Posts Tagged ‘PCB Design’

08_74156

Manufacturing Process of Printed Circuit Boards

Written by Rush PCB Inc on . Posted in PCB, PCB Manufacturing

For anyone involved in the electronics industry, the process followed for manufacturing Printed Circuit Boards (PCBs) is very important. This is because PCBs are the very basis of all electronic circuits, being used to provide the mechanical basis on which circuits are built. PCBs come in different forms—rigid, flexible, rigid flex, and High Density Interconnect (HDI), the difference being mainly in the materials used to fabricate them, and give them their ability to flex or remain rigid. To some extent, the PCB manufacturing process depends on the material being used, but the main stages in the PCB manufacturing process remain the same, irrespective of the nature of the PCB. However, before coming to the manufacturing stage, the designer has to make a few choices depending on the application. These are:

  • Selecting the type of PCB required by the application
  • Deciding whether the board will have single, double, or multiple layers
  • Deciding the mechanical layout, the stackup, and the routing of tracks on different layers
  • Producing relevant documents and files for the manufacturing process.

The above steps are important for the success of the final product, and its reliability when used in the application. For instance, if the application demands a component moves back and forth during operation, such as the head of a printer does, a flexible circuit must supply it. Most wearables are shrinking in size, and HDI technology is most suitable for the rigid-flex boards they use. At this stage, the designer selects the material for the PCB design. The complexity of the electrical circuit that goes into the design decides the number of layers on the PCB. With the area available on gadgets shrinking as they tend towards miniaturization, the density of PCBs also increases proportionately. The only option for the designer is to have multiple layers on the PCB to contain the design within the specified mechanical boundary. Depending on the nature of the application, the designer has to decide on the stackup, or the design of consecutive layers. For instance, if the application has high frequency circuits, the designer has to define the impedance and minimize crosstalk. For this, he may have to use power and ground layers alternately, with traces carrying signals in between. While doing this, the designer has also to decide on the width, spacing, and routing of traces, placing of vias and test pads, and more. Once he/she completes the design process, the designer produces an output in the form of standard documentation, which helps the manufacturer fabricate the specified design. This documentation can follow either the Gerber X2 format or the IPC-2581 format.

Selecting Suitable Material for Printed Boards

Material for Substrate

The most widely used material for rigid boards is based on glass fiber known as FR-4. Being reasonably priced, FR-4 also has an appropriate degree of stability under temperature variations, and does not break down easily. Other cheaper materials such as paper phenolic are also available and used for low cost commercial products. On the other end of the spectrum are substances such as Teflon or PTFE used for substrates, offering very low losses and a stable dielectric constant for high performance high frequency designs. Flexible, rigid-flex, and HDI PCBs typically use Polyimide based substrates, although Teflon and Kevlar are also used for high frequency and high performance applications.

Material for Cladding

A PCB requires copper traces and planes, and these originate in the form of copper cladding on the substrate—a thin sheet of copper bonded to the substrate. The designer can specify the thickness of the copper cladding, as the manufacturers make them in a few standard thicknesses. Selecting the copper on a circuit board is also critical to reaching and maintaining desired performance levels, especially for adhesion to the substrate and performance at high frequencies. Commercially, manufacturers use two types of copper foils for PCBs—Electrodeposited (ED) copper and Rolled-Annealed (RA) copper. The two types of copper foils have different forming processes, and undergo different treatments for improving and preserving adhesion to various substrate materials. For instance, formation of ED copper follows electrical deposition of copper on a slowly rotating polished stainless-steel drum, placed in a solution of copper sulfate. The copper foil is removed in a continuous roll, with the side against the drum providing the smoother finish. On the other hand, ingots of solid copper successively passing through a rolling mill produce RA copper foils. The two types of copper possess different qualities. ED copper used on PCB substrates are useful for applications where mechanical stress may be critical, while RA copper is suitable for applications involving thermal shock. When subjected to thermal cycling conditions, PCBs with ED copper may develop cracks in narrow conductors. The grain structure of PCB copper varies according to its manufacturing process. For instance, manufacturers offer ED copper foils in coarse, moderate, and fine grain structures, and is linked to the finish on the copper surface. Although a copper foil with a coarse structure also has a coarse surface finish, and enables a stronger bond between the copper foil and the dielectric laminate, it typically exhibits a greater insertion-loss performance at higher frequencies. Additionally, skin effect at higher frequencies worsens the insertion-loss performance of the rough surface of copper traces on a PCB.

