Pradeep Shamanna- Principal Applications Engineer, Microchip Technology Inc.
The 2.4GHz band has become a popular choice for short-range radio applications in homes, offices and factories. Usually, 2.4GHz channels are part of unlicensed Industrial Scientific Medical (ISM) frequency bands. Many protocols such as ZigBee (IEEE 802.15.4), Bluetooth (IEEE 802.15.1), Wi-Fi (IEEE 802.11 b/g/n), Wireless Universal Serial Bus (WUSB), proprietary protocols (such as MiWi) and few cordless phones occupy this space. However, operation in the 2.4GHz ISM band induces the radios to interfere with other protocols using the same spectrum.
It is important therefore to evaluate the range and performance of wireless transmission to create models for estimating the path loss for short range modules in indoor and outdoor environments to give designers an initial estimate on a wireless communications system’s performance. The performance parameters include range, path loss, receiver sensitivity, bit error rate (BER) and packet error rate (PER), which are critical in any communications system.
To do this, consider three modules with varied specifications related to power and type of antenna – Microchip’s MRF24J40MA, MRF24J40MB and MRF24J40MC. The MRF24J40MA is a certified 2.4GHz IEEE 802.15.4 radio transceiver module with integrated PCB antenna and is suitable for wireless sensor networks, home automation, building automation and consumer applications. The MRF24J40MB is similar but better suited to longer range applications such as automatic meter reading. The MRF24J40MC has an external antenna (shown in Fig. 1) and also suits longer range applications. All three connect to microcontrollers through a four-wired SPI interface and have various regulatory and modularly certified on board.
Path loss models
Large-scale models predict behaviour averaged over distances. The large-scale model is a function of distance and significant environmental features that are roughly frequency independent. This model exorbitantly breaks down as the distance decreases but is useful for modelling the range of a radio system and rough capacity planning. Small-scale (fading) models describe signal variability on a scale of one to one. They have dominating multi-path effects (phase cancellation). The path attenuation is considered constant but is mostly dependent on the frequency and bandwidth.
However, usually the initial focus is on small scale modelling with rapid change in the signal over a short distance or length of time. If the estimated received power is sufficiently large (typically relative to the receiver sensitivity), which may be dependent on the communications protocol in use, the link becomes useful for sending data. The amount by which the received power exceeds receiver sensitivity is called the link margin.
The link or fade margin is defined as the power (margin) required above the receiver sensitivity level to ensure a reliable radio link between the transmitter and receiver. In favourable conditions (antennas are perfectly aligned, no multi-path or reflections exist, and there are no losses), the necessary link margin would be 0dB. The exact fade margin required depends on the desired reliability of the link, but a good rule of thumb is to maintain 22 to 28dB of fade margin at any time. Having a fade margin of not less than 15dB in good weather conditions provides a high degree of assurance that the RF system continues to operate effectively in harsh conditions due to weather, solar and RF interference.
The path loss due to propagation between the reception and transmission antennas is normally written in dimensionless form by normalising the distance to the wavelength. However, it is sometimes convenient to consider the loss due to distance and wavelength separately. In this case, it is important to track the units being used, since each choice involves a differing constant offset.
As an example, estimate the feasibility of a 1km link (range) with RF nodes one and two of MRF24J40MB modules with 20dBm output power. Node one is connected to an omnidirectional PCB antenna with 1dBi gain, while node two is also connected to a similar PCB antenna with 1dBi gain. The transmitting power of node one is 100mW (or 20dBm) and its sensitivity is -102dBm. The transmitting power of node two is 100mW (or 20dBm) with a similar sensitivity as node one. The cables are short and are approximated with a loss of 1dB on each side. Then add all the gains and subtract all the losses from the node one to node two link considering only the free space loss for a path loss of a 1km link.
Since -60dB is greater than the minimum receive sensitivity of node two (-102dBm), the signal level is just enough for node two to communicate with node one. There is a 42dB margin (102dB – 60dB), which is suitable for good transmission under good weather conditions, but may not be enough to protect against harsh weather conditions.
The path loss is the same on the return path. Therefore, the received signal level on the node one side is -60dB. Since the receive sensitivity of node one is -102dBm, this leaves a fade margin of 42dB (102dB – 60dB). Additionally, there are losses due to environment (fading) even at LoS (line of sight) and could further reduce by 20dB, which is within the requirement for communications without any additional gain.
Now let’s substitute node two with an MRF24J40MA module with 0dB gain (output power). Since the receive sensitivity of node one is -95dBm, this leaves a fade margin of 35dBm (95dB – 60dB). Additionally, there are losses due to environment (fading) even at LoS and can further reduce by 20dB, which communicates only with some additional gain of 15 to 20dB.
The Fresnel Zone is the area around the visual LoS that radio waves spread out after they leave the antenna, as shown in Fig. 2. It is good have the LoS to maintain strength, especially for 2.4GHz wireless systems. This is because the 2.4GHz waves are absorbed by water. The rule of thumb is that 60% of the Fresnel Zone must be clear of obstacles. Typically, 20% Fresnel Zone blockage introduces little signal loss to the link, and beyond 40% blockage the signal loss becomes significant.
It is important to enumerate the extent to which the Fresnel Zone can be blocked. Typically, 20 to 40% Fresnel Zone obstruction introduces little to no interference into the communications link. It is better to have an inaccuracy up to more than 20% blockage of the Fresnel Zone.
The propagation losses for indoors can be significantly higher in buildings because of obstructions such as walls and ceilings. This occurs because of a combination of attenuation by walls and ceilings, and blockage due to equipment, furniture and human intervention.
Trees attenuate around 8 to 18dB of loss per tree in the direct path. This attenuation depends on the size, shape and type of tree. A dry wood wall on both sides can result in about 6dB loss per wall. Comparatively older buildings may have greater internal losses than new buildings due to materials and LoS issues. Concrete walls account to 10 to 15dB depending on the size and shape of the construction. Floors in buildings account for 12 to 27dB of loss. Concrete and steel floors attenuate more than wooden floors. Mirrored walls have very high loss because the reflective coating is conductive.
The Fresnel Zone is sometimes a good indication of an indoor environment range measurement. Generally, the LoS propagation is valid only for about the first 3m. Beyond 3m, the indoor propagation losses can go up to 30dB per 30m in dense office environments. Conservatively, it overstates the path loss in most cases. Actual propagation losses may vary significantly depending on the building construction, structure and layout.
Some of the possible reasons for propagation losses through the Fresnel Zone are collisions with other transmitters, weak error vector magnitude (EVM) from the transmitter generally in the range of 20 to 24% rms, and reflections from moving objects or people.
Take care when choosing the path loss model for predicting the RF system performance. Serious errors can occur by using the free space path loss (FSPL) model for most cases except few restricted cases. A more realistic model to use for urban environments is the ITU indoor propagation model.
For urban environments, the use of 10 to 12dB is a good rule of thumb for predicting the required increase in the link budget to double the transmission distance. Receiver sensitivity is the first variable in a system that must be taken care of and optimised to increase the transmission distance. Other variables in any wireless system also affect distance but must be changed by a greater percentage to equal the effects presented by changing the receiver sensitivity.
Fading due to multi-path can result in a signal attenuation of more than 30 to 40dB, and it is highly recommended that sufficient link margin is factored into the link budget to overcome this loss while designing a wireless system.