Author: Aleksej Cegner, Electronics Engineer
IRNAS has been actively involved in environmental and animal conservation since its founding in 2014. Last year, we have been working with Smart Parks to develop a modular LoRaWAN® GPS tracker for use as a collar payload on wildlife like Lions, Wisents, Elephants and also park rangers and vehicles.
Of course, one of the highest priority features of such a device is the ability to withstand extreme weather conditions as well as regular mechanical impact due to everyday animal behaviour. These devices are with the animals 24/7 and are something the animals learn to live with. Human interaction with the devices occur only during mounting and dismounting, done by trained veterinarian professionals.
Because the device has a high exposure to environmental and impact factors, we needed to ensure high mechanical robustness. One of the techniques we often rely on is potting electronics in an epoxy resin (keep an eye out for more on this topic in one of the following blog posts). However, this technique brings some challenges with regard to the RF signals that travel in and out of the potted enclosure. We now need to keep an extra eye on good antenna tuning.
In case of the so-called “Lion tracker”, we are operating on the 868 MHz frequency and from our first antenna performance tests, we saw a frequency shift of 100 MHz compared to a free-air radiating set-up. Based on our positive experience with the Antenova SR42I010 antenna from the past, we started to tune it for this particular application.
Potting the electronics in an epoxy resin was only an addition to the list of regular challenges that usually come our way in antenna tuning, like having a rechargeable Li-Ion battery near the antenna and an EMC shield can (we will talk about why it was necessary to use it later). The SR42I010 antenna uses an inductor on the band select pin in order to finetune the frequency. According to the datasheet, a frequency of 868 MHz should be achievable by using an inductor of 4.2 nH towards ground, but keep in mind that this is on Antenova reference board. From the datasheet and very responsive technical support from Antenova we discovered that using a lower value inductor gives us a higher resonance frequency but the problem was that even with an 1 nH inductor we could only shift the frequency 50 MHz upwards in a frequency spectrum which was still far from 868 MHz. You may find the following article very informative by Qorvo “4 Things to Know About Antenna Tuning in 4G / 5G Smartphones” which explains why antenna tuning is needed and how to use the right component.
In our case, we have used a capacitor instead of an inductor, which gives the results we were looking for. The only downside of using a capacitor was a worse S11 parameter of around 5 dB, compared to the Antenova reference board, but we could still stay within the specified signal efficiency and consequently signal range.
Now we turned to our next challenge: the EMC shield. Because RF components and traces are typically impedance matched to air, potting the circuit changes the relative permittivity of the surroundings and the impedance drastically changes. You could decide to tune the electronics to compensate that, however for us it was not a practical solution because the insides of the MURATA CMWX1ZZABZ-091 module we use here can not be adjusted. Therefore, we have decided to try and use an EMC shield that can create an air cavity around the RF components. This way we have positioned all RF components inside the EMC shield including the MURATA CMWX1ZZABZ-091 LoRa module and the U-Blox ZOE-M8G GPS receiver. Since the battery and EMC shield are soldered manually and it is impossible to ensure repeatable positioning i.e. achieve the same distance to the antenna, we have added an auto-tuning circuit with VSWR measurements.
For the auto-tuning capability, we have chosen the TDK HHM22106C1 for VSWR measurements and the pSemi PE64102B-Z digital tunable capacitor to be connected in parallel with a fix value capacitor near the antenna, following the advice of Fabien Ferrero, an expert in antenna design (check out his great lectures). This solution provided us with some significant benefits, like tuning the antenna frequency for each device separately. Now each device will always work in optimum conditions, plus anything that is added to the potted device and detunes the antenna, the device can fix itself automatically (within reasonable limits, of course).
We hope you have enjoyed this short insight into the making and performing of this LoRaWAN® GPS tracker. If you would like to know more about the use and performance of these devices in the field, please check the website of Smart Parks and the OpenCollar Initiative.
By request of Smart Parks, this project is offered fully open-source and can be found on GitHub page.
About Smart Parks
Smart Parks is a Dutch social enterprise based in Utrecht. A Smart Park is a nature reserve that uses smart sensor technology to collect information for the improvement of nature protection and management. This may include information about movement patterns of humans and animals, but also other issues, such as rainfall and temperature. Sending this data is via a LoRaWAN® network, among other things; a Long Range, Low Power Internet of Things connection, with which simple data packets can be sent. Due to the low energy consumption of this network, sensors can be developed with a long battery life. This makes their use very suitable for monitoring animals, people and material in nature reserves.
Smart Parks is active in the Netherlands, Tanzania, Malawi, Rwanda, Kenya, Zambia, Namibia, Congo and India.
Explore pictures about our campaigns: https://www.smartparks.org/press/