Multilayer PCB Fabrication Process

The actual process for PCB fabrication can begin on receipt of the necessary documentation from the designer regarding the proper choice of materials for the substrate and cladding, the number of layers and stackup, the mechanical layout and routing. The documentation must have individual details for each layer of the PCB.

Preparing the Central Panel

The fabrication process starts with obtaining the copper clad board, with the specified substrate material and copper cladding. For a multilayer board, the cladding will be on both sides of the substrate, and this forms the innermost or central layer. Usually, such copper clad boards come in large sizes of standard dimensions, and the necessary panel of a size matching the mechanical layout specified has to be cut out by shearing the board. Depending on the individual size and total number of PCB units to be made, the panel may have to be dimensioned to hold multiple PCB units. The copper cladding usually comes with a thin coating of protective layer to protect the surface from oxidation, and the protective layer must be removed by immersing the panel in a bath containing the solution of a weak acid.

Drilling and Etching the Central Panel

The panel is dried and is usually heated to remove excess moisture. Using details from the drilling files submitted by the designer, the fabricator proceeds to drill the necessary holes for the central layer. This starts with drilling the registration or locating holes on the periphery of the panel, and the CNC machine changes its drills to match the diameter for individual holes specified in the drill file. The registration holes are necessary for aligning subsequent layers. To obtain the correct pattern of tracks on the two sides of the panel, fabricators use a combination of a photographic process followed by a process of chemical etching. Typically, the copper surfaces on the drilled panel are covered with a thin layer of photoresist. Each side is then exposed to UV light through a photographic film or photo-mask detailing the optically negative pattern of tracks specified by the designer for that layer. UV light falling on the photoresist bonds the chemical to the copper surface, and the rest of the unexposed chemical is removed in a developing bath. This stage is usually supplemented with a visual inspection. When placed in an etching bath, the etchant removes the exposed copper from the panel. This leaves behind the copper traces hidden under the photoresist layer. During the etching process, concentration of the etchant and time of exposure are both critical parameters to obtain optimum results. Stronger than necessary concentration and a longer time of exposure can result in over-etching copper from under the photoresist, leading to tracks with widths thinner than specified by the designer. After successful etching, the photoresist is washed away to leave the necessary copper tracks on both sides of the central layer. Although the photographic process is very popular, other methods are also available. For low volumes or prototype PCBs, an etch-resistant ink may be transferred onto the copper surface using a silk screening process, with the silk screen representing the optically negative pattern of the tracks required. Another dry process uses a specialized, highly accurate milling machine to remove the unwanted copper from the surface of the panel. The machine uses inputs from the documents supplied by the designer to drive the automated milling head. However, as the process is time-consuming, it is suitable only for very small quantities of PCB.

Plated Through Holes

Holes in a PCB are necessary for connecting traces on one side to traces on the other. Sometimes, these holes may also be required to hold the leads of a leaded component, although this is becoming a rare requirement due to the availability of SMD components. Additionally, the PCB may require some holes to enable it to be mechanically mounted. To facilitate interconnection between the layers, PCB fabricators line the inner surface of the holes with a copper layer, using a plating process. On completion, these are called plated-through holes or PTH. A round of visual inspection and electrical continuity testing at this stage verifies the process. For single or double-layer type PCBs, the panel now goes for solder masking (coverlay for flexible PCBs) to cover those parts of the tracks that will not be soldered, and surface finish for the exposed tracks and pads. Multilayer boards will have further layers added on to the central or innermost layer.

Adding on Subsequent Layers

Fabricators add subsequent copper layers to the central layer with a layer of insulation in between each. For rigid PCBs, this insulation is usually the prepreg, while for flexible PCBs this is an adhesive layer. Fabricators add insulation and copper layers on to each side of the central layer, using heat and pressure to bond them together. The copper surfaces on both sides now undergo the same treatment of protective layer removal followed by drilling. Only this time, the drilling depth is controlled so that the copper on the inner layers remains undamaged. Eminent PCB manufacturers such as Rush PCB Inc. use ultrasonic or laser drills in place of mechanical drills for achieving greater accuracy and reliability. The same process of photoresist and etching follows as above, and this leaves only the traces required by the designer for the specific layer. The drilled holes are then electroplated to provide the necessary connections. Usually, a round of visual inspection and electrical testing to verify the process follows. If no further layers are required, the panel now goes for the solder masking/coverlay process. For additional layers, the above process is repeated.

