At the heart of the motherboards and graphics cards that we review on this site is the printed circuit board (PCB). On one side, you have the cold, hard facts and data of benchmarks and testing. On the other hand, you have the ingenuity of manufacturing, coupled with the artistry of designers. The PCB is an instantly recognizable symbol of both the beauty of electronics design, and its overwhelming sophistication. The following primer assumes no real knowledge of PCBs, but we've tried to keep it comprehensive, and go from A to B to create a general reference for all our readers.
What Is A PCB?
A PCB is found in almost every electronic device. If you have electronic components in a device, they are mounted on a PCB, big or small. Besides keeping the components in place, its purpose of a PCB is to provide electrical connections between the components mounted on it. As electronic devices have become more complex, and require more components, the PCB has become more populated, and dense with wiring and components.
Typical PCB. The bare board (without components) is also referred to as a 'Printed Wiring Board'.
The substrate of the board itself is an insulating and non-flexible material. The thin wires that are visible on the surface of the board are part of a copper foil that initially covered the whole board. In the manufacturing process this copper foil is partly etched away, and the remaining copper forms a network of thin wires. These wires are referred to as the conductor pattern and provide the electrical connections between the components mounted on the PCB.
What Is A PCB?, Continued
To fasten the components to the PCB their legs are soldered to the conductor pattern. On the most basic PCBs (single-sided boards) the components are located on one side of the board and the conductor pattern on the opposite side. This requires holes in the PCB for the component legs to penetrate the board. Hence, the legs are soldered to the PCB on the opposite side of where the components are mounted. The top and bottom side of a PCB is therefore respectively referred to as the 'Component Side' and 'Solder Side.'
If a component needs to be removable from the PCB after it is manufactured, it is mounted on the board with the use of a Socket. The socket is soldered to the board while the component can be inserted and taken out of the socket without the use of solder. The one shown below is a ZIF (Zero Insertion Force) socket, which allows the component (here a processor) to be inserted easily in place, and be removable. The lever on the side of the socket is used to fasten the component after it is inserted.
To connect a PCB to another PCB an edge connector is often used. The edge connector consists of small uncovered pads of copper located along one side of the PCB. These copper pads are actually part of the conductor pattern on the PCB. The edge connector on one PCB is inserted into a matching connector (often referred to as a Slot) on the other PCB. In a PC, graphic cards, sound cards and other similar products are connected to the main board with the use of edge connectors.
What gives the PCB its green or brown color is the solder mask. This is an insulating and protective coat that protects the thin copper wires and prevents solder from attaching outside the connection points for the components. On top of this colored mask a silk screen is printed. This is text and symbols (often white) printed on the board to label the locations for the different components that are to be mounted. The silk screen is also referred to as the legend.
Green PCB with white silk screen
Brown PCB without silk screen
Types Of PCBs
As mentioned earlier the most basic boards have the components mounted on one side of the board and the conductor pattern on the opposite side. Since there only is a conductor pattern on one side, this type of PCB is called 'Single-sided.' This type of board has severe limitations when it comes to routing the wires in the conductor pattern (since there is only one side no wires can cross, and they have to be routed around each other), it is only used in very primitive circuits.
Single-Sided PCB, top view
Single-Sided PCB, bottom view
These types of boards have a conductor pattern on both sides of the board. Having two separate conductor patterns requires some kind of electrical connection between them. Such electrical 'bridges' are called 'vias'. A via is simply a hole in the PCB that is filled or plated with metal and touches the conductor pattern on both sides. Since the surface available for the conductor pattern is twice as large compared to a single-side board, and that wires now can cross (by routing them on opposite sides of the board), double sided PCBs are much more suited for complex circuits than the single-sided.
Double-Sided PCB, top view
Double -Sided PCB, bottom view
To increase the area available for the wiring even more these boards have one or more conductor pattern inside the board. This is achieved by gluing (laminating) several double-sided boards together with insulating layers in between. The number of layers is referred to as the number of separate conductor patterns. It is usually even and includes the two outer layers. Most main boards have between 4 and 8 layers, but PCBs with almost 100 layers can be made. Large super computers often contain boards with extremely many layers, but since it is becoming more efficient to replace such computers with clusters of ordinary PCs, PCBs with a very high layer count are less and less used. Since the layers in a PCB are laminated together it is often difficult to actually tell how many there are, but if you inspect the side of the board closely you might be able count them.
