Publisher's Note: Products purchased from Third Party sellers are not guaranteed by the publisher for quality, authenticity, or access to any online entitlements included with the product. Written by a team of experts from around the world, this encyclopedic resource has been thoroughly revised and expanded to include the latest printed circuit tools and technologies — from design to fabrication. Hundreds of illustrations and charts demonstrate key concepts, and valuable tables provide quick and easy access to essential information. Chapter 1. Coombs, Jr.
Holden 3 1. Chapter 2.
Chapter 3. Chapter 4. Design for Manufacturability Tim Rodgers 43 4. Chapter 5. Holden 63 5. Chapter 6. Supplier Selection and Qualification Tim Rodgers 6.
Printed Circuits Handbook
Chapter 7. Chapter 8. Product Acceptance and Feedback Tim Rodgers 8. Chapter 9. Chapter Holden Ritchey Montrose Embedded Components Vern Solberg Drilling Processes Matthias Stickel Multilayer Materials and Processing C. Don Dupriest and Happy T. Preparing Boards for Plating Michael Carano Electroplating George Milad Direct Plating Hayao Nakahara Solder Mask David A. Vaughan Etching Process and Technologies Gareth Parry Considering the fact that a multilayer printed circuit may cost more than a thousand dollars, and have that much more value added in other components, it behooves the buyer, designer and user of printed circuitry to have at least rudimentary understanding of how they are made.
What fol- lows here is a brief description of the steps followed by most printed circuit manufacturers. Printed circuits boards are dielectric substrates with metallic circuitry photo- chemically formed upon that substrate. There are three major classifications:. Dielectric substrate with circuitry on one side only. There mayor may not be holes for components and tooling drilled into them. Dielectric substrate with circuitry on both sides. An electrical connection is established by drilling holes through the dielectric and plating copper through the holes.
Two or more pieces of dielectric material with cir- cuitry formed upon them are stacked up and bonded together. Electrical connections are established from one side to the other, and to the inner layer circuitry by drilled holes which are subsequently plated through with copper. Other materials are glass with polyimide, teflon, or triazine resins; and paper covered with phenolic resin. All of these substrates have copper foil bonded to both sides with the resin.
Planning receives all documentation from the customer: artwork, drawings, manufacturing specifications, sales order. Together with the Photo department, the planning engineer reviews all the documentation for completness and accu- racy. If it is all acceptable the following happens:. Once the drilling tape has been bought off the complete job can be drilled. The panels are pinned together in stacks from one to four high, depending on panel thickness.
These are loaded onto the drilling machine. The drilling pro- gram tape is fed into the numerical control NjC drilling machine memory. This tape commands the NjC drill to drill all the holes in the proper location with the correct drill size. Most modern drilling machines will automatically change drill bits when the hole size changes. After drilling, the panels are deburred, then go to Plating to have copper chemically deposited in the holes; this process is called electroless copper plat- ing, or through hole plating.
The electroless copper deposit is only about 20 millionths of an inch thick. The deposit serves two purposes: 1 It provides the electrical connection between the sides of a panel and to the inner layers of multilayer boards. After the holes have been plated, the circuit pattern is imaged onto the panels. The image defines the circuit for plating and etching purposes. This image is commonly applied using one of the following methods:.
The image is screened on with an ink plating resist. Both sides of the panel are coated with a thin layer 0. The circuit artwork is layed on top of the photoresist and exposed to ultraviolet light. The u. The unexposed photoresist is washed away dur- ing development, to leave the circuit pattern. After imaging, the circuitry is electroplated with copper to a thickness of 0. This second metal performs two functions: 1 It preserves the solderability of the circuitry by protecting the copper from oxidation.
Tin- lead is not the only metal used for these purposes, but it is the most common and easiest to use. Nickel, tin, and tin-nickel are used for special purposes. After the circuit has been plated to the correct thickness, the resist dry film photoresist or screened on ink must be stripped by a solvent. This will bare the unwanted copper foil which must be etched away to leave the plated circuit.
Copper foil is etched by spraying the etchant to both sides of the panel as it moves on a conveyor. The speed of the conveyor and the type of etchant used play major roles in yielding straight side walls with minimum undercut. Contact finger plating is the next operation in manufacturing the printed circuit. Contact fingers are the rows of tabs along one or more sides of the boards. These tabs fit into connectors in the electronic equipment. They must be durable and resistant to tarnishing and oxidation.
A strip of plater's tape is applied to mask off the contact fingers from the rest of the board. Tin-lead is chemically stripped from the contact fingers, and nickel, then gold are plated. Nickel serves as a wear resistant barrier between the copper and the gold. It prevents atoms of gold and copper from migrating into each other; this pre- serves the electrical conductivity of the gold.