Solder Mask/Coverlay and Surface Finish

Fabricators protect areas of the PCB not to be soldered by covering them with a protective layer. For rigid PCBs, this is the usual green layer of solder mask. However, this is a brittle layer, and not suitable for flexible PCBs. Therefore, flexible and rigid-flex PCBs use a polymer adhesive layer called coverlay for the purpose. The solder mask/coverlay protects the board from other contaminants as well, as the PCB goes through the assembly process. To enable leaded or SMD components to be added to the board by soldering, the solder mask/coverlay has openings at appropriate places, exposing the copper surface. To prevent the exposed copper from oxidizing, fabricators tin or plate them with solder, plate them with gold, or use other combinations of different metals to achieve a surface finish, as specified by the designer.

Silk Screen

The last step in the PCB manufacturing process consists of printing text and other idents on the PCB. Usually, this helps in identifying the board, marking component locations, and fault finding instructions. After a final inspection, the PCB is ready for dispatch.

095015276

Problems with the Gerber File Format and Solutions

Written by Rush PCB Inc on . Posted in PCB Assembly and component, PCB Design, PCB Manufacturing, Uncategorized

The world over, a majority of designers and fabricators follow the Gerber RS-274X as the de facto standard when designing and fabricating their PCBs. The evidence of its popularity notwithstanding, Gerber has a number of practical limitations. Often, these limitations lead to a variety of problems when fabricating PCBs.

Brief History of the Gerber File Format

Ucamco developed the Gerber file format in the 1960s, when it was the Gerber Systems Corporation, and a leading provider of early photo-plotter systems using numerical controls. Their first format, RS-274D, was a subset of EIA RS-274-D, supporting their vector-based photo-plotters. Widely adopted, RS-247D remained the standard format for vector-based photo-plotters until the 1980s.

Raster scan plotters began replacing vector-based photo-plotters in the 1980s. These newer plotters were bitmap-based, requiring a completely different data format. Consequently, in 1998, Barco ETS, who had acquired Gerber Systems, released a single standard image format, and named it the Extended Gerber or GerberX. This was later renamed as the RS-274X format and is still in use today.

The latest Gerber RS-274X presents a complete image description format. Therefore, the Extended Gerber file holds the complete description of a layer of the PCB, and provides the operator with everything necessary to generate a PCB image, including the definition of any aperture shape. Requiring no external aperture files, painting, or vector-fill, the RS-274X standard specifies all pads and planes clearly and simply. Its simplicity has made it the de facto standard followed by nearly 90% of the world’s PCB designers and fabricators.

Problems with the Gerber File Format

Despite its wide acceptance and use, the RS-274X Gerber file format has its own shortcomings. The trouble is the standard does not address all aspects of fabrication and assembly, as required by the PCB fabricator.

Although Gerber RS-274X is extremely accurate and reliable when rendering images of copper shapes precisely on signal and plane layers, it does not transfer the layer stackup order accurately. Moreover, data sets and information regarding materials, drill data, netlist, pick-and-place data, bill of materials, test point reports, and more need to be generated by separate processes by different utilities. This means, the Gerber RS-274X format is incapable of transferring the complete information from the design domain to the manufacturing domain.

In the absence of a defined layer order being transferred to the manufacturer, fabrication has no way of deciding the order of copper layers and may miss a layer or two altogether. With the layer order missing, the drill data may generate holes relative to an incorrect layout. This mismatch can happen with the entire assembly data, and at all aspects of the fabrication process. Usually, with Gerber RS-274X, there is no defined way a fabricator can know about missing output data, wrong source file version, and these can render boards useless.

Designers usually get over the above shortcomings using a well-maintained design methodology and best practices shared with the fabricators. In general, they utilize Gerber RS-274X with minimum fabrication issues. However, maintaining ideal conditions all the time is difficult, things can slip up, causing problems to the fabricator and assembly houses, and now they have to face the brunt of the responsibility and sort through the issue. This also leads to fabricators and assembly houses being forced to spend a great deal of time and resource in inspecting and verifying the entire data for all incoming jobs, simply to minimize manufacturing issues.

Solutions and Alternatives to the Gerber File Format

Eminent PCB manufacturers such as Rush PCB Inc., eliminate the problems by adopting design transfer standards that addresses all aspects of the fabrication and assembly process. Two new open standards are available, and these enable efficient and accurate data exchange from the PCB designer to the manufacturing fabricators and assemblers. Ucamco administers one of these standards, the Gerber X2, while the IPC Consortium administers the other, the IPC-2581. Both are open standards, free from any proprietary restrictions.

