Design for eXcellence

Unit 3: Design for assembly

Section 2: Applying DfA to the PCA process

Photo 1: An example of a PCA with BGA technology and an array of connectors
Source: Agilent Technology

An example of a PCA with BGA technology and an array of connectors

Section Contents


Methods of assessing assemblability of electronic assemblies

Boothroyd-Dewhurst PCA DfA

Printed Circuit Board (PCB) assembly primarily involves the onsertion and insertion plus soldering of components onto PCBs. Component placement and insertion is achieved manually or by high-speed dedicated machines.

For high production volumes, most manufacturers use a combination of automatic and manual processes due to odd shaped components or components that cannot withstand the temperatures during the mass reflow processes. However, it is desirable to use automatic machines wherever possible as they operate at higher production rates and with greater reliability.

Boothroyd-Dewhurst provides a method for estimating PCB assembly costs. In addition, because PCB assemblies contain mechanical parts, it may be used in conjunction with the previous manual assembly assessment methods.

Estimation of PCB assembly costs

Refer to Figure 1 for the example of a logic board analysed using Boothroyd-Dewhurst’s method. This method is very comprehensive because it involves estimation of assembly operation costs and rework costs. A full explanation (and this example) can be found in Product Design for Assembly by Geoffrey Boothroyd and Peter Dewhurst (1991). The PCB DfA method uses a set of comparison metrics:

Figure 1: An example of PCA DfA with Boothroyd Dewhurst

An example of PCA DfA with Boothroyd Dewhurst

Design Report Cards

In his paper, Predicting Cost Tradeoffs in Design for Manufacturing (DFM), Happy Holden describes a system used in Hewlett-Packard called the ‘Assembly Report Card’. Developed in 1992, this is a matrix supplied by the assembler that relates the process assembly and test choices for various component sizes, orientations, complexities and standards of quality to the cost of providing for these design choices.

Typical factors that affect assembly costs are:

By collecting all the costs associated with assembly test and repair and then normalising these costs with the smallest non-zero value a matrix such as shown in Table 1 can be produced.

Table 1: A sample of the point allocation for H-P design report card
factor pts highest pts middle pts lowest
solder process
35
1 pass reflow
20
2 pass reflow
0
reflow & wave
placement
8
100% auto
5
99–90% auto
0
> 90% auto
digital test coverage
9
< 98% auto
3
98%–90%
0
> 90%
manual attachment
8
100% auto
25
brackets
0
post solder assembly

Also in the November 2000 issue of SMT Magazine, Robert Rowland of RadiSys Corp. suggests a DfM Rating Index and describes a mathematical tool used to measure the manufacturability of PCAs. RadiSys has developed a model based on four topics:

The rating index calculator was created in MS Excel. As in Hewlett Packard’s model, in its scoring scheme high points equates to low manufacturability, with scores below 2000 being desirable.

Agilent scorecard

A similar system is currently used by Agilent Technologies. Table 2 shows part of a DfX Scorecard. It has 4 parts P1–P4:

Each issue is rated between 1 and 9, 9 being the most desirable. For example, in the process checklist 100% machine placement scores a 9 and an assembly with more than 20 hand placed components scores 1. Each set has between 5 and 7 guidelines.

The Agilent DfX engineer averages each set and takes the product of the four averages of checklists for a DfX rating. If the rating exceeds a certain level, the DfX engineer returns the design to the designer with recommendations for redesign.

Table 2: Part of the Agilent scorecard
Part of the Agilent scorecard

[back to top]


Principles of DfA as applied to the PCA process

At the start of this unit, some generic DfA principles were listed. Here they are again:

  1. Reduce the number of components in the product

  2. Optimise the ‘assemblability’ of the remaining parts

  3. Optimise the ‘handlability’ of the remaining parts

  4. Improve quality, increase efficiency and reduce assembly costs

Component reduction

Minimising the number of part numbers and component placements drives down material costs, assembly costs, increases yield and results in a superior design. It is difficult in PCB design to reduce the number of components without affecting functionality of the product. But a couple of options are available:

The fact that a generic single board computer is made up of 5% integrated circuits, 4% connectors, 40% capacitors, 33% resistors and 18% miscellaneous parts, gives an idea of the potential for integrating passive components within or on the surface of a substrate. The downside however, is that, even if your regular board fabricator has the facility to integrate components into the board, then the board will be more expensive and have longer lead times. Mounting bare ICs directly onto the board is also an option that can impact on assembly cost, as well as giving smaller, cheaper, and more powerful designs compared to conventional THT and SMT assembly techniques. Both these are topics to which we will return in the Technology awareness module.

Optimising assemblability and handlability

Each component has to be assembled onto the board, and the method of assembling a component and its subsequent handling depends both on the type of component and its location on the board.

Type of component

The type of component will partly dictate its assembly process. Generally, THT components will be wave-soldered and SMT components will be reflow soldered. To be most efficiently built, assemblies will generally nowadays be designed for reflow soldering.

Self Assessment Questions

Since its introduction into widespread use in the 1980s, SMT has become the preferred component technology. Can you suggest reasons why?