Gold is used because of its excel- lent conductivity and resistance to oxidation. Reflow tin-lead fusing is the next operation. Tin-lead is a dull gray metal, very porous and easily oxidized. Tin-lead plating is sometimes called solder plating, because solder is an alloy of tin and lead. However, the actual alloy is not formed until reflow. Reflow is accomplished by two common methods, infrared and hot oil. A conveyorized infrared oven is the fastest method of all. However, there are times when hot oil is easiest and more effective to use: small quantities, and thick panels with large ground shield areas.
After being reflowed the panels will exhibit shiny traces and holes, with gold contact fingers, and begin to look like what most board users think of as a printed circuit. Inspection is usually performed next. Broken traces or short circuits can be fixed easily at this stage. There is no substitute for quality workmanship and no one is more qualified to inspect the work being done than the operator and supervisor at each operation. However, accidents in processing and handling do occur. The in-process inspection affords an opportunity to take stock of qual- ity factors and highlight areas needing attention.
Soldermask is an epoxy barrier applied to one or more sides of the panels. It prevents solder bridges from forming during the assembly wave sol- dering operation performed by the board user. It is silk screened on and baked to cure. After soldermask, an epoxy nomenclature, or legend, is silk screened on to one or both sides also. This silk screening operation is performed using artwork supplied by the board user to identify locations on the boards.
It, too, is baked to cure. Fabrication is the final operation performed in manufacturing printed circuits. Fabrication is the operation of cutting the board from the panel. A panel may contain one board one up or contain several boards two up, three up, etc. Each board has its own special shape and must be routed from the panels to meet tight dimensional requirements.
Beveling and second drilling are also performed in Fabrication. Printed circuit boards contain holes which are plated through, and some which are not to have any plating in them at all. Since the electroless copper process, will deposit copper inside all drilled holes, a second drilling step is needed. Second drilling is simply the operation of drilling holes which were not drilled when the panels were drilled the first time. This is typically done for the case of tooling holes to align boards for automatic component insertion machines.
It is extremely time consuming. The decision to second drill or to plug is made by the planning engineer, after reviewing the drawings for that printed circuit board. Beveling, also called chamfer, puts a tapered edge on the contact fingers; this facilitates loading printed circuit boards into connectors. If hardware is required by the board user-terminal lugs, standoffs, eye- lets, connectors, etc.
The boards are given a final inspection, usually on the basis of an A. Each board manufacturer must deter- mine the A. All that remains is to package and ship. Multilayer printed circuits are manufactured in substantially the same way as double sided boards-once the inner layers have been laminated into a panel. The basic process is similar in the Planning and Photo departments. There are added considerations, since all the layers must register virtually perfectly: 1.
Special tooling holes are punched, or drilled, into the artwork and lami- nate. These tooling holes are used to align the artwork to the laminate, and to align the layers during lamination into panel form. They assure proper registration and orientation throughout the process, including drilling. Programming Film 2. Inner Layer Film 3. Outer Layer Film 4. Step-and-Repeat 6. Touch Up 7. Circuit Spread and Choke. Basic sequence of manufacturing operations. The inner layer laminate is imaged, usually with dry film photoresist.
The image is called a print and etch pattern. There will be no plating. The layers are etched to remove unwanted copper foil, and the resist is stripped. This leaves copper circuitry on the epoxy jfiberglass substrate. The etched layers are immersed in a hot caustic oxidizing bath. This bath forms dendritic crystals of black oxide.
The black oxide improves bonding during multilayer lamination. The layers are carefully sequenced and oriented during the stackup oper- ation also called booking. During booking, sheets of prepreg fiberglass coated with epoxy resin are placed between the inner layers. This will be used to bond them together. The number of sheets is determined by the thickness requirements for the board.
The inner layers and prepreg are pinned together between caul plates, and loaded into the multilayer press. Pressure and temperature are applied to cure the epoxy and bond the layers into one panel. After the panels are removed from the press, they are drilled and processed like a double sided board except for smear removal prior to electroless copper.
Drill bits reach hundreds of degrees in temperature. This smears epoxy around the inside of the holes. If this happens on a multilayer board, the smear could prevent an electrical connection from being made during electroless copper plating. The smear must be removed by chemical means, such as concentrated sulfuric acid or chromic acid. Plasma is also used today for removing epoxy smear. For a summary of the entire procedure, see Fig.
A final word about tin-lead: Tin-lead is preferred by board manufacturers and users for three reasons: 1 It is plated over the copper circuitry to serve as an etch resist during removal of unwanted copper foil. Electrical components are soldered into printed circuit boards; if the boards are already covered by a layer of solder, a reliable joint can most readily be formed. As mentioned before, other metals are used, including gold.
None have the outstanding solderability of solder, though. A trend is underway in the elec- tronics industry toward printed circuits which have solder only on the pads and holes. The rest of each board is soldermask over bare copper. The big advan- tage of this type board is greater reliability for fine line, tightly spaced circuitry. One service increasingly offered by the board manufacturer is that of bare board electrical continuity testing.
