 Hello, welcome to RF guidelines for STM32WB session. Watching this video you get basic information about the needs and design of RF circuit for STM32WB. These are frequently asked questions about correct design respecting standards compliance. In this part we will show what the RF part of STM32WB consists of and we will explain function of each block. This is described in the schematic. The RF part of STM32WB consists of these blocks. The chip consists of whole transceiver that ends with ballon so it has only one RF output that incorporates transmitter and receiver. It is as simple as possible to ease the hardware implementation. It is followed by a few components that is not able to integrate into chip. It is matching network that serves for matching output of the chip to the input of the filter. The filter suppresses an unwanted signal comes from the chip or into the chip. It can be low pass or band pass. Low pass filter has advantage in lower insertion loss. The standard filter should be matched to the output of the chip otherwise the transmission characteristic of the filter would be influenced by wrong impedance. This matching network should be as close as possible to the chip. It is followed by further matching network due to matching output of the filter to the antenna. Antenna is the most sensitive part of RF path. Antenna's performance and parameters depend on surroundings, proximity of the metal or absorptive materials. The antenna should be matched in operational state. In these conditions the antenna has certain reflection coefficient and the matching network has to respect its impedance. Now we show this RF path in real implementation on the application board MBE1355C. Significant parts of the schematic described in the following slides are based on this board. Dimension schematic represents the RF path of nuclear board. The RF output can be routed to the printed antenna if C35 is soldered or to the SMA connector if C38 is soldered. Note that the SMA connector is not mounted by default. The capacitors C35 and C38 are not necessary in your application. These serve as switch of the output path. If you want to preserve this switch in your application just to verify that it hasn't impact on RF performance, for example filtering, sometimes the connector is useful for measuring testing or for certification process. As we mentioned, the STN32WB requires a filter between its output and the antenna to ensure that its transmitted harmonics are compliant with the FCC regulation. Clearing rejections has to be fulfilled for passing the certification process. These are quite strict therefore you have to care about well performed matching and filter selection. The integrated filter on nuclear board is a band pass filter more selective than the low pass filter. Thanks to this filter it is possible to use the nuclear board in an environment more disturbed by neighbor communication bands in presence of higher insertion loss. Now we will note another important part of design. The STN32WB incorporates the switch mode power supply operation. There are several requirements for components and way of using. If you will use this SMPS function you have to care about right selection of components and their correct placement. We will focus on layout in detail in second part of this video. The inductor L1 should be selected as 10 microhenry associated with a bulk capacitor or 4.7 microfarad for 4 MHz operation. If you choose 8 MHz operation then the inductor L1 should be selected as 2.2 microhenry. The bulk capacitor 4.7 microfarad remains the same. It allows a smaller footprint especially very low profile inductor. For all packages it is advised to add an extra 10 nanohenry inductor L2 in series with the 10 microhenry or 2.2 microhenry 1. This is needed to filter the RF harmonics that can degrade the receiver performance. Always you should check if the SMPS frequency isn't in high levels in the output spectrum. You will find them in 8 or 4 MHz from the center frequency. In this part we will deal with crucial components of the RF system, of their selection and correct placement. So let's do it. You can meet with two types of filters implemented in STN32WB boards. Here are some basic facts about the filters. This information could help to select right filter into your board. Ceramic filter is usually made as low temperature cofiled ceramics or based on ceramic oxides. The structure is low pass or bent pass. In fact the main feature is that spurious and harmonics has to be decreased. It has larger package in comparison of IPD filter we talked about after a while. Also necessary matching network increases the needed space. Ceramic filter is need to connect to 50 ohm for proper functionality. If not correctly matched it decreases the stop band rejection. IPD integrated passive device is the structure made on glass. It consists of matching network for chip STN32WB and low pass filter with high rejection of harmonics and spurious. Need to be connected to 60 ohm at the input and 50 ohm at the output. The input RF track must fulfill the certain length because it is a part of matching network. It is a bit dependent on thickness of the substrate. It is minimalistic solution that takes account the rejection is stop band of the filter and high performance in Bluetooth band. More details are mentioned in second part of this video. What about the differences in implementation of both filter types? Let's have a look. We will start with Ceramic filter. The design of the matching network between output of the chip and the filter must fulfill optimal matching between non 50 ohm output of the chip to the filter. The matching network has to respect the power adjustment of the output either the impedance of the filter. That is most important component for rejection of unwanted spurious and harmonics. Please note the output of the chip is usually measured in power on mode with maximal power level. The design of the matching network between the filter and the antenna has to respect the impedance of the filter and the impedance of the antenna. The antenna is very sensitive therefore it should be measured in operation conditions. It is good to check also the transmission characteristic of the whole RF trace. It means the attenuation of the trace from the chip to the antenna and unwanted signals rejection capability. Because the antenna matching often influence the filter its stop band mainly. Performing an impedance matching at the output of the filter often moves the impedance at its input. So it is advisable to make a subsequent correction of the input matching network. Such simpler case is IPD implementation. There is no matching network between output of the chip and the IPD filter. Only the length and impedance of the track must be maintained. The matching network is inside the APD and this input track is part of the matching network. As in previous case the output impedance measurement is recommended. And the rest is the same. The design of the matching network between the IPD filter and the antenna has to respect the impedance of the filter and the impedance of the antenna. The antenna should be measured in operation conditions. The advantage is that if input line is designed well the output is close to 50 ohm and for many designers it is much straightforward way to propose matching network. Also in this case it is highly recommended to check not only transmission characteristic of the passband but also rejection performance of the filter. Now we will discuss the possibilities in antennas. It is often neglected component although one of the most important. If we will talk about PCB antenna we mean DIPOL or F-antenas not those that request