1) In the design of low-power RF PCBs, standard FR4 materials are used (good insulation characteristics, uniform materials, dielectric constant ε = 4, 10%). It mainly uses 4 ~ 6 layers of boards. In the case of very cost-sensitive double panels with a thickness of less than 1mm, it is necessary to ensure that the opposite side is a complete ground layer. At the same time, the thickness of the double panels is more than 1mm, which makes the ground layer and signal layer The FR4 medium between them is thick. In order to make the RF signal line impedance reach 50 ohms, the width of the signal trace is usually about 2mm, which makes it difficult to control the spatial distribution of the board. For a four-layer board, the top layer normally only uses RF signal lines, the second layer is the complete ground, and the third layer is the power supply. The bottom layer generally uses digital signal lines that control the status of RF devices (such as setting the clk of the ADF4360 series PLL, data, LE signal lines.) The power supply of the third layer should not be made into a continuous plane, but the power supply traces of each RF device should be distributed in a star shape, and finally connected to a point. The power traces of the third layer RF device should not cross the bottom digital lines.
2) For a mixed-signal PCB, the RF and analog parts should be far away from the digital and digital parts (this distance is usually more than 2cm, at least 1cm guaranteed), and the ground of the digital part should be separated from the RF part. It is strictly prohibited to use a switching power supply to directly supply power to the RF part. The main reason is that the ripple of the switching power supply will modulate the signal of the RF part. This modulation often severely disrupts RF signals, leading to fatal results. Under normal circumstances, the output of a switching power supply can pass through a large choke coil, a π filter, and then a linearly regulated low-noise LDO (Micrel's MIC5207 and MIC5265 series. For high-voltage, high-power RF circuits, You can consider using LM1085, LM1083, etc.) to get power to the RF circuit.
3) In RF PCB, each component should be closely arranged to ensure the shortest connection between each component. For the ADF4360-7 circuit, the distance between the VCO inductor on the pin-9 and pin-10 and the ADF4360 chip should be as short as possible to ensure that the distributed series inductance brought by the connection between the inductor and the chip is the smallest. For the ground (GND) pins of each RF device on the board, including the pins that are connected to the ground (GND), resistors, capacitors, and inductors, holes and ground should be punched as close as possible to the pins (second Layer) connectivity.
4) When selecting components that work in high-frequency environments, use surface-mount devices whenever possible. This is because surface-mount components are generally small and the component pins are very short. This minimizes the effects of additional parameters from component pins and internal traces. Especially for discrete resistors, capacitors, and inductive components, using smaller packages (0603 \ 0402) is very helpful to improve the stability and consistency of the circuit;
5) Active devices working under high frequency environment often have more than one power supply pin. At this time, we must pay attention to setting a separate decoupling capacitor near the pin of each power supply (about 1mm), with a capacitance of 100nF. about. Where board space allows, it is recommended to use two decoupling capacitors for each pin, with capacitance values of 1nF and 100nF, respectively. Generally use ceramic capacitors made of X5R or X7R. For the same RF active device, different power pins may power different functional parts of the device (chip), and each functional part of the chip may work at different frequencies. For example, the ADF4360 has three power pins that supply power to the on-chip VCO, PFD, and digital parts. These three parts implement completely different functions and work at different frequencies. Once the low-frequency noise of the digital part is transmitted to the VCO part through the power supply trace, then the VCO output frequency may be modulated by this noise, causing spurs that are difficult to eliminate. In order to prevent this situation, in addition to using a separate decoupling capacitor, the power supply pin of each functional part of the active RF device must be connected together through an inductive magnetic bead (about 10uH). This design is very beneficial to the improvement of the isolation performance of the active mixers LO-RF and LO-IF that include LO buffer amplification and RF buffer amplification.
6) For the input and output of RF signals on the PCB, be sure to use a special RF coaxial connector. The most commonly used is the SMA type connector. For SMA connectors, they are divided into in-line and microstrip. For signals with a frequency below 3GHz, and the power of the signal is not large, and we do not take into account the weak insertion loss, we can use a straight-through SMA connector. If the frequency of the signal is further increased, we need to carefully choose the RF connection wire and RF connector. At this time, the in-line SMA connector may cause a relatively large signal insertion loss due to its structure (mainly turning). At this time, you can use a microstrip SMA connector with better quality (the key is that the connector uses PTFE insulator material) to solve the problem. Similarly, if your frequency is not high, but you are demanding indicators such as insertion loss and power, you can also consider microstrip SMA connectors. In addition, small RF connectors include SMB, SMC and other models. For SMB connectors, this type of connectors generally only support signal transmission below 2GHz, and the snap-in structure used by SMB connectors will appear in high-vibration occasions. "Flashing" situation. Therefore, you should consider it carefully when choosing an SMB connector. Most RF connectors have a limit of 500 mating cycles. Too frequent mating may permanently damage the connector, so when debugging the RF circuit, do not play with the RF connector as a screw. Since the part of the PCB socket of the SMB is a pin structure (male), the connector that is frequently plugged and butt-welded to the PCB end has relatively little loss and reduces the difficulty of maintenance. Therefore, in this case, the SMB connector is also a kind of Good choice. In addition, for those places that require high space, there are micro-connectors such as GDR for selection. For those analog signals with impedances other than 50 ohms, low frequency, small signal, precision DC, or digital signals such as high-frequency clocks, low-jitter clocks, and high-speed serial signals, SMA can be used as a feed-in connector. .
