Silicon Labs' sub-GHz reference designs for EFR32 chip family utilize an external ceramic balun in the RF-FE matching network. This matching approach is documented and well-detailed in the application note AN923.
However, mostly due to cost reasons, full discrete matching designs might be more desirable. The KBA shows several options about how to apply a matching network for EFR32 with utilizing SMD discrete components only.
Since the EFR32 wireless Gecko has differential TX and RX ports, the matching circuit has to have a balun function too, upon the impedance matching, so the standard 4-element matching balun approach can be applied here as well as shown in the application notes AN369/643 (however, these are discussed with other radios). The impedance goals for the matching network for EFR32 can be found in AN923.
Some extra details for simulations: TX bonding wire inductance is around 2 ... 2.5 nH; RX bonding wire inductance is around 1 ... 1.5 nH. LNA capacitance is around 1 ... 1.1 pF; PA capacitance is adjustable but it is recommended to use the min. value of it for the best efficiency which is around 2.5 pF. Impedance goals are e.g. 125 ohms in TX mode for +20 dBm running at 3.3 V and for +13/14 dBm too running from the on-chip DCDC converter (1.7 ... 1.8 V); and 500 ... 600 ohms in RX mode.
Here is a recommended schematic topology for a full discrete match in TX-RX direct-tie configuration:
- differential-to-differential L-C match for the first section of RX path [LGATE; CSER-1; CSER-2]
- 4-element matching balun approach [L1-1; L1-2; L2; C1; C2] + common mode suppressor [CH] applied on the TX path (+ rest part of RX match) + LPF [CHF0/1/2; LHF0/1]
Here is a recommended schematic topology for a full discrete match in split (separate TX and RX paths) configuration:
- TX path: 4-element matching balun approach [L1-1; L1-2; L2; C1; C2] + common mode suppressor [CH] applied + LPF [CHF0/1/2; LHF0/1]
- RX path: 4-element matching balun approach [LR1; LR2; CR1; CR2]
The split matching configuration can easily be re-used for designs with external SAW filter, FEM or RF switch utilized.
The common mode suppressor (CH) improves the balun function of circuit and can also be tuned for a specific even harmonic (typically 2nd or H2) where enhanced suppression can be achieved by the given notch filters composed by the L1-1 -- CH and L1-2 -- CH series resonances to GND.
The component values of LDC (RF choke inductor), CC (RF bypass - DC block capacitor), LGATE, CSER and LPF elements can be found in the application note AN923.
The simulated component values (simulated only, so bench tuning can likely be required) of the discrete matching balun networks - applicable for both schematics shown above - are summarized here:
TX path: (L1 here below is shown as total value, so L1 = L1-1 + L1-2; while L1-1 = L1-2)
These designs shown above are single- and typically narrow-band solutions, especially in the RX path due to the higher-Q impedance transformation needed. For dual-, multi- or wide-band matching solutions please contact Silicon Labs' Technical Support (and make sure there is an NDA in-placed with Silicon Labs).
The available reference matching networks for the EFR32 wireless Gecko family at the sub-GHz frequency region utilize the so-called direct-tie topology where the TX and RX paths are directly connected to each other without using external RF switch. In order to be able to insert a SAW filter / RF switch or FEM into this matching structure it is recommended to separate the TX and RX paths, since it is not suggested utilizing SAW filter in the TX path, because of:
- the expected power efficiency degradation in TX mode due to the considerable insertion loss of the SAW filters,
- SAW filters are typically designed for low power levels, i.e. would yield TX power limitation,
- SAW filters typically do have weaker attenuation at the higher frequencies, i.e. at the RF harmonics, so discrete LPF would always be recommended.
The recommended schematic structure for EFR32 with SAW filter is shown below.
- TX path match is kept from the reference matches available, recommended component values are shown in AN923 or refer to the existing reference designs.
- The order of the LPF section may be changed based on the power level, harmonic suppression requirements.
- RX path match utilizes a standard 4-element discrete matching balun network, similarly as detailed in AN643. Simulated component values are shown in the table below.
- RF switch is also being used in order to separate the TX and RX paths' matches, while being connected to the same antenna port.
- SAW filter's separate matching network may not be needed (LW1, LW2, CW1, CW2, CW3 and CW4) - refer to the given SAW filter datasheet.