The vias described in the section about double-sided PCBs always penetrate the whole board. When there are multiple layers of conductor patterns, and you only want to connect some of them, such vias waste space that could be used to route other wires. 'Buried' and 'Blind' vias avoid this problem because they only penetrate as many layers as necessary. Blind vias connect one or more of the inner layers with one of the surface layers without penetrating the whole board. Buried vias only connect inner layers. It is therefore not possible so see such vias by just looking at the surface of the PCB.
In multi-layer PCBs whole layers are almost always dedicated to Ground and Power. We therefore classify the layers as Signal, Power or Ground planes. Sometimes there is more than one of both Power and Ground planes, especially if the different components on the PCB require different supply voltages.
Technologies For Component Packing
Through Hole Technology
The components that are mounted on one side on the board while its legs are soldered on the opposite side are called 'Through Hole' (THT: Through Hole Technology). Such components takes up a large amount of space and require one hole to be drilled in the PCB for every leg. Hence, their legs occupy space on both sides of the board, and the connection points for them are also fairly large. On the other hand, THT components are fairly good mechanically connected to the PCB compared to Surface Mounted devices, which will be discussed below. Connectors for cables and similar devises also have to withstand mechanical stress and are usually THT.
Through Hole Components (soldered on bottom side)
Surface Mounted Technology
The legs of components that are made using 'Surface Mounted Technology' are soldered to the conductor pattern on the same side of the PCB as the component is mounted. This technology does therefore not require a hole in the PCB for every leg of the component.
Surface Mounted Components
Surface Mounted Components could even be mounted on both sides of the PCB directly underneath each other.
Surface Mounted Components mounted on the solder-side of a PCB
SMT components are also much smaller than THT components. This makes PCBs with SMT components much more dense compared to similar PCBs with THT components. Today SMT components are also cheaper than THT components. It is therefore no surprise that most components on main boards nowadays are SMT.
Since the connection points and component legs are so small it becomes very hard to solder on a SMT component manually. Considering that machines do almost all assembly, this issue only becomes important when repairs have to be done.
The design of a PCB is a process that starts long before the actual routing of the conductor pattern. These are the main steps in the design:
A System Specification of the electronic device that is to be made must be formulated. This includes specifying all functions of the system, cost limits, size, operating conditions, etc.
System Block Diagram
A Block Diagram of the system's major functions must be created. How the different blocks are related must also be specified.
Partition The System Into Separate PCBs
Both the reduction in size and the ability upgrade/exchange separate parts of the system are advantages of dividing the system into separate PCBs. The system block diagram gives a good indication of how this should be done. A PC would be divided into main board, graphic card, sound card, floppy drive, power supply, etc.
Determine The Technology To Be Used And The Size Of Each PCBs
When the technology and amount of circuitry on each PCB is determined, the board size must be estimated. If space limitations apply and it turns out that a PCB will be too large, the technology must be changed or the partitioning must be redone. When choosing the technology the quality and speed of the circuit must also be considered.
Schematic Of The Circuitry On All PCBs
A schematic is a detailed drawing of all connections between the components in a circuit. This must be done for all PCBs in the system, and is nowadays done using Computer Aided Design (CAD). Below is an example of a schematic that was made using CircuitMakerTM
Schematic of the PCB
Simulating The Design
To make sure the designed circuit work properly it must be simulated with a computer program. Such programs take the schematic as input, and can than display the operation of the circuit in numerous ways. This is much more efficient than building a prototype on a breadboard and doing the measurements manually.
Placing The Components On The PCBs
How the components are placed on the board is dependent on how they are connected. They must be placed such that the wires that link them together can be routed as efficient as possible. Efficient wiring is as short as possible and uses as few layers as possible (which also keeps the number of vias at a minimum), but we will come back to this under Routing. Below is a picture of busses routed on a PCB. It is important to place the components such that they allow such nice routing of wires.
Wires forming a Bus
Testing for routability and proper functionality when running at high speeds
At this point certain computer programs can actually check if the placement of the components will make it possible to route the wires such that the circuit will work when running at high speeds. This is referred to as proper sequencing or scheduling of the components, but we will not go into detail about this here. If the circuit appears to be malfunctioning, the components could be rearranged before the actual routing is done.
Routing The PCBs
The connections in the schematic are now translated into a model of the actual conductor pattern. This process is usually automated, but manual modifications are often necessary. Below is a screenshot from the routing of a two-layer board. Red and blue lines respectively represent wires on the component and solder side of the PCB. The yellow text and rectangles are part of the silk screen that labels the locations of the components. Holes and vias are showed as dark red dots and circles. At the far right of the circuit we can see an edge-connector at the solder-side of the PCB. This final model of the PCB is often referred to as the Artwork.