Compare your list with the description we give in the next section.

Why use SMT?

In some applications, SMT is superior to through-hole in a number of ways: size, weight, better efficiency in both price and speed, and component availability. The evolution of electronics components has followed the demand of the consumers, increasing the power of the components through higher I/O pin counts, whilst also decreasing their size.

If the circuit boards in modern devices like cellular phones and laptop computers were made using through-hole technology, the extra bulk of the leads and the extra circuit board space required would make those devices impossibly huge and heavy. Instead, the reduction in the sizes of those leads and boards, along with the new soldering method, allow higher I/O pin count, more power, and reduced size, all at the same time.

The small size of the new components leads to a much more efficient use of board space. The area of the SMT boards is greatly reduced, causing a corresponding cost reduction: unless there is a major increase in complexity as a result of the technology change, board costs are closely correlated to board area. By reducing the size of the boards, customers can reduce the size of the enclosures that most of these will go into, saving money also on the material costs of the enclosures. Reduction in size increases the efficiency in the speed of the PCB. Because of the reduction in size of the components, there is a reduction of the distance the signals have to travel, enhancing the operating speed of critical, high-speed circuits.

All of the advantages of SMT are resulting in the slow phasing out of THT components. If customers want to design with the latest component, the latest technology, it’s most likely going to be available from the component manufacturer only in a SMT package. While the strength, durability and size of THT still has its applications, the efficiencies found with surface mount technology, that are not possible with through-hole, make SMT the preferred technology for many applications today.

SMT advantages

Much smaller components can result in a reduced PCB size, with associated cost savings. A typical SMT conversion can reduce component space on the board to 30% of the through-hole size.

SMT disadvantages

Use of through-hole components

The use of THT components should be minimised in any ‘mixed technology’ applications, that is, those assemblies that use both THT and SMT technology. Unless the assembly has sufficient THT components to justify auto-insertion (for example 10 or more), placement of THT components is a manual process that involves significantly more labour than SMT components. Also THT components typically require some degree of preparation for use (lead forming, trimming, etc.), and this adds to overall assembly cost.

Pin in Hole Reflow

Today, typical components such as resistors, ICs, capacitors, and connectors with straight terminal pins are almost exclusively fitted using SMT in mass production. In contrast, angled SMT connectors at the edge of the board have not been successful because of tolerance problems (co-planarity) and stresses during mating. Modified solder connectors for assembly with Pin-in-Hole Intrusive Reflow (or intrusive reflow) process offer a better solution. These can be mounted at low cost, utilising existing SMT production lines.

Pin in Hole Reflow soldering is a technique that eliminates the need for the wave solder process. In this process, the connector is inserted into plated through holes in a comparable way to conventional component mounting. All other components can be assembled on the PCB surface.

The components are positioned using pick-and-place machines. These automatic assembly machines differ according to whether the components are small, lightweight or bulky. Connectors are considered bulky (odd form) because of their comparatively heavy weight and large volume that makes them more difficult to grip. Furthermore, machines for odd form components must have higher insertion power to fit the components into PCB holes, which are filled with solder paste. As a rule, modern SMT production lines are equipped with both types of machines; therefore the Pin in Hole Intrusive Reflow process generally entails no extra investment costs for the user.

Odd-form components

An odd-form component is any device that can’t be placed automatically by an SMT placement machine. Prime examples include transformers, connectors, DIPs, SIMM and DIN sockets, along with conventional radial and axial TH devices.

Odd-form placement also tends to be one of the few remaining manual production areas for many manufacturers. This makes it prone to all the quality and throughput problems that stem from hand assembly.

This is why growing numbers of European and American firms, for example high volume mobile phone manufacturers, have installed odd-form placement machines and final assembly systems onto their production lines.

Odd-form placement machines range from dedicated radial or axial THT inserters to highly flexible placement machines capable of handling almost any shape and size of component package. But to cater for a wide range of diverse assembly needs, there are a range of different placement techniques and machine features offered.

A typical example is PMJ’s HiSAC odd-form placement machine. This deploys a flexible servo gripper placement head that uses different pick-up tool attachments to place a wide variety of odd-form components supplied in almost any feeder format. The placement head automatically selects the tool it needs to pick up a component, while sensors and a vision inspection camera are used to verify that the right device is placed correctly. In addition the HiSAC machine can cut and form components prior to placement and automatically clinch THT component legs after insertion to keep components safely in place during the assembly process.

[back to top]


Minimise the number of process steps and technologies

Each process technology and process step adds labour, tooling and non-reoccurring engineering costs and increases opportunities for defects. Reducing the overall number of process technologies used in the design/assembly of a product dramatically reduces the cost of assembly and can enhance product reliability with fewer major thermal cycles. Elimination of an entire process technology step will increase ease of assembly and decrease assembly cost more than any other DfA effort.

For example:

Each of the following are considered process steps and technology listed from least to most labour-intensive:

Assembler’s capabilities

The designer must ensure the component technology levels comply with the assembler’s process capabilities. The assembler will have minimum process capabilities depending on the equipment they have. For example, the developer may have older equipment that may need upgrading or adjustment for smaller discrete components like 0402 or 0201 body size. The assembler will be able to advise designers of their capabilities through either encouraging visits and discussions or issuing guidelines.

