How the O3 Senses Temperature

How is it possible to measure the temperature at occupant height when the O3 is mounted on the ceiling?

The O3 is actually modelling the occupant height temperature based on the readings it gets from the three internal temperature sensors. Two of the sensors are traditional temperature sensors. They are directly measuring air temperature up at the ceiling. The third sensor is an infrared sensor, which measures a large area directly underneath the sensor hub. This IR sensor covers an area that is roughly the diameter of the mounted height. For example, if you mount the unit on a 2.4 m (8 ft) ceiling, the IR sensor covers roughly a 2.4 m diameter. At 3 m (10 ft), it’s looking at 3 m diameter.

So it’s true, we’re not really measuring the temperature at occupant height. We’re modelling what the temperature will be down at the occupants level based on everything the sensor hub senses up on the ceiling.

How It Works

Using a technique called Kalman filtering, we’ve produced a model for temperature measurements. We take the three temperature readings and feed them into the algorithm in real time. Each sensor reading is given a weighting towards the final temperature value. It’s not just the current measurements that matter, the previous temperature readings are also fed into the model with their own weighting.

By running this model constantly, we achieve two things. The O3 is able to respond to temperature changes much more quickly than a traditional thermostat while also rejecting noise - that is, sudden spikes in temperature due to random events. Using this approach, we get the best of both worlds. We see the temperature rise or fall as it’s really happening, without confusing the device because somebody walked underneath it.

In practice, what the modelling does very well is track the changes in temperature. The O3 makes an initial prediction as to what the temperature is, then follows along as the temperature moves up and down from there. However, because each room is different (air flow, heating sources, height of the ceiling, etc.), that initial prediction is likely to be off. We find that it is typically off by one to two degrees Celsius. Fortunately, this offset is fairly constant and can be calibrated out.  For details on the measurement method refer to the whitepaper published in the New Horizons in Civil Engineering Conference 2020 - O3 White Paper.

Frequently Asked Questions

Can I read temperature outside the specified range of the O3?

Yes, the O3's area of influence is specified as the circular area under the device with the diameter equal to the mounting height of the hub. If you go outside this area the temperature will not be unknown as the temperature will propagate into the area of influence of the O3.

What if sunlight shining through a window hits a desk or other area in the O3’s field of view? Won’t this affect the temperature readings?

If the sunlight is shining for a significant amount of time, and covers enough area, yes it will be picked up by the unit. Bear in mind, though, in this situation the temperature in the space would actually be increasing anyway. This is not the same as a thermostat being exposed to direct sunlight for an extended period of time. Sunlight directly on a thermostat heats up the thermostat itself, which can cause false readings. Sunlight on an area under a sensor hub can cause the room itself to heat up, which will eventually be shown by the O3.

Isn’t the IR sensor actually measuring the surface temperature of the floor or furniture beneath the O3?

Not exactly - the IR sensor is measuring heat energy. Some surfaces do a very good job of reflecting heat energy and therefore show up to the sensor, but it is not a direct measurement of surface temperature.

What happens if a heat source is moved into the O3’s measurement range after calibrating?

This will have an effect on your temperature reading accuracy. How much error is dependent on the amount of heat that is produced. If it’s a small source that’s not active very often, it won’t be noticed. If it constantly produces a significant amount of heat, then you would need to recalibrate under the most common conditions.

What happens if I calibrate with a large number of people in the space, but in normal use there is only one person using the room?

Your calibration will be off. The extra people will show up as extra heat energy when you calibrate. How much extra is almost impossible to predict. The best approach is to calibrate as close to your set point as possible with the space being used as it would be under normal conditions.

Can I calibrate the temperature reading outside of the range of the IR sensor?

This is not recommended. The IR sensor is a major component of the temperature model and calibrating outside of its measurement area means that there is a weak relationship between the point you’re controlling to and what goes into the model.

Can I use the O3 with in-floor radiant heating?

Yes. For best results calibrate the O3 at occupant height when the radiant heated floor is on.