external sources of interference mean that I introduce thermal radiation in the infrared region into the measurement room, then the measurement results won't be reliable. Strong daylight and direct sunlight must be avoided too. The same applies to the walls.

What do you mean exactly?

Let's say theoretically you have a wall which surface temperature is not homogeneous because there is a heater fitted to it every three metres. If the dummy passes through this area, the detector sees the dummy's temperature in relation to whatever is behind the dummy. If this difference changes simply because the temperature of the walls isn't homogeneous, this substantially affects the measurement results.

This is what the raw data looks like after the measurement In this case, with a presence detector with a 32-metre tangential and 11-metre radial detection range as an example.

Apart from the full-body dummy, there's also a motorised test arm in the laboratory. What's it for?

This is designed specifically for presence measurement. Unlike with motion detectors, presence detectors also include an interior presence area, in addition to the tangential and radial areas. Here, they detect even the tiniest movements. The test arm simulates a person's forearm while sitting at a desk. Normally, the forearm is horizontal, pointing in the direction of the detector. Then it is moved vertically upwards by 90° before returning straight back to the horizontal start position. This is done three times and then we calculate the average. After each individual measurement, the test arm is moved within a stipulated grid in the supposed presence area of the particular detector.

The results are also displayed graphically once all the measurements are complete. In the subsequent communication with the customer, these results always have very uniform shapes. But given the direction-dependent measurement deviations you mentioned earlier, that probably doesn't correspond to reality.

The round fields of detection you see in the planning and communication documentation are measurement results that have been rounded according to a certain rule. As I mentioned, the measurement results can normally deviate wildly in certain cases. But the planner cannot do anything with this raw data. That's why we've created a rule that displays this raw data in an idealised form. This data is then displayed in geometric basic shapes, such as a circle.

What is the rule?

Only 15 % of trigger results may be shorter than the specified idealised line. If I have a circle for a ceiling detector with a field of detection of 360°, which I have measured with tangential and radial measurements every 10 degrees, or 36 times in all, only 15 % of these 36 measurements may be within the idealised circle. With the presence area, the presentation is slightly different due to the grid. In this case, 15 % is the maximum figure for all the measurement results that are not triggered within the idealised presence area.

The test arm is moved within a grid to measure the presence area. The yellow circle shows the idealised line that subsequently facilitates needs-based planning. A maximum 15 % of the measurement results within the circle can fail to trigger a response.

In parallel with sensNORM, the international IEC 63180 has also emerged as a standard for measuring PIR sensors. That is slightly surprising and initially seems problematic. Are they related at all?

Absolutely. The sensNORM group helped shape and substantially influence the wording of the IEC standard from the start. Two people from the sensNORM group were also in the IEC standard working group and collaborated in the drafting process.

How do the two standards differ?

The biggest difference is that the IEC allows a walking test with a real person as an alternative to the automated test, while sensNORM does not. There are people that walk a bit faster and people that walk a bit slower. That means my measurement results are different. That's why as the people involved in sensNORM we believe that a walking test with humans doesn't provide reliable measurement results. sensNORM measurements are automatic only. So it doesn't matter who measures which detector and when. The results are always reproducible.

And why does the IEC differ in this regard?

The IEC version is the lowest common denominator that we could reach in the committee. Those of us involved in sensNORM set out more stringent requirements, but didn't always manage to get them through in the global forum. You normally just have manufacturers sitting on these kinds of standardisation committees. Nonetheless, they do believe that you must use these standards in countries without the financial resources of major industrialised nations. The kind of automatic measuring device we defined can be very expensive. And countries with limited resources are supposed to apply this standard too. We wanted to be able to say to them: OK, this is a standard that you can use, and all you need to do to take the measurements is walk. You just need to make sure that you walk at a constant pace.

Are there any other differences?

On some important issues, sensNORM placed more emphasis on reliability and the relevance of the values to practical applications. But for the IEC, it was more important that the measurements were as inexpensive as possible and restricted to the essentials. With tangential detection, we decided that at least two out of the three back and forth motions of the dummy had to trigger a detection. As such, we measured an area where you can say that the probability of being detected there is very, very high. But you might also find that triggers occur at even greater distances.

In the IEC, one trigger is sufficient, with only two back and forth motions altogether. There's a different way of thinking behind it. The measurements are taken so that you determine the greatest possible distance. In our view, that has less relevance to what happens in practice. With sensNORM, you get a result with a shorter range but this result is considerably more reliable. With IEC, you get a greater range, but you're not sure at all if the detector also spots you effectively if you are even closer to it. It might be that it doesn't spot you at all.

As such, our measurement method is more suitable for practical applications and easier for the planner to use. It's no use for planners knowing that the detector scans 40 metres in the best-case scenario when they want to cover a much smaller space and know for sure that a response is triggered in 20 metres. Why should I care how far the detector can scan? I want it to work reliably where I need it.

The measuring facility at METAS is programmed so that you can use both test methods there. But your answer to the question of whether I as a manufacturer should have my sensors tested according to sensNORM or the IEC standard should be unequivocal.

ESYLUX supports sensNORM and its detectors can of course measure according to this standard. The detector is pre-conditioned before the measurements. Before taking any measurements, this means that the detector is checked to see whether it is triggered erroneously with extreme temperature ranges. This test is much more difficult to pass with sensNORM than according to the IEC's simple requirements.

The detectors are tested twice with sensNORM. Once with the minimum temperature specified by the manufacturer and once with the maximum possible temperature. The detector must not be triggered erroneously within 24 hours. However, the IEC only tests one temperature: the normal room temperature. And the duration is only 12 hours. So the quality requirements are much easier to meet with the IEC. That's why good manufacturers test their detectors according to sensNORM.

Looking back for a moment, what did you find most interesting about your involvement with the standards?

The most interesting thing was working in the international standardisation group because I got to meet lots of interesting people that also have different ways of thinking about measurement methods. It's exciting getting to know people and their different perspectives and mindsets at an international level, being able to travel to other countries and ultimately working together as a team to come up with a joint result despite all the differences.

Mr Riebenstein, many thanks for sharing these interesting insights.