Once the bed of nails test fixture has been built, little expense is involved in testing. Unwanted short circuits or open circuits are detected and their location printed out on a tape. This makes it easy to find problems without depending on visual inspection methods. Considering that some printed circuit boards can cost more than a thousand dollars, and may have thousands more loaded as components, it makes little sense not to use bare board electrical testing for large board quantities.
Printed circuits are built on a variety of substrates; the particular substrate used for any given design is chosen on the basis of function, operating environ- ment, and cost. Functional considerations include electrical, mechanical, and flammability factors. Environmental operating considerations include temper- ature and humidity conditions, as well as vibration and thermal stresses. Cost must be considered, since it is desirable to use the least costly material which will fulfill the preceeding requirements.
Probably the majority of printed circuits used in radio, television, computers, and telecommunication equipment are manufactured on this material; as are printed circuits for many other uses. Epoxy jfiberglass is the single most impor- tant laminate in the U. There are variations which enhance fire extinguishing properties, drilling, punching, and machining prop- erties, properties of moisture absorption, chemical and heat resistance, and mechanical strength.
These are listed below. FR-5 is used at higher operating temperatures than FR-4, where it is important to maintain electrical and mechanical properties. It is difficult to drill and machine because of its hardness. It has poorer electrical and flame retardant properties than does FR However, it is far easier to machine and to drill. Like FR-5 and polyimide, it is used in high temperature environments and is especially used for burn-in boards. It is a black laminate and will not show discoloration due to prolonged heating.
These are manufactured to reduce cost, and to alter the mechanical properties. Typically fiberglass and epoxy resin are applied over a paper core. Laminates which are made on paper will tend to absorb moisture to a greater extent. They are less expensive, more punchable substitutes for FR This material is one of the cheapest laminates made. It has been widely used for circuits which are punched from the panel form.
With greater need for enhanced chemical resistance, wave soldering, and greater overall reliability in electrical components, the use of phenolic resin over paper has diminished greatly in the U. This is a fairly cheap laminate, with good moisture absorption and electrical properties; although it is inferior to epoxy substrates, especially as regards mechanical strength and heat resistance.
- The French and Indian: War Deciding the Fate of North America.
- Printed Circuits Handbook, Seventh Edition!
- Climatic and Environmental History of Isla de los Estados, Argentina.
- State Building and International Intervention in Bosnia: After Dayton (Security and Governance).
- Navigation menu?
- Follow us on.
It is finding wider acceptance for cir- cuits used in extremely high volume, such as electronic game cartridges and disk drives. It has been manufactured primarily with nonwoven fiberglass, which has an esthetically unappealing surface. This material has excellent electrical and mechanical properties, which are maintained at elevated operating temperatures.
Polyimide is very hard and requires frequent drill bit changes during drilling. A major application is for bum-in boards, where electrical components are tested for reliability. Polyi- mide is also used for' multilayer construction to eliminate z-axis expansion in high reliability applications-usually for the military. None GI Polyimide resin on woven fiberglass.
This material is used for printed circuits requiring very low dielectric constants. Teflon is soft and difficult to drill and machine cleanly. The only difference between them is the testing method and certification supplied. The dissipation factor and dielectric constants for GX must be known, fall within a narrow tolerance, and be so tested in the X-band frequency.
Material labeled GX meets closer tolerances than does material labeledGT. The nonwoven glass offers slightly lower dissipation factors at a slightly more uniform dielectric constant. Teflon over ceramic, for electronic applications requiring high dielectric con- stants of TABLE Most common copper clad laminates. Commercial electric aJKl Tan, 0. Natural 0. Good punc:IIilll at room temper- ature. Good dimcllsional characteristics. FR-2 phenolic! Fire resistant. Very Sood Tan 0. SOmm humidity ancI temperature green 3.
Excellent chamical resistance, low warpage, and predictable maclrinabiHty. Dimensionally stable. SOmm sreen 3. Good 2. Very good Tan 2. Retention Natural 0. SOmm tures and humidity. Very high mechanical strength and very low dimensional change. OSOmm chemicals. See Table I-I. OhmegaPly is the brand name of laminates, both rigid and thin clad, which contain a layer of resistor material.
The resist is etched onto the board surface. Using a material such as this would eliminate surface mounted resis- tors to a large degree and reduce the overall size of the board. The polyimide resin is coated over DuPont Kev- lar fibers, not over fiberglass. Vectron Graphic Systems, Inc. Santa Clara, California. This is a concise introduction to the basics of computer aided design CAD and design automation for printed circuits. It is not "everything you ever wanted to know about computer aided design," but it will provide enough infor- mation for you to find out the rest.
It will not tell you whether or not you should install a computer system in your company's printed circuit design facility or tell you which one to install if you decide you need one, but it will give you some general considerations to think about, which will help you make these decisions. After reading this, you can find more specific information about how these computer systems perform, and how much they cost, by discussing them with the sales people for computer aided design systems. First, let us discuss what is meant by the two terms computer aided design, and design automation.