For designs that use RF switch only, then the above schematic can be used without having the SAW filter and its matching components (LW1/2, CW1/2/3/4) mounted. However, a series RF bypass / DC blocking capacitor can be suggested between the 4-element discrete balun SE port and RF switch.
A typical design with FEM is being shown in the figure below. Here, the RF switch can basically be replaced by the FEM while the SAW filter is not shown below, but can be utilized between the FEM RX ports, if applicable. The recommended placement of SAW filter in FEM (with LNA) designs is between the separate RX ports of FEM that ensures the following order of RF blocks in the receiver path: Antenna --> discrete LPF (required due to TX harmonic suppression reasons) --> SAW filter --> FEM LNA --> 4-element discrete matching balun network --> EFR32 LNA. This approach will ensure robust receiver operation in an even noisy environment. Despite the fact that in a non-noisy environment, better link budget could be achieved if the SAW filter were placed between the 4-element matching balun and FEM LNA, Silicon Labs do recommend to use the approach described above, since introducing LNA in the receiver path in general yields however better sensitivity, but worse linearity and blocking performance. If the non-noisy environment can be ensured then the SAW filter is not necessary in the design.
For discrete matching solution on the TX path as well (i.e. eliminating the external ceramic balun between the EFR32 and FEM), please refer to the following KBA link reference.
The achievable RF range is affected by many factors as listed below.
- Transmit power and TX antenna gain
- Receive sensitivity and RX antenna gain
- Frequency: It is related to the gain and effective area of the antenna. It can simply translate into that the lower frequency the link operates at the better RF range can be achieved.
- Antenna radiation pattern: The best RF range can be achieved if the TX and RX antennas are facing to each other in their maximum radiation lobes. There could be some directions where the antennas' radiation patterns do have minimum notches and thus the RF range could be poor in these directions.
- Interference, noise: Any in-band noise does have severe negative effects on the range since it can mask out the wanted signal at the RX side (see the co-channel rejection parameter). But, stronger out-of-band noise can also degrade the RF range based on the receiver's ACS and blocking performances.
- Frequency offset between TX and RX: It can become more critical in narrow-band systems where the exact carrier frequencies must correctly be set.
- Final product placement and enclosure: The antenna performance can be affected by any material in the close proximity of the antenna and by the antenna placement. In order to avoid any de-tuning effect (and thus RF range degradation) make sure about the recommended antenna (or i.e. module) placement and clearance.
- Environment: Ideal case is an outdoor environment where there are no reflections (e.g. no walls, big obstacles, trees, houses) and there is a direct line-of-sight (LOS) between the TX and RX and there isn't any obstacle in the Fresnel ellipsoid too (see online calculators for the Fresnel zone/ellipsoid). Less ideal case is an urban area, or when there is no LOS between the TX and RX. The worst situation is an office indoor environment where there is typically no LOS and there are walls, obstacles and thus reflections. Propagation constant can describe the environment which is typically 2.5...3.5 in an outdoor environment with LOS between the TX and RX nodes, while can even be 4...6 in an indoor environment.
- Transmitter and receiver heights: This is also related whether there is any obstacle, e.g. ground, in the Fresnel ellipsoid. If so, the RF range is negatively affected. Thus, the higher the nodes are placed at the bigger RF range can be achieved.
See a related KBA link on this topic below which describes an example estimator/calculator for the RF range.
How should I route the traces on more-layer RF designs for optimal performance?
In order to achieve the possible best RF radiated performance the followings are suggested for more-layer RF board designs:
- Top Layer: Components and short traces. Top layer should use as large and continuous GND plane metallization as possible (with many stitching GND vias) on the entire PCB.
- 1st inner layer: GND plane and traces if necessary. The most important rule is to keep the GND pour metallization unbroken beneath the RF areas (between the antenna, matching network and RF chip). Traces can be routed under the non-RF areas and use GND pour where possible.
- 2nd, 3rd... inner layers: Traces. VDD and all other traces are suggested to be routed on these layers. Use GND pour where possible.
- Bottom Layer: GND plane. Use as large and continuous GND plane as possible. Do not route traces on this layer, just if it is necessary, e.g. short connection traces to connectors.
- Generic for each layer: Try to avoid routing traces along or close to the board edges. It is recommended to place ground stitching vias with GND pour along the PCB edges.