ground plane under it as patches. It is very difficult to say whether it is preferable to use a chip antenna or a PCB antenna. Each has its pros and cons however for selection it is important to know this. The PCB antenna will always be slightly larger due to the dielectric constant of the substrate. However it can be optimized for given conditions of use and can be placed in non-standard positions. Even a metallic or absorptive surrounding can be compensated. However this requires an appropriate simulation tool and qualified operators. Chip antenna is usually smaller however be careful about fine tuning stops of the resonance frequency. For a given space and location you will need to use a matching network to fine tune the resonance frequency. However its use is simple. For optimal functionality it is necessary to keep the layout of the manufacturer. Its modification has impact on performance. Now a couple of basic rules for matching and placement. The antenna should be matched by matching network that is as close as possible to the antenna. Commonly P or T structures are used. As you can see from picture it is placed just at the antenna port. The antenna itself is necessary to place in the corner or on the edge of the board. Placing inside the board affects the antenna's performance. The reason is metal plane around. It should not be disturbed by any metallic part that impacts the impedance of the antenna. It is valid for PCB either for the chip antenna. These types of antennas have to have sufficient ground plane. The ground behaves as a mirror and it is necessary condition for its functionality. Small ground plane or small PCB significantly impacts the efficiency of the antenna. We can say that it is the best to keep the layout of the manufacturer. But be careful, under the antenna must not be any metal. It means PCB layers. Also around the antenna must be preserved space for no or negligible impact. Antenna has to have enough space for radiation. Take care about absorptive materials close to the antenna. It has attenuation and so it decreases the distance range of the link. We always target to maximize radiated power. You often meet with plastic case or enclosure. Also know that plastics aren't significant if they don't touch the antenna. The matching is an integral part of the RF path. Without it the RF system would not be able to operate well. Although this is probably an annoying part for many developers it's a necessity. Here is the example of the Smith chart. These define four areas in Smith chart A, B, C and D. Each impedance that lays inside the certain area is possible to match by certain structure by matching network. Fold matching networks are the most simple ones. In most cases it is enough. These matching networks are commonly used. For each area of Smith chart is determined to certain network to match to the normalized impedance, in this case 50 ohm. P or T structure can be used for realization of more complicated functions. For example DC separation or father filtration or higher order filtration. However, lower stability and higher loss are to be expected. It is basically a transformation of complex impedance to another complex impedance. Here we work with reactance components that do not have parasitic properties. In the real world components affect the result a bit. They add an attenuation and you need to measure written loss and fine tune component values. Here is the example of simple matching procedure. Let's have load impedance ZL 30.4-J32 ohms for 2.45 GHz. Adjust this impedance to 50 ohm. We will do the transformation of load impedance. First we mark the load impedance in Smith chart. The impedance have its real and imaginary part. You will find them as an intersection of appropriate circle and curve. We need to get from this point to the circle of constant conductance. It is possible to do it by connecting the serial conductor. So we move along the circle of constant resistance. Then we need to get to the middle of the Smith chart. It can be done by moving along the circle of constant conductance. So the parallel capacitor is need to be connected. And that's it. Of course it is matched only in this point on other narrow frequencies. Is the matching result a bit worse? In this case we will match the impedance 44-J60.7 ohm to 28.7-J4.3 ohm. It is matching of complex impedance to another complex impedance. It means that we must transform the load impedance ZL to the complex conjugated impedance ZG. It is 28.6 plus J4.3 ohm. First we move along the circle of constant conductance by parallel inductor. And then move along the circle of constant resistance by serial capacitance to complex conjugated point of wanted impedance. We have fulfilled the matching law. The real parts are equal and imaginary parts are with opposite sign. The first case is in fact the same. Only the imaginary part of the wanted impedance is equal to zero. This is the summary of consequences of poor matching. Poorly imperfect matching or no matching are often the reason for poor performance. When looking for an error on a real sample, remember to check the adjustment. Let's list several issues we can meet with. Mismatch power losses due to reflections. The power reflects to the source and only part of the power comes through. The amount of the power depends on the return loss, so decreased output power is the consequence of poor matching. Less rejection of harmonics is caused by input or output mismatch. Because it influences the transmission characteristic of the filter. The impact on pass bound transmission characteristic of the filter is observable in large mismatch cases. It shows as change of output power over the frequency. Less radiated power from the antenna due to less power come into it. Resonance frequency of the antenna can be shifted by matching network. Why not with poor matching? Poor matching can even impact the internal bound, its transmission. There's sensitivity of the receiver due to higher attenuation, uneven transmission change in phase or additional resonances. You can see that it is high important to take care about matching. Poor or no matching leads to degradation of RF performance. If it's difficult for you, don't worry. There's a lot of tools on the web for designing an RF path. They are very useful and powerful. Of course, there are also a couple of constraints. There are some assumptions you have to know as parameters of each component used in RF path to be possible to adjust them quite fast and accurately. In case of active component, the sparameter file is in appropriate power supply and power level condition. In most designs, you need to measure them. If you have them all, then it is quite simple work. Even these tools have the tuning and optimization features. I recommend to not use optimization tool to avoid misleading. But what about the antenna? The antenna radiates the signal. It behaves as EM generator and it is almost impossible to describe it by sparameter file. There are many surrounding effects that influence the impedance and resonance frequency. Using the antenna requires the measurement of reflection at the antenna port in working conditions. For these, the calculated matching network is only valid. And how to deal with stn32wb output? The output reflection of the chip is changing versus power level. Therefore, even a reflection at the chip output has to be fine out for working conditions. It is much simpler work, if you consider IPD, that it has matched output when well provided its input connection to the chip. I'm looking forward to you by second part of this session. We will deal with PCB layout and several hints for correct ARF design. Thank you for watching.