3. Measurement Setup
The measurement setup contains a single signal generator capable of playing back predefined modulation patterns and various RF pico
boards for the different frequency bands to be tested.
169 MHz 4362CPRXB169
434 MHz 4362CPRXB169
868 MHz 4362CPRXB868
All the receiver tests have been performed utilizing the on-chip packet handler. Wireless Development Suite (WDS) can be used to
configure the radio for the given WMBUS modes. WDS configuration files are provided to help configure the radio for various WMBUS
modes. These modes can be open in the Radio Configuration Application. Refer to AN796: Wireless Development Suite General De-
scription and AN632: WDS User's Guide for EZRadioPRO Devices for more details regarding WDS.
Upon sync word detection, the packet handler places a predefined number of bytes into the FIFO. It also does a CRC calculation on the
payload data and compares it to the received CRC bytes located at the end of the packet. If the CRC check is successful, the chip
asserts its packet_received status bit; otherwise, it will assert the CRC_error status bit. During the measurement, a packet reception is
deemed successful if the packet_received status bit is set and is deemed unsuccessful if either the CRC_error status bit is set (corrup-
ted packet) or neither of the above two status bits are set (missed packet).
The test procedure for one packet is as follows:
1. Start receiver (only once at the beginning of the test).
2. Wait one packet length’s worth of time.
3. Fire packet on the generator.
4. Wait one packet length’s worth of time.
5. Check reception. Then go back to step #2.
Packet-related parameters:
• The preamble and sync word length are set according to the requirements of the actual MBUS mode.
• Payload length: 20 bytes
• Payload bytes: 0x0F, 0x44, 0xAE, 0x0C, 0x78, 0x56, 0x34, 0x12, 0x01, 0x07, 0x44, 0x47, 0x78, 0x0B, 0x13, 0x43, 0x65, 0x87,
0x1E, 0x6D
• CRC length: 2 bytes
• CRC polynomial: CRC-16 (IBM): X16+X15+X2+1
Note:
1. Whenever coding is required on the data, the length of the payload is adjusted accordingly.
2. In the case of Manchester coding, the payload is twice as long (40 bytes) in the air.
3. In the case of 3-out-of-6 coding, the payload is 1.5 times as long in the air (30 bytes).
4. The CRC used at the tests does not match the CRC specified in the standard. The CRC check is merely there to help qualify the
packet reception.
For each mode a PER curve, sensitivity vs. frequency offset, deviation offset and DR offset curves (where applicable) are presented in
the next sections. On the sensitivity curves the minimum sensitivity limit (taken from the standard) is drawn as a horizontal red line and
the offset limits are drawn as vertical red lines. The traces should always travel below the horizontal lines in the region bordered by the
vertical lines. Deriving the DR and deviation offset limits are straightforward from the Tx side specifications in the standard. The fre-
quency offset limits, however, deserve a few words here.
In the standard only the Tx side frequency accuracy is specified; we simply need an RX solution that can receive Tx signals with the
extreme frequency offsets. The receiver, however on its own has its own frequency inaccuracy that must also be taken into account.
Throughout the tests conducted for this application note, however, the receiver was calibrated to have close to 0 ppm frequency accu-
racy with regards to the signal generator. So in the test setup we do not have the aforementioned inaccuracy at the Rx side. The resolu-
tion to this issue is that we "place" the Rx inaccuracy to the Tx side and draw the offset limit lines at twice the specification on the Tx.
This logic assumes that the Rx has the same inaccuracy as the Tx. This assumption became a design goal when the receive configura-
tions were put together.
As an example N2a mode requires a ± 1.5 kHz frequency accuracy on the nodes. On the sensitivity vs. frequency offset graph this
number is translated to ± 3kHz as the receiver has no frequency error at all in the tests. This also practically means that in an applica-
tion that the same reference source (XO/TCXO) can be used in the Rx node as specified in the Tx node. In our example it means a ±
1500 [Hz] /169 [MHz] = ± 8.87 ppm reference source accuracy at either side of the link.
This logic is adhered throughout the document at the sensitivity vs. frequency offset graphs unless otherwise stated.
AN805: Si446x Wireless MBUS Receiver
Measurement Setup
silabs.com | Smart. Connected. Energy-friendly. Rev. 0.3 | 17