A CAD system is one which makes printed circuit board artwork from a layout, which has already been designed. Specific infor- mation about the layout is entered into the computer system by one of two methods discussed below; and the computer draws the artwork which is used to manufacture the printed circuit. The computer aided design system is a sub- stitute for the process of hand-taping the printed circuit board artwork. Such systems are referred to as digitizing systems, after the digital process of enter- ing the layout into the system. A design automation system, by contrast, is one in which the computer does all or part of the designing task for the printed circuit boards.
The information from a schematic diagram a list of what is connected to what is entered into the computer, and the computer goes about the task of designing the printed circuit board. The computer may ask for help in placing certain packages, or it may leave some traces unrouted to be completed by hand; but, in general, the computer is doing the designer's job. Of course, when the designing job is finished, the design automation system draws the artwork, just as the computer aided design system does.
Those who work with printed circuit boards know that taped-up artwork is no joy. Computer aided design systems produce artwork that eliminates some of the chronic problems of tape-ups. First, the artwork produced on a computer aided design system is more accu- rate than a hand-made tape-up artwork. The accuracy of a tape-up is limited by the steadiness of the hand of the taper. If the taper is unskilled, or negligent, or ill, the artwork quality suffers.
By contrast, computer aided design artwork is accurate to within about half a mil, regardless of the skill level or attentive- ness of the operator. With the computer drawing the artwork, lines that are supposed to be on 25 mil centers are on 25 mil centers, not 24 and not If the computer is told to place 8 mil lines with 8 mil air gaps, that will be just the result. More accurate artwork brings a number of advantages.
The boards will be more manufacturable. The printed circuit manufacturer will have no problems due to artwork induced short or open circuits, there will be no layer-to-Iayer misregistration, or metal bridges. Also, the printed circuit assembler will be able to make use of auto-insertion equipment for loading IC's and other com- ponents into the board. In fact, the computer aided design system should be able to punch the tape which drives the auto-insertion machinery. The greater accuracy of the artwork will allow the designer to meet military specifications, which are impossible to meet with hand taping.
A second advantage of CAD artwork is improved stability over the hand tape-up. With time, the tape of a hand tape-up can creep along the Mylar base, leaving artwork which is even less accurate than it was initially. An even worse experience is that of removing a tape-up from its folder, only to find two or three pads and pieces of tape still in the bottom of the folder. Where were they on the artwork before falling off? Since CAD artwork is made from a single sheet of photographic film, these problems do not occur. A third advantage of CAD artwork over taping is speed. Most people can enter a design into a computer aided design system faster than they can tape it.
A very fast taper can keep up with a CAD operator, but only at the expense of accuracy. Even the fastest taper has no chance, however, if any section of the board pattern is repeated. While the taper must tape the repeated section, the computer aided design operator simply tells the computer to do it. In the extreme, a memory array which would require several days of taping, can be completed by a computer aided design operator in ten to fifteen minutes. A related benefit of a computer aided design system is the ability to edit quickly, easily, and accurately. How many times have you lifted a piece of tape from artwork, only to have a tangle of other pieces come off with it?
CAD system, the computer always makes it clear which line is to be moved or removed. By entering layouts into a computer aided design system, two other benefits can be had. First, the CAD system can punch a paper tape that will drive a numerically controlled drilling machine for drilling the printed circuit boards. This tape will direct the drill to the centers of the pads on the board. Second, the CAD system can draw the fabri- cation and assembly drawings with a minimum of difficulty. Many CAD sys- tems can draw on vellum, or Mylar, with a variety of pen types; documentation grade drawings are made available as fast as the artwork is completed.
In addi- tion to using less drafting time, a CAD system always produces drawings with perfect, even lettering, correctly drawn symbols, and information that matches the printed circuit artwork-because it came from the same data base.
Responding To A Promotion?
What is the catch to computer aided design? The advantages of CAD artwork do not come unless you have a computer aided design system. To decide if you need a CAD system, you must compare the costs of the system with your costs of personnel time, defective boards from inaccurate artwork, and rework time. In contrast to hand taping, where only a light table, a stool, and some supplies are needed; the CAD system requires a large investment in equipment, procedures, and personnel training. The following section discusses the equipment and people needed for computer aided design.
Computer aided design for printed circuit artwork requires a number of special pieces of equipment. The first of these pieces is the device that actually draws the finished artwork, the photoplotter. A photoplotter uses a light beam to make marks-pads and lines-on a piece of photographic film. It moves the light beam around the film on a mechanical arm, to draw the printed circuit artwork directly onto the film.
The film is then developed, to yield the finished artwork. Most photoplotters get their data directly from a computer tape. The tape is mounted on the photoplotter with the film, the "start" button is pushed, the photoplotter reads the tape, then draws the artwork. A photoplotter generates actual size artwork, rather than a two-to-one enlargement typical of tape-ups. The photoplotter must be able to plot different pad sizes and line widths; hence, it must be able to change the size of the light beam which it uses to draw upon the film.