Routing the PCB using CAD tools
For every design a set of rules that specify minimum clearance between wires, minimum wire width and similar physical properties of the conductor pattern must be followed. These rules depend on factors like the speed of the circuit, the power of the signals that are to be transmitted, how sensitive the circuit is to leakage currents and noise, and the quality of the materials and the manufacturing equipment. The thickness of a conductor must for example be increased the more power it is required to transmit. To reduce the cost of the PCBs as few layers as possible must be used without breaking any of these rules. If more than two layers are necessary, Ground and Power planes are usually used to avoid routing wires carrying supply voltages on the signal layers, which can cause unwanted leakage currents. They also act as shields for the signal layers.
Test Of The Routed Circuit
To ensure that the circuit works properly after the wires have been routed, it has to pass a final check. This check will also verify that no connections have been incorrectly routed, i.e. that all components have been connected according to the schematic.
Creating Manufacturing Files
Since there are numerous different CAD tools for designing PCBs, the manufacturer needs a standardized set of files as input to the machines that produces the boards. There are a couple of different standards, but the most common is Gerber files. A set of Gerber files includes photo plots for all signal, power and ground layers, photo plots for the solder mask and the silk screen, drill files, and pick-and-place files.
Electromagnetic Compatibility Issues
An electronic device that is designed without considering Electromagnetic Compatibility (EMC) is likely to radiate electromagnetic energy that can cause undesirable interference in nearby electronics. EMC is a design requirement that has maximum limits for Electromagnetic Interference (EMI), Electromagnetic Fields (EMF) and Radio Frequency Interference (RFI). This requirement ensures both the proper operation of the electronic device itself and other nearby devices. It compels the design to limit the radiative or conductive emission from one device to another and reduce the device's susceptibility to external sources of EMF, EMI or RFI. In other words the goal is to prevent stray electromagnetic energy from entering or leaving a device. This is a rather difficult issue to deal with. Common techniques are the use power and ground planes, and place the PCB inside a metal box. Power and Ground planes tend to shield emission to and from the signal layers, while a metal box also shields the components. We will not go more into detail about that here.
The maximum speed of a circuit depends on how well the EMC requirement is met. Internal EMI like leakage currents between conductors increases in magnitude when the frequency of the circuit increases. The distance between connectors must therefore be increased if there is a big potential difference across them. This also tells us that it is important to avoid high voltages and keep the power consumption of the circuit at a minimum. The latency in the wires is also crucial, so their length must be as short as possible. Hence, a small and well-routed PCB tends to be capable of running at higher speeds than a large PCB.
The process of manufacturing a PCB starts with a board of Glass Epoxy or similar substrate. This is referred to as the base substrate.
Imaging (Forming The Conductor Pattern)
The first step is to create the conductor pattern that provides the electrical connection between the components. We will here give an introduction to a 'Subtractive transfer' of the artwork into metal conductors. This technique involves covering the whole base substrate with a thin copper film and then remove the superfluous copper. 'Additive PatternTtransfer' is another less common way of creating the conductors. The copper is then only added where the wires are to be formed. We will not discuss Additive Pattern Transfer in more detail here.
If a double-sided PCB is to be made, the base substrate is covered on both sides with the copper film. For multi-layer PCBs several of these boards are made and laminated together at a later stage.
The following flowchart shows how the conductor pattern is formed on the base substrate.
Imaging (Forming The Conductor Pattern), Continued
The (positive) Photoresist is made of a light sensitive material that dissolves when it is illuminated and developed (negative photoresist dissolves in the development process if it has not been illuminated). There are many ways to apply the Photoresist material onto the copper surface, but the most common is to heat and roll on a film containing the Photoresist material (referred to as Dry Film). It could also be sprayed on as a liquid, but the Dry Film offers higher resolution, resulting in thinner wires.
The mask is just a photo plot of the layer that is to be made. When this mask is placed over the photoresist before it is exposed with UV light, it prevents certain areas of the photoresist to be illuminated (assuming a positive photoresist). The copper covered by the photoresist in these areas is later turned into wires in the conductor pattern.
After the photoresist is developed the copper that is to be etched away is left uncovered. The etching is done by either lowering the board in an etching solution, or spraying the etching solution onto the board. Common chemicals for the etching solution are Ferric Chloride, Alkaline Ammonia, Sulfuric Acid + Hydrogen Peroxide, and Cupric Chloride. When the etching is complete the remaining photoresist is removed. This is referred to as 'Stripping' the photoresist.