The Gerber X2 File Format

The Gerber X2 is an expanded version of the GerberX format. In addition to the layout image data, Gerber X2 now includes design data as well. The X2 fabrication files now include the board layer order and stackup information that so far, fabricators had to interpret and verify manually. In the same way, a set of drill files is also included within the X2 fabrication files, detailing the location, drill size, plated/non-plated information, and the layer span.

The X2 attribute system qualifies objects with specifications such as file function, part, pad function, and more that add intelligence to the traditional image data improving the automation process. For instance, the file function specifies a file as top copper layer, top solder mask, while part specifies whether the PCB is a single or a panelized array, and pad function defines the pad as belonging to a via, through-hole, SMT, or fiducial. The Gerber X2 format directs all outputs to one single folder.

As the Gerber X2 is both forward and backward compatible with the RS-274X standard, it helps any X2 reader to interpret Gerber RS-274X files correctly. Therefore, fabricators using the Gerber X2 process will have no trouble interpreting legacy fabrication files created in the Gerber RS-274X format and vice versa.

The IPC-2581 File Format

Contributors from a wide range of PCB industry segment initiated, developed, and drove the IPC-2581 standard. These industries included MES, CAD/CAM and PLM vendors, PCB fabricators, contract manufacturers, as well as OEMs.  The IPC-2581 is a single data format and within a single file, contains all aspects of the PCB design, such as layer stackup, materials, assembly, and test details.

With the IPC-2581 standard, the designer can include details of layer stack and information on materials to ensure proper layer order. The standard is suitable for stackups of complex board design such as related to rigid-flex boards, and is capable of handling special materials. It can also include drill and mill data for blind, buried, and filled via types. It also supports information on back drilling, V-grooves, slots, and cavities. For bare board testing, designers can include the net-list as well.

In addition to a complete set of fabrication data, the IPC-2581 can also hold assembly data. Therefore, it can contain not only the pick-and-place information, but also the information on polarity and rotation of a component, enabling support for both stacked and embedded components.

In addition to assembly drawings, the IPC-2581 standard has the capability to generate the documentation for bill of materials and purchasing. Therefore, the standard can tie up with PLM/ERP system data to create links between design and supply chain facilities. The greatest advantage of the IPC-2581 is one single file containing the entire data related to fabrication and assembly.

How does Conductor Surface Roughness matter?

Written by Rush PCB Inc on . Posted in PCB, PCB Manufacturing

In RF circuit design, it is necessary to select the proper Printed Circuit Board (PCB) material for the application. In this context, modern simulation tools such as Computer Aided Engineering (CAE) help engineers by predicting the electrical behavior that circuits exhibit on various types of PCB materials. The tools use material parameters in their calculations, with dielectric constant (relative) being one of the important parameters. However, most such tools overlook an equally important material parameter of PCBs during the process of design—the roughness of the conductor’s surface. Contrary to popular belief, conductor surface is never perfectly smooth, and this has consequences in high-frequency PCB design.

Conductor Surface Roughness

Eminent PCB manufacturers such as Rush PCB Inc. understand the influence an imperfect conductor surface has on the performance of a PCB. In fact, scientists have since long studied the effect of grooves present on the conductor’s surface on a PCB, having noted the increase in losses caused by the grooves. Under worst-case scenarios, the grooves caused losses that were twice the original. The explanation the researchers extended was electromagnetic (EM) waves travel mostly along the conductor’s surface, such as along the copper signal trace on a PCB. The grooves actually cause the signal paths to become longer, as the EM waves, while traveling along the surface, have to enter into, and then exit out of the grooves.

Skin Effect

The surface roughness of the conductor thereby causes the EM waves taking a path with a longer mean resulting in the increase in losses. Effectively, higher the degree of surface roughness of the conductor, higher is the resistance from skin effects. When an EM wave propagates in a conductor, skin effect tends to change the EM wave’s current distribution to accumulate more towards the conductor’s surface rather than remaining deep inside the conductive material.

When using EM simulators or other commercial CAE tools, designers overcome the surface roughness of a conductor by relying on the traditional Morgan Correlation. They do this by involving a numerical factor for correcting the surface roughness, depending on the ratio of a smooth surface to a rough one. While calculating the loss of high-frequency microstrip lines, using Kr mostly does a good job of matching closely the results of measurement of conductor losses. However, there are cases where the measurements fail to match the computer predictions so closely.