This change is accomplished by shining the light beam through an aperture; which makes the beam the shape and size of the desired pad or line width. The photoplotter contains a wheel, in which is mounted an entire set of apertures. The computer tape which tells the photoplotter where to plot must also tell it which aperture on the wheel to use. Although the wheel has only 24 positions on it, it can be replaced with another source of apertures. Wheels containing letters or logos are also available.
Photoplotters are notoriously temperamental pieces of equipment, requiring constant care and attention. The light beam must remain focused, and the mechanism clean, in order to maintain the desired accuracy. Light intensity must be varied by using filters, in order that exactly the correct amount of exposure may be achieved. Even computer aided design shops which maintain a full time professional to run their photoplotter expect to botch a few pieces of film now and then.
Photoplotters must reside in photographic darkrooms, with attendant requirements for lightproofing, clean air, and constant temper- ature and humidity. Because of these drawbacks, many shops equipped with CAD systems do not bother with a photoplotter. Instead, they use their com- puter aided design system to make tapes which are sent out to photoplotting services. The next piece of equipment needed is a computer. The computer makes the tape which is later given to the photoplotter.
A computer system for CAD work needs the standard list of computer accessories: disk drives to store data, mag- netic tape unit to make photoplotter tapes, a console for directing the computer, a printer for lists which must be printed, and so on. The next piece of equipment needed is a digitizer, or digitizing terminal, for entering printed circuit layouts. The digitizing terminal has a keyboard, a dig- itizing tablet or table , and a video screen. The digitizing tablet is a special type of drafting table with imbedded electronic sensors.
These sensors allow the computer to detect the location, on the table, of a special stylus or cursor which comes with the table. Data about the circuit layout is entered into the computer by pointing to the layout with the stylus. The video screen draws a picture of the layout that has been entered.
This is necessary for verifying the correctness of the data which has been entered. Another required piece of equipment for a computer aided design system is a pen plotter. Pen plotters are similar to photoplotters, except that they draw on paper or vellum or Mylar with ink, instead of on film with light. The pen plotter is plugged directly into the computer for running, rather than being driven by a tape like the photoplotter.
Pen plots are indispensable for checking correctness of the final circuit which has been entered. It is cheaper and faster to make a pen plot, and then make any needed corrections, before using the photoplotter. The final and most important item of equipment requirement for a computer aided design system is the software. It is the software which transforms the equipment from just a computer system into a computer aided design system.
Without the software, the computer will not know how to draw on the plotters. Without software, the digitizing terminal will not know how to use the tablet. The system will not put out photoplotting tapes without software. The software makes the system run well or poorly. There are four job classifications for people working with CAD systems: two of these are already working in the printed circuit design shop, the two others will be new.
The printed circuit designer is one of the existing classifications. The computer aided design system does not actually design printed circuits, it just draws the artwork; the designer is still required. The next type of person needed is the digitizer. It is the job of the digitizer to enter the designer's lay- outs into the computer system. The digitizer will take the place of the taper, already found in the printed circuit design shop.
The skills required for digitiz- ing are a little different than those required for taping. Although both need a rudimentary knowledge of printed circuit artwork, the taper also needs the manual dexterity to work with the tape. The digitizer needs to be trained in the use of the computer system; training which can usually be accomplished in a few weeks.
A new type of person who will be required is the computer operator. The computer operator is responsible for general system maintenance, performing data backups, telling the computer to make photoplot tapes, or plots, and so on. In very small operations, the computer operator may also be the same per- son as the digitizer, but a system of any size requires a separate person to take care of the computer operator's tasks. The computer operator needs about the same amount of training as the digitizer, if not a little more.
The fourth type of person needed for a computer aided design system is the manager. The manager keeps the work flowing smoothly through the system. The manager must keep the other people working as a well oiled machine. The following is a discussion of the process by which a printed circuit designer's layout is turned into finished artwork, with the use of a computer aided design system.
The first process which must occur is that of digitizing the layout, entering it into the computer. A layout of the circuit is taped onto the digitizing ter- minal, which includes the special electronic drafting table with the stylus Fig. Next, the digitizer uses the stylus to point to items of the layout. To insert a pad, the digitizer points to it, and the computer adds it to the board. To insert a package, the digitizer points to the first pin and tells the computer what kind of package to insert; and the computer inserts the package.
To repeat some feature on the board- an array of memory traces, for example-the digitizer indicates what part of the layout is to be repeated, and tells the computer where on the layout it is to be duplicated. The computer can step and repeat even very complex patterns instantly. If the board contains a ground plane, the digitizer indicates this fact, digitizes an outline of the plane, draws in the tie bars; and the computer does the rest. While the digitizer is busy entering the layout, the computer is cleaning up the entry automatically. One of the most valuable functions a CAD system performs is to evenly space lines and pads on the board.
The digitizer tells the computer that, for example, everything on the board is to be on 25 mil centers. Subsequently, if the digitizer tries to insert a line that is 23 mils from the pre- vious line, the computer will simply move the line 2 mils over to make the required spacing. If the digitizer then indicates a line that sags a little at one end, the computer will straighten it out automatically.
