From the device manufacturer’s perspective, sensitivity measurements are typically done in a calibrated shield room with an RF signal generator. All system losses are accounted for and the sensitivity measurement is rolled back to the pin of the device. This is a very controlled environment avoiding any outside interference the can impact this measurement. This is not that practical from the customer’s standpoint. Customers or OEMs usually do not have specialized RF test equipment nor large shield rooms to house at the equipment and boards to be tested.
About the software application and tools:
Freescale provides test software (Connectivity Test App) to perform simple RF testing allowing the User to make the same sensitivity measurement with the Connectivity Test App delivered within BeeKit in the SMAC folder. BeeKit is a standalone software application that provides a graphical user interface (GUI) in which the user can create, modify, save and update wireless network solutions based on Freescale's protocol stacks.
About the reference design hardware:
Most hardware reference designs offer two RF output paths to either a SMA connector or to a PCB trace type antenna since RF measurements can be made by either. However the direct connection (SMA) is highly
recommended because over-the-air measurements (trace antenna) are very difficult to perform as well as quantify since over-the-air performance is highly dependent on antenna efficiency, direction, distance between device nodes and environmental interference. The purpose is to measure the quality of the RF signal being feed to the antenna.
Test setup and procedure:
Any simple sensitivity measurement requires an adjustable attenuator, RF cables and two boards; one used for transmitting the RF signal (TX) and the other for receiving the signal (RX). Although any of the IEEE802.15.4 reference hardware can be used it is recommended to use 2 TWR-KW2x boards since this is the topic of discussion. The hardware comes preprogrammed with the Connectivity Test App so Users can make the sensitivity measurements simply by selecting the Packet Error Rate test (PER) within the tool menu using any PC terminal program. If an adjustable attenuator is not attainable, inline attenuators can be used and stacked as required.
The RF path loss between the SMA port and the balun is 1 dB for each board. Also there will be up to 1 dB loss in any cables used between boards and attenuator being used for the test. So for 2 boards and cables there is about 4 dB loss which has to be accounted for in the final measurement. The receiver board will need to be isolated from the transmitter board and any other interference. There are small RF chambers that can be purchased to assist with isolation.
To help with both isolation and attenuation levels, set the TX board to its lowest power level (about -31 dBm, PA step = 03) and adjust the attenuator until the PER is below 1% while running the PER test with 1000 packets. Once this adjustment is made, run the PER test several times to validate the attenuation settings.
As an example PER measurement with the TX board set to -31 dBm and adjusting the attenuator to -65 dB then adding in the path loss of 4 dB, this equates to a Sensitivity (PER) = -100 dBm (at the balun or delivered to the PCB trace antenna). This measurement is not perfect but can give the users some idea of how they can test their own design if calibrated RF measurement equipment and a shield rooms are not economical.