Note that, even if the assembler has the capability, it is not advisable to use the smallest components or latest technology unless absolutely necessary. Assembly yields must be considered when designing a product.

[back to top]


Component location

There are design rules and guidelines that, if used, will produce designs with acceptable assembly yields and quality. Component location guidelines form a large part of these, especially location in relation to other components and features. The second major consideration is whether the component is placed on the top or bottom of the PCB. Along with component type, this will dictate the process used, as you will have seen in the last section.

Component spacing

The closest distance a designer will separate the components to be assembled is determined by the process flow of the product. During assembly the product passes through a number of processes, each of which constrain the component spacing. Assemblers can offer spacing matrices that detail recommended component spacing, depending on the assembly processes.

An example of an assembler’s matrix for designer reference is shown in Table 3 This is a look-up table that categorises components into package types and sizes and allows the user to cross-refer different types to one another. For example, in this table, the minimum spacing recommended between an 0603 discrete capacitor and a 0805 resistor is 0.95 mm for wave soldering and 0.49 mm for reflow soldering.

Table 3: Component spacing matrix for various components for both wave solder and reflow process

component spacing (mm)

0402 cap. 0402 res. 0603 cap. 0603 res. 0805 cap. 0805 res. ≥1206 cap. ≥1206 res. electrolytic
0402 capacitor
0.70
0.35
0.70
0.35
0.95
0.49
0.70
0.35
1.20
0.65
0.75
0.38
1.45
0.78
0.81
0.40
5.75
5.50
0402 resistor
0.50
0.23
0.95
0.49
0.70
0.35
1.20
0.65
0.75
0.38
1.45
0.78
0.81
0.40
5.75
5.50
0603 capacitor
0.95
0.49
0.95
0.49
1.20
0.65
0.95
0.49
1.45
0.78
0.95
0.50
5.75
5.50
0603 resistor
0.70
0.35
1.20
0.65
0.75
0.38
1.45
0.78
0.81
0.40
5.75
5.50
0805 capacitor
1.20
0.65
1.20
0.64
1.45
0.78
1.20
0.60
5.75
5.50
0805 resistor
0.75
0.38
1.45
0.78
0.81
0.40
5.75
5.50
≥1206 capacitor
1.45
0.78
1.45
0.80
5.75
5.50
≥1206 resistor
0.81
0.40
5.75
5.50
electrolytic
3.27
1.52
Component-to-component spacing shown is wide edge to wide edge Figures shown in bold on the top half of each cell are for wave soldering;

those for reflow soldering in the bottom half of the cell in normal type.

A number of process issues can affect the spacing recommendations in the table. The recommended spacing will be the largest of the following:

  1. In general, line-of-sight must be maintained for manual or automated inspection of the finished assembly. If the solder joints cannot be easily inspected and reworked the spacing is inadequate.

  2. In pick and place the nozzle will pick up the component using suction by the top surface. Each part is placed sequentially and therefore there will be no fouling of one component on another, providing the components are placed accurately.

  3. At reflow, the solder paste is heated until it liquefies and forms a joint between the component pad and its pin. The solder should reflow within the land area, and solder mask between consecutive pads will prevent solder bridges unless too much solder was deposited onto the pads. Otherwise the component spacing may be very close and other factors will determine spacing requirements.

  4. At wave soldering there are two effects:

    • Shadowing, where a tall component will ‘shadow’ a smaller component, preventing it from touching the solder wave. In this case greater separation spacing must be designed in to allow for this.

    • Short-circuits due to many reasons, for example, the pins of a component being too close to each other, or to vias, test points or any other exposed copper areas.


  5. Hand-fitting components, either before wave soldering or afterwards as in press-fitted components. The spacing around these components must be enough to allow human fingers to fit the components. This would affect taller components like through-hole mounted electrolytic capacitors. Where components are snap-fitted or press-fitted to the board, there must be sufficient clearance between these components and the surrounding parts, so that any bending or stressing of the board does not snap copper tracks or tear component lands from the substrate.

  6. Auto-insertion of THT components. Figure 2 below shows the mechanism for cutting and clinching THT components. This requires a keep-out area above and below the PCB. The dimensions of this keep-out area will vary depending on surrounding component height.

    Figure 2: Auto-insertion cut and clinch

    Auto-insertion cut and clinch

  7. Rework/repair design considerations. The majority of component spacing guidelines are dictated by rework and repair accessibility. Different assembly technologies require unique considerations for rework. For components with visible leads, adequate clearance must be provided to allow access with a soldering iron for removal and replacement. Special requirements apply to those device types without visible leads (such as BGAs and µBGAs). BGAs are particularly difficult because the method for removing a device involves heating the area beneath the part as well as from above. Any components in that area will have to be removed. Ideally the designer should try to avoid placing any components there at all.

Recommended component spacing matrix

A common method of presenting minimum and recommended component spacing for various component types is through a matrix. Can you describe the factors to consider when deciding the minimum distance?

Compare your answer with this one.

[back to top]