The computer is told what line width to use, and it uses just that width.
Handbook of Printed Circuit Manufacturing | Raymond H. Clark | Springer
After the digitizer has completed entering the layout, the CAD system pro- duces a pen plot of what the entered circuit will look like on film. This plot can be made in multiple colors, to look like the original layout, or the system can plot each side of the board separately. This plot is used to check that the data entry has been done properly. The digitizer, or a person other than the digitizer, must verify that all the lines and pads have been entered, and that all the traces go to the correct locations.
Typically, the plot is checked against the original schematic diagram. It is not necessary to check spacing of the plot; the com- puter system will already have corrected that. There is no chance that something will be a few mils off one way or the other. With digitizing systems, things are either right, or they are very obviously wrong. After the plot has been checked and corrected, it is given back to the digi- tizer, who goes back to the terminal and makes any required changes. The changes are made in a manner similar to the original data entry.
The digitizer points to the offending line or pad, and indicates where it should be, or that it should be removed. The computer will move or remove discrepant items. The digitizer can check on the video screen that the changes have been made cor- rectly. When changes are completed, the job is ready for photoplotting; a sec- ond pen plot may be run first, to spot check the corrections. An alternative method of entering a layout is to enter the information from the keyboard, using a net list.
The digitizer looks at the connections which are indicated by the schematic diagram Fig. The advantage of working with a net list is that layouts can be quickly entered. However, this is done at the expense of accuracy. Generating a net list increases the possibility of error.
When the tape is done, it is taken to the photoplotter-either in-house, or an outside photoplotting service-and the artwork is made. Some last minute changes can be made at this point, if desirable. The board is usually reduced from 2: 1, the scale at which it is digitized, to 1: 1 for photoplotting. However, 2: 1 artwork can also be photoplotted if required. The sizes of pads and traces can also be changed en masse if necessary. After photoplotting, the artwork is checked. It must be verified that the pho- toplotter was in focus, that the right pad sizes and line widths were loaded into the photoplotter, and that the film has been properly developed.
If no problems come to light, the artwork is ready to be sent to the printed circuit board man- ufacturing shop. The CAD system draws the pen plots of the fabrication and assembly drawings as the artwork is being photoplotted so that these drawings are ready at the same time Fig. BAR J1 32 1. The circuit artwork must contain all the connections shown in the schematic. Computer aided design systems for your printed circuit artwork cost a lot of money.
Obviously, a lot of thinking and evaluation are needed to determine which one is best. In addition to complete stand alone systems, your options include smaller systems; timesharing services; service bureaus; and of course, continuing to perform hand taping of the artwork. Depending upon the specific requirements of your company, any of these might be the correct decision. The first decision which must be made is to determine if a CAD system is needed at all.
Compare the cost of a CAD system with the costs of continuing to cope with current artwork generation problems. The costs of the CAD sys- tem are the costs of the original installation, plus the cost of ongoing mainte- nance, and the cost of additional people which may be needed to run the sys- tem. These are all costs available from CAD salesmen. The cost of not using CAD is more difficult to calculate, because most of the costs are indirect and intangible.
How much is it worth to be able to get artwork faster-hence get your products to the market faster? The CAD system will speed up the process dramatically. How much is it worth to avoid downstream manufacturing difficulties caused by inaccurate artwork? CAD systems help improve manufactur- ing yields and reduce field failures. Example of photo-plotted artwork. How much is the ability to make design changes quickly worth, in order to be able to respond to changing market demands?
How much is it worth to be able to hire and train a relatively unskilled digitizer when you need one, rather than scour the countryside for a new taper when the work load increases? The CAD system requires less expe- rienced people than does a manual taping operation. It is possible for a CAD system to increase your department's efficiency enough that there will be no need to hire additional people to handle increasing work loads. What are the critical factors which make one system better than another? Before answering this, it is important to dispell some common myths about the components of the system.
The photoplotter does not determine how useful the system is. All photoplotters are, for the most part, very much alike in terms of accuracy. There are differences in speed of photoplotting, but the photoplotter will probably never be the bottleneck in any operation, as long as it is not out of commission. An exception to this would be when there are a dozen or so digitizers all working on different circuits at the same time.
Also, the type of computer in the system should not be used to judge the system. A salesman may say that a given system is faster and more accurate because it contains a 32 bit minicomputer; but in truth, the speed and accuracy of a computer aided design system are more dependent upon the ingenuity of the people who designed and wrote the software than upon the particular hardware upon which the system runs. Lastly, the features, such as color graphics, should not be used to judge the system.
Although color may be impressive, and useful in some applications, it is not that useful for a double sided printed circuit, and it is not that expensive. Do not assume that any particular feature is all that useful, until you have demonstrated it for yourself. The prime consideration for a computer aided design system must be throughput. CAD systems are justified on the basis that they generate more artwork than competing methods.