7) When designing the RF PCB, the width of the RF signal traces is strictly specified. According to the thickness and dielectric constant of the PCB, the impedance of the trace at the corresponding frequency must be strictly calculated and simulated during design to ensure that it is 50 ohms (the CATV standard is 75 ohms). However, not all of us need strict impedance matching at all times. In some cases, a small impedance mismatch may not be a big deal (such as 40 ohms ~ 60 ohms); and even if your simulation of the board is based on Ideally, when it is actually handed over to the PCB factory for production, the process used by the manufacturer will cause the actual impedance of the board to be different from the simulation result. So for the problem of impedance matching of small-signal RF PCBs, my suggestion is: Step-1: Communicate with the PCB factory appropriately to obtain the width range of the board with a thickness of 50 ohms for the corresponding thickness and number of layers; Step-2: Select a suitable width within this width range and apply it uniformly to all 50 ohm RF signal lines; Step-3: When the PCB is delivered for production, indicate on the script that all lines of this width do impedance matching for 50 ohms. At this point, there is no need to point out a lot of lines that need to be impedance-matched. (For PCB manufacturers, they will make an impedance strip on the PCB extension you designed, and leave it at the factory. When testing the impedance of a corresponding width sample trace on an impedance bar to roughly determine the impedance of the same width trace on the board. Finally this impedance bar is cut and recycled by the PCB factory without being seen by you). At different frequencies, the impedance of a line of the same width will be slightly different, but this difference is generally within 10%. Of course, you can also write a very complex impedance setting script to let the cardboard factory fine-tune the width of the traces working at different frequencies so that its impedance is strictly set to 50 ohms, and then ask the PCB factory to The root line is filtered. This results in a logarithmic cost increase, and a large number of rejects will be generated; and after such PCBs are installed, the resistance distribution will still cause deviations in impedance due to factors such as solder distribution and the RF components themselves. This kind of situation is extremely rare, because even a precise RF test and measurement instrument, the error caused by the small mismatch of the trace impedance of the RF signal (within 5%) can be easily corrected by software; As for the telecommunications machine, it doesn't even have to pay attention to the 5% difference. But I want to emphasize that for the RF circuits of the LNA (low noise amplifier) and PA (power amplifier) parts, the impedance of the RF trace is very sensitive, but fortunately, whether it is the LNA circuit or the PA circuit, the trace The frequency must be the same, and the number of traces is small (there are only two nodes, input and output). At this time, I suggest that in sensitive occasions, LNA and PA are used as separate boards, and high-quality RF dedicated PCB boards (Rogers / Arlon / Taconics) with uniform dielectric constant distribution are used, and no solder resist oil (also (Referred to as green oil), to avoid the resistance drift caused by soldering resistance; and require the PCB manufacturer to provide an impedance test report. Because the signal power of the input part of the LNA circuit itself is already very small (below -150dBm), the insertion loss caused by impedance mismatch further reduces the valuable signal strength. For PA circuits, because they work at very high power, The insertion loss caused by impedance mismatch can consume a lot of energy (comparatively, the insertion loss is also 1dB: the difference between the energy consumed by 10dBm signal attenuation at 9dBm and 50dBm attenuation at 49dBm, hehe, the latter can generate 20W of heat ) In some PAs with a kilowatt power, the insertion loss of 1dB may bring the effect of fire and splashes.
8) For those RF microstrip circuits that are simulated in ADS, HFSS and other simulation tools on the PCB, especially those directional couplers, filters (PA's narrowband filters), and microstrip resonant cavities (such as yours in Design VCO), impedance matching network, etc., it is necessary to communicate well with the PCB factory, using a plate with strict specifications such as thickness and dielectric constant and the indicators used in simulation. The best solution is to find an agent for the microwave PCB sheet by yourself, and then commission the PCB factory to process it.
9) In RF circuits, we often use a crystal oscillator as a frequency standard. This crystal oscillator may be a TCXO, OCXO or an ordinary crystal oscillator. For such a crystal circuit, be sure to stay away from the digital part, and use a special low-noise power supply system. What's more important is that the crystal oscillator may produce frequency drift with the change of ambient temperature. For TCXO and OCXO, this situation will still occur, but to a lesser extent. Especially those small packaged crystal oscillator products are very sensitive to ambient temperature. For such cases, we can add a metal cover to the crystal circuit (do not directly contact the package of the crystal) to reduce the sudden change of the ambient temperature causing the frequency of the crystal to drift. Of course, this will lead to an increase in size and cost.