Wearable PCB design should be cautious, these three major issues need to be considered
Due to their small size and size, there are few ready-made printed circuit board standards for the growing wearable IoT market. Before these standards came out, we had to rely on the knowledge and manufacturing experience learned in board development and think about how to apply them to unique emerging challenges. There are three areas that require our special attention. They are: circuit board surface materials, RF / microwave design, and RF transmission lines.
PCBs generally consist of laminates, which may be made of fiber-reinforced epoxy (FR4), polyimide, or Rogers (Rogers) materials or other laminates. The insulating material between the different layers is called a prepreg.
Wearable devices require high reliability, so when PCB designers face the choice of using FR4 (the most cost-effective PCB manufacturing material) or more advanced and more expensive materials, this will become a problem.
If wearable PCB applications require high-speed, high-frequency materials, FR4 may not be the best choice. The dielectric constant (Dk) of FR4 is 4.5, the dielectric constant of more advanced Rogers 4003 series materials is 3.55, and the dielectric constant of brother series Rogers 4350 is 3.66.
Figure 1: Laminated view of a multilayer circuit board showing the FR4 material and Rogers 4350 and the core layer thickness.
The dielectric constant of a stack refers to the ratio of the capacitance or energy between a pair of conductors near the stack to the capacitance or energy between the pair of conductors in a vacuum. At high frequencies, it is better to have a small loss, so Roger 4350 with a dielectric coefficient of 3.66 is more suitable for higher frequency applications than FR4 with a dielectric constant of 4.5.
Under normal circumstances, the number of PCB layers for wearable devices ranges from 4 to 8 layers. The layer construction principle is that if it is an 8-layer PCB, it should be able to provide enough ground and power layers and sandwich the wiring layer in between. In this way, ripple effects in crosstalk can be kept to a minimum and electromagnetic interference (EMI) can be significantly reduced.
In the circuit board layout design stage, the layout plan is generally to place a large block of ground close to the power distribution layer. This can form a very low ripple effect, and the system noise can be reduced to almost zero. This is especially important for RF subsystems.
Compared with Rogers materials, FR4 has a higher dissipation factor (Df), especially at high frequencies. For higher performance FR4 stacks, the Df value is around 0.002, which is an order of magnitude better than ordinary FR4. However, the Rogers stack is only 0.001 or less. When FR4 materials are used in high-frequency applications, significant differences in insertion loss can occur. Insertion loss is defined as the power loss of a signal transmitted from point A to point B when using FR4, Rogers, or other materials.
Wearable PCBs require more stringent impedance control. This is an important factor for wearable devices. Impedance matching can produce cleaner signal transmission. Earlier, the standard tolerance for signal-bearing traces was ± 10%. This indicator is obviously not good enough for today's high-frequency and high-speed circuits. The current requirement is ± 7%, and in some cases even ± 5% or less. This parameter and other variables can severely affect the manufacturing of these wearable PCBs with particularly tight impedance control, which in turn limits the number of merchants that can manufacture them.
The dielectric constant tolerance of the stack made of Rogers UHF material is generally maintained at ± 2%, and some products can even reach ± 1%. These two materials can find Rogers' insertion loss to be particularly low. Compared with traditional FR4 materials, the transmission loss and insertion loss of Rogers stacks are half lower.
In most cases, cost is the most important. However, Rogers can provide relatively low-loss high-frequency stacking performance at an acceptable price. For commercial applications, Rogers can be used with epoxy-based FR4 to make hybrid PCBs, some of which use Rogers materials and others use FR4.
When choosing a Rogers stack, frequency is the primary consideration. When the frequency exceeds 500MHz, PCB designers tend to choose Rogers materials, especially for RF / microwave circuits, because the above traces can provide higher performance when the impedance is strictly controlled.
Compared with FR4 material, Rogers material can also provide lower dielectric loss, and its dielectric constant is stable over a wide frequency range. In addition, Rogers materials can provide the ideal low insertion loss performance required for high frequency operation.