A system that cannot get the work through is worth nothing. Here is a list of considerations that can be used to help judge the value of a CAD system: 1. Decide which editing features are useful for your application. If your boards use silk screens, check on how easy it is to do silk screens. If you use buried power and ground planes, find out how much trouble they will be using a given system. One feature of the CAD system that dramatically affects the speed at which work can be done is the speed at which the system can display the layout on the video screen of the digitizing terminal.
Whenever a change has been made, the operator will want a display to check the correctness of the change. It is also important to know how fast you can get pen plots. This is a function not only of the pen plotter, but also of how good a duty cycle you can expect. Will the plotter run all night unattended? Will the paper have to be changed after every plot? If you run a one or two shift operation, these two questions may be more important than the actual speed of the pen plotter.
How many digitizing terminals will be needed to get the work done? If the workload is large, the CAD system must be able to handle as many digi- tizing stations as is needed. Will it be necessary to have terminals at long dis- tances from each other? Do not assume that a given system will handle it; not all of them will. Make a determination, as best you can, of how much trouble you will have with a given system. All complicated machinery breaks down occasionally. Find out from other users how reliable various CAD systems are. Find out how long before the CAD manufacturer gets the repairman out to fix it.
How much work will be lost if the system goes down? The last thing entered by the digitizer? Everything since lunch? A whole days work? Find out. A design automation system differs from a computer aided design system. The design automation system does the actual printed circuit board design, in addi- tion to generating artwork from designs prepared by printed circuit designers.
It does this with the addition of a large block of software written for that pur- pose. The software is intended to take the place, at least in part, of the printed circuit board designer. In other respects, design automation systems are similar to computer aided design systems: it takes about the same staff to run them, and they run with the same computer hardware, with the possible exception that a somewhat more powerful computer and a disk drive may be necessary.
A design automation system usually has a CAD system embedded within it, so that the system can be used purely as a CAD system, which is necessary to make changes to digitized boards, or to input design automation into the system. Using a design automation system for designing printed circuits offers several advantages over the manual design process. Computers are well suited for this work, and function well in the printed circuit design environment. A design automation system offers fast design turnaround.
Boards that might take a designer several weeks can be designed by the computer in a day or two. There are many reasons why computers are so much faster than a designer. First, the computer can check a large number of possibilities very quickly. Although most computers tend to bog down toward the end of a design when the last few traces are being routed , in the beginning they are literally as fast as light- ning, routing traces by following rote rules of good board design.
Next, the computer spends no time thumbing back and forth among the pages of a large schematic. Design automation systems first compile a list of the interconnects to be made on a board; after this, the system needs only a few thousandths of a second to find out what is to be connected where. Finally, the computer spends no time with pencils. Once it decides where a trace is to go, it places it there almost instantly. There is no sharpening of pencils, no aligning of rulers, and no fuss with other mechanical aspects of drawing a layout.
A design automation system produces printed circuit designs that contain no rule violations. The computer patiently checks every line it places on the board against every other line and pad for spacing violations. The design automation system also avoids careless mistakes in connecting things. The computer never looks at a 14 pin package and mistakes it for a 16 pin package. It never misreads the signal pin assignments out of the manufac- turer's specification book.
For all these reasons, printed circuit board design departments can save money with design automation. Since the computer can design routine boards, the designers can turn to designing very complex boards, or boards in which signal shielding or signal length are of paramount importance-boards that the computer is likely to have trouble with.
Manually designed boards must be exhaustively checked for correctness. With design automation, this is greatly simplified, since if the input to the computer is checked and is correct, then the design can be relied upon to be correct. Design automation systems provide all the advantages of a CAD system. A design automation system photoplots the same accurate, stable artwork as a computer aided design system, yielding the same manufacturability and reli- ability in the artwork as boards from a CAD system.
The design automation system also produces paper tapes for numerically controlled drills. Fabrication and assembly drawings are also provided from the same data base that creates the artwork. To use a design automation system for creating printed circuit designs, you must follow certain procedures. The net list of the board must be entered into the computer. This can be done by typing in a prepared net list, or by using the digitizing terminal to enter the information from a schematic diagram. If this latter method is used, then the design automation system must have a pro- gram that can produce a net list from a schematic diagram.
If this latter method is used, then the design automation system must have a program that can produce a net list from a schematic diagram which has been entered. Next, physical data about the board must be entered. The computer must be told the size of the board, the location of power and ground buses if you want them located in a certain place on the board , the locations of connectors to the board, locations of areas where no traces should be routed, and so forth.
Again, there are two ways to enter these data: the first is to type in the x- and y-coordinates of the various features of the board, and the second is to use the digitizing system to draw in the various features. If the digitizer is being used, it may be possible to design in a few critical traces before the computer gets started on the rest of the board. Next, the package placement scheme must be decided.
When this is the case, packages must be placed by the designer, who types the information in, or uses the digitizer. Systems which do have package placement programming will consult the interconnect list then make decisions on how the packages are to be placed in relation to one another. Some systems request guidelines regarding how packages should be placed on the board; for example, whether the long or short axis of a chip should be placed along the horizontal direction, or how many rows and columns of ICs should be used.