Furthermore, such deviations between the calculated and the measured values can be expensive at the design stage, especially as achieving the desired performance requirements can lead to additional design iterations. Avoiding such delays in design might mean considering carefully the choice of an RF PCB laminate based on its conductor’s surface roughness.

Types of Copper Cladding

Manufacturers have to use some form of a cladding of copper conductor on the PCB substrate. Three types are most common—electrodeposited (ED) copper, rolled-annealed (RA) copper, and reverse-treated (RT) copper.

Forming RA copper foils involves rolling the copper ingot through a rolling mill, where subsequent passes through the rollers of the mill results in a copper foil consistently thin.

ED copper formation requires depositing copper onto a slowly rotating, highly polished drum made of stainless steel, within a bath containing a solution of copper sulfate. While the roughness of the copper surface where it meets the stainless-steel drum is analogous to that of RA copper, the copper surface of the deposition side facing the solution is much rougher.

RT foil production starts by plating the ED copper foil on the drum side, when the foil on the bath side is still low profile.

As the copper foil has to adhere to the dielectric material, which may range from FR-4 to polytetrafluoroethylene (PTFE) substrates, the copper surface has to be treated to increase its adhesion. The reason being a reasonably smooth copper surface does not adhere ideally to the dielectric. Whether formed by the RA or ED processes, an untreated copper film has a surface typically covered with tiny teeth-like imperfections, and the jagged surface is perfect for forming a powerful bond between the dielectric material and the copper.

However, this is in direct contrast with the requirements of a good transmission line, as the rough surface is then not suitable for transmission of high-frequency EM waves. On the other hand, a surface with a mirror-like finish on an utterly smooth copper foil is inadequate for foil-to-dielectric adhesion.

That means fabricating PCBs with low-loss copper traces while keeping good adhesion between the dielectric material and copper depends on accepting a compromise in the surface roughness of the copper foil.

Effect on Dielectric Constant

Another important factor involving the design and manufacture of PCBs is the relative permittivity of the dielectric material—commonly referred to as its dielectric constant Dk. In reality, Dk, rather than being a constant, varies with frequency.

The value of Dk, as the dielectric manufacturer’s data sheets report, is often assumed as the intrinsic property of the material. However, manufacturers generate the effective dielectric constant using a specific test method, sandwiching the dielectric material between two copper plates. When comparing simulation against measurements, this often causes a discrepancy in insertion loss—caused by increased phase delay resulting from surface roughness.

The explanation for the above is that surface roughness decreases the effective separation between the parallel plates, thereby increasing the electric field strength leading to an increase in capacitance, and that accounts for the increase in effective dielectric constant.

Laminate suppliers commonly use a method called the clamped stripline resonator test method, described by IPC-TM-650, to measure the effective dielectric constant of their material. As the measurement is highly dependent on the test apparatus and the measuring conditions, it does not guarantee the values are accurate for design applications. This is mainly due to reason that the copper foils used for the test are not physically bonded to the laminate, leaving small air gaps in between the layers. This affects the measurement results.

Designers have to get around this mismatch during simulations by using a multiplication factor for the dielectric constant for their impedance calculations, rather than using the Dk directly, as published in the data sheets.

Commercial PCB Laminates

Recognizing the effect of surface roughness on PCB performance at high frequencies, suppliers offer commercial laminates and copper foils in numerous profiles. They produce these laminates with copper treatment at different levels. For instance, they offer PCB materials with low profile (LP) copper conductors that provide excellent adhesion between the dielectric material and copper, while the smooth conductor surface improves etch definition and reduces conductor losses.

Other suppliers offer materials with reverse-treated copper foils of low profile, which are suitable for high-frequency analog and digital circuits. They come in a variety of panel sizes and dielectric thicknesses, with 1- or 0.5-oz. cladding of reverse-treated ED copper in low profile. Two popular models of laminates have dielectric constants of 3.38 & 3.48, while their dissipation factors at 10 GHz in the z-direction are 0.0027 & 0.0037. Both the materials are suitable for high-density circuits and are appropriate for low passive intermodulation distortion, low insertion loss, and superior signal integrity.

Conclusion

Although special material can help overcome the effects of surface roughness of conductors at high frequencies, selecting a PCB material for minimizing the effects of surface roughness is not a simple task. When targeting to minimize the effects of surface roughness, PCB materials copper foils of lower profile will perform better at higher frequencies showing low conductor losses, rather than with materials using foils of higher profiles.