The thermal expansion coefficient (CTE) of Rogers 4000 series materials has excellent dimensional stability. This means that compared to FR4, when the PCB undergoes cold, hot, and very hot reflow cycles, the thermal expansion and contraction of the circuit board can be maintained at a stable limit at higher frequencies and higher temperature cycles.
In the case of mixed lamination, Rogers and high-performance FR4 can be easily mixed together using general manufacturing process technology, so it is relatively easy to achieve high manufacturing yields. Rogers stacks do not require special via preparation processes.
Common FR4 cannot achieve very reliable electrical performance, but high-performance FR4 materials do have good reliability characteristics, such as higher Tg, still relatively low cost, and can be used in a wide range of applications, from simple audio design to Complex microwave applications.
RF / microwave design considerations
Portable technology and Bluetooth pave the way for RF / microwave applications in wearables. Today's frequency range is becoming more dynamic. A few years ago, very high frequency (VHF) was defined as 2GHz ~ 3GHz. But now we can see ultra-high frequency (UHF) applications ranging from 10GHz to 25GHz.
Therefore, for wearable PCBs, the RF part requires closer attention to wiring issues. Signals must be separated separately to keep the traces that generate high-frequency signals away from ground. Other considerations include: providing a bypass filter, adequate decoupling capacitors, grounding, and designing the transmission and return lines to be nearly equal.
The bypass filter can suppress the ripple content of noise content and crosstalk. Decoupling capacitors need to be placed closer to the pins of the device carrying the power signal.
High-speed transmission lines and signal loops require a ground plane between the power plane signals to smooth the jitter generated by noise signals. At higher signal speeds, small impedance mismatches can cause unbalanced transmission and reception signals, which can cause distortion. Therefore, special attention must be paid to the impedance matching problems related to RF signals, because RF signals have high speed and special tolerances.
RF transmission lines require impedance control in order to transmit RF signals from a specific IC substrate to the PCB. These transmission lines can be implemented in the outer layer, the top layer and the bottom layer, or they can be designed in the middle layer.
The methods used during PCB RF design layout include microstrip lines, suspended striplines, coplanar waveguides, or ground. A microstrip line consists of a fixed-length metal or trace and the entire or partial ground plane directly below it. The characteristic impedance in general microstrip line structures ranges from 50Ω to 75Ω.
Figure 2: Coplanar waveguides provide better isolation near RF lines and those that need to be routed close together.
Suspended striplines are another way to route and suppress noise. This line consists of a fixed-width wiring on the inner layer and a large ground plane above and below the center conductor. The ground plane is sandwiched between the power planes and therefore provides a very effective grounding effect. This is the preferred method for routing wearable PCB RF signals.
Coplanar waveguides can provide better isolation near RF lines and lines that need to be routed close together. This medium consists of a central conductor and ground planes on or below it. The best way to transmit RF signals is to levitate striplines or coplanar waveguides. These two methods can provide better isolation between signals and RF traces.
It is recommended to use so-called "via fences" on both sides of the coplanar waveguide. This method provides a row of ground vias on each metal ground plane of the center conductor. The main traces running in the middle are fenced on each side, thus providing a shortcut to the underlying formation for the return current. This method can reduce the noise level associated with the high ripple effect of RF signals. The dielectric constant of 4.5 remains the same as the prepreg FR4 material, while the prepreg—from microstrip, stripline, or offset stripline—has a dielectric constant of about 3.8 to 3.9.
Figure 3: Via fences are recommended on both sides of a coplanar waveguide.
In some devices that use the ground plane, blind holes may be used to improve the decoupling performance of the power capacitor and provide a shunt path from the device to the ground. The shunt path to ground can shorten the length of the via, which can achieve two purposes: you not only create the shunt or ground, but also reduce the transmission distance of devices with small pieces of ground, which is an important RF design factor.