In situations where several electronic devices are packaged in one IC, the design automation system must also make decisions about logical groupings for the devices within the available ICs. Ideally, this package assignment should be done in conjunction with the package placement; but some systems do the assignment first, as a separate step. After package assignment, the system must do pin assignment in situations where there is a choice of pins for use by certain signals. Although printed circuit designers typically do this as they are designing, computer systems usu- ally do it first.
This not only reflects the fact that it is easier to program this way, but also that computers do an adequate job of pin assignment in advance.
The printed circuit designer must thumb through pages of schematics to make reasonable guesses about what pin assignments would be advantageous to min- imize trace crossings. The computer can find out where each of the signals to a given IC are connected, and make pin assignments in a second or two. Routing is the last phase of the automated design cycle. The computer now places traces on the board to make electrical connections. This program is usu- ally the most time consuming part of the operation for the computer.
The com- puter must look at interconnects one at a time, and insert each onto the board. It uses factors such as how crowded areas of the board are becoming, where previously placed traces are located, and which traces go in the same general direction, to optimize the route for the given trace it is trying to place. Most computers sort traces according to some criteria, such as routing the shortest one first, or perhaps the longest, or perhaps traces which run to a given con- nector on the board. When the computer has finished with the board, a pen plot is run of the resulting design to see if any problems arose, and to print a list of traces that the computer could not insert, if any.
If there is a list of "fails," the computer's design is turned over to a designer for completion. While computers are very good at placing ordinary traces in the early stages of designing a board, they are not very good at squeezing the last few traces onto a board. The more out of the ordinary a job is, the more difficulty the computer will have with it. Designers usually have little trouble placing the last few traces on the board, since their thinking is more flexible than the rules followed by computers.
When the board has been completed, additions and changes which may be necessary are entered into the system by the computer aided design functions. After the design has been checked for accuracy, it is photoplotted just as it would have been with a CAD system. Since design automation systems are a viable alternative to designing by hand, they must be given consideration.
To do this, it must be determined whether the company can make economic use of a design automation system, by com- paring the costs of the two methods. The costs of a design automation system can be obtained from a salesman, and the criteria and numbers evaluated. The costs include the cost of the initial purchase price of the system, the costs of ongoing maintenance, and the cost of additional people that may be needed to run the system.
The costs of design- ing printed circuit boards by hand is more difficult to determine. The cost of the designers' salaries and the overhead to sustain them are obvious. But, there are equally valid intangible costs to factor in. How much is it worth to a com- pany to be able to get products to market several weeks ahead of the compe- tition, by virtue of being able to save weeks on printed circuit design time? How much field maintenance will be saved by having printed circuit designs which are right, rather than ones which have had a few wires added to them?
Convert to and from PDF
How much is it worth to be able to handle a surge in the work load by having the computer work overtime, instead of having to scour the countryside for extra designers, hire in job shoppers, or shed critical designs to an outside service? There is so much information available on design automation systems from competing sources, that the problem is to figure out which information is important, and which is not.
First, the statistics about computer speeds, disk drive speeds, programming techniques used on construction of the design auto- mation programs, algorithm names, and so on, are unimportant. A design auto- mation system for designing printed circuit boards requires a complex inter- action of computer, disk drive, and software.
There is little reason to believe that a design automation system with a" nanosecond cycle time" for its computer is any faster in designing printed circuits than a design automation system whose computer has a slower cycle time. There is no reason to believe that a router using "Lee's algorithm" will produce better routed boards than one using a nameless algorithm. The most important thing about a design auto- mation system is whether or not it will design your boards effectively. Adver- tising puffery and sales rhetoric have never designed a printed circuit.
There is a great disparity of capabilities among the various design systems. Is data entered by constructing a net list and then typing it in, or can a schematic be entered from which the computer will extract the net list? Can physical data about the Board size and location of connectors, and so forth, be entered through digitizing, or must the coordinates of each feature of the board be entered by typing?
Does the computer decide package location on its own, or must the designer do this? The question which must always be in mind is whether or not the system can help design your boards. To help you answer, most design automation man- ufacturers tell you their "completion rates," which is the percentage of traces the computer designs before the job must be turned over to a designer. Most design automation systems cannot design every board to completion. There are two reasons to be skeptical of this number as a measure of how well the design automation system will design your boards.
First, unless a benchmark is run on the design automation system, the completion rate told to you by the manufacturer will be the completion rate on someone else's boards. Since your boards may be totally different from someone else's, the completion rate on your boards may not be as great. It is not true that a router which does well on one kind of board will do equally well on another; very minor changes in board styles can markedly affect completion rates.
Any of the following fac- tors, for example, will affect completion rate:. Whether or not the board is long and skinny. Whether or not it contains edge connectors. Whether or not it contains buried power and ground. Whether or not it contains chips rotated 90 degrees.