The figure below illustrates how the copper wires are formed.
This process is done simultaneously on both sides of the board.
Drilling And Plating
If the manufactured PCB is a multi-layer board that contains buried or blind vias, each layer has to be drilled and plated before they are laminated together. If not, the layers can be laminated together first.
After drilling holes in the boards, which is done by machines according to the drill-files, the inside of the holes must be 'Plated' (Plated-Through-Hole technology: PTH). This metallization of the holes' inner walls creates an electrical connection through the board and to all conductors in the inner layers that the holes touch. Before the plating can start 'Drill Smear' on the inside of the holes must be removed. This is a resinous epoxy coating caused by the heat from the drilling, and must be removed because it covers the conductors in the inner layers. Both Drill Smear removal and Plating are done in chemical processes.
Lamination Of Multi-Layer PCBs
Single layers must be laminated together to form a multi-layer PCB. Lamination involves gluing the layers together with an insulating film in between. For holes that go through all the layers the drilling and plating must be repeated. The conductor pattern of the two surface layers of a multi-layer PCB is often created as described above after all layers have been laminated together.
Solder Mask, Silk Screen And Plating Of Edge Connectors
A Solder Mask is applied over the wires on the outer layers such that solder will not attach outside the solder pads. The Silk Screen is printed on top of this mask to label the component locations. It is important that the silk screen does not cover any solder pads or edge connectors, something that would respectively reduce the solderability and electrical connection. The edge connectors are often plated with gold to ensure a high quality electrical connection when they are inserted in a slot.
Testing the PCB for short circuits and breaks (broken connectors) can be done both optically and electrically. Optical tests involve scanning the layers to detect defects, while electrical tests most often are done by a 'Flying-Probe' that verifies all connections. An electrical test is more reliable when searching for short circuits and breaks, but the optical test can more easily detect incorrect spacing between conductors.
Component Mounting And Soldering
The final step is to mount and solder the components. Both THT and SMT components are placed on the PCB by machines.
THT components are most often soldered in an automated process called 'Wave Soldering'. This enables all components to be soldered simultaneously. Their legs are first cut near the board and slightly bent over to keep the component in place. The PCB is then moved over a wave of liquid flux, such that the bottom side strikes the flux. This removes any oxide from the metal surfaces. After heating the PCB it is similarly moved over a wave of melted solder. The solder attaches to the solder pads and component legs, and the soldering is complete.
A common way of soldering SMT components automatically is 'Over Reflow Soldering.' A solder paste containing both flux and solder is then applied to the solder pads before the components are placed on the PCB. The PCB is then heated in an oven such that the solder in the paste melts. Cooling the PCB completes this type of soldering and the PCB is ready for final testing.
Where The Cost Savings Come In
To make the cost of the PCB as low as possible a lot of factors have to be considered:
- The size of the board is of course significant. The smaller the board is, the cheaper it is. Some PCB sizes have become standard sizes for manufacturing, and sticking to one of these sizes helps reduce the cost. The website of CustomPCB has info about some standard sizes.
- Using SMT is cheaper than THT because it makes the PCBs more dense (and therefore smaller).
- On the other hand, if the board becomes very dense the wires in the conductor pattern must be thinner, and more high tech machines have to be used to manufacture the boards. Higher quality materials must also be used, and the routing of the wires must be done more carefully to avoid any leakage currents that could affect the operation of the circuit. All this could increase the cost of the PCB more than what is gained by reducing its size.
- The cost increases with the number of layers, but fewer layers will often increase the size of the PCB.
- It takes time to drill the holes, so as few vias as possible is desirable.
- Buried vias are more expensive than vias that go through all the layers. This is because buried vias makes it necessary to drill each layer separately before they are laminated together.
- The size of the holes in the PCB depends on the diameter of the component legs. If components with different types of legs are required on the same board the machine that drills the holes cannot use one single drill to drill all the holes. The more times the drill has to be changed while processing one board, the more expensive the PCB is to manufacture.
- An electric test with a 'Flying-Probe' is more expensive than an optical test. Often an optical test is sufficient to make sure that the PCB does not have any defects.
It all adds up to a significant amount of work for the manufacturer as devices become more complex. The process of creating a PCB is useful to understand because, it gives us an indication of the abilities of a manufacturer when we are comparing like boards, which may deliver similar performance, but may vary in cost, or reliability.
A good engineer can look at a board and draw a conclusion as to the quality of the design. You may not want to go that far, but the next time you look at a motherboard, or graphics card, you just might have an appreciation of the art of PCB design.