In the last article we described what a PCB actually is. It was about the materials that are used and the process of doing a PCB design to have a proper layout. We will now describe the engineering process starting from scratch until manufacturing a PCB. But, one has to notice: Until the printed circuit board is ready, it mostly takes a minimum of three iterations of all steps of the development cycle, which is described in the following. The whole cycle has mostly to be done manually. However, Celus automation software offers a lot of options to reduce the effort, which will be mentioned shortly in the article. Here is an overview of the described process. You can compare the manual process like it mostly is nowadays to the one with support of Celus.
1. Define what you want
The first phase is about analyzing requirements. You need to exactly know what you need the PCB for and which function it should fulfil to be able to define a project scope. For this purpose, you need to read and understand the standards of this special project. Celus can help to reduce this effort through integrated data management for a database, where those standards are saved. This can save 10% of the development time.
2. Establish your concept
The first thing to do here is to research available topologies. Every electronics engineer knows: This is a great piece of work including reading long sheets with much information. Afterward, basic topologies have to be dimensioned and first block diagrams are created with different components, which helps to select the most appropriate ones. After choosing a variety of components and specifications, more detailed block diagrams have to be created to get an idea of the circuit that is created on the PCB. This is a lot of manual and repetitive work, but it is necessary to have a good basis for the concept. Celus can help with these steps as it has an UI developed for diagrams and additionally, it offers a database, which allows PCB designers to use previous work done by others instead of starting from scratch.
3. Select the best-fitting components
In the third phase, the concept the engineer created in the second phase can now be used to select the best fitting components. This all pays to the goal to fulfill the requirements that were defined in phase one. What the electronics engineer has to do now is browsing through websites from manufacturers and distributors and an internal database to find pages of data sheets with specifications for components. They have to be compared to each other to find the one fitting best for the project, which again is a lot of manual work that can be quite repetitive. However, the design process can only proceed if there are the best fitting components for every purpose found as later changes will cause much extra work. This whole process is automatically done by Celus: The software will automatically compare component specifications in its database and select the best one for a special purpose.
4. Draw a schematic of your printed circuit board
Of course, drawing does not mean using a pen and paper rather than an electronic solution. The most common methodology to draw a PCB schematic is to use electronic CAD-tools, a PCB design software. You can read more about CAD-tools as PCB design software in this article. What the PCB designer has to do here is a partition block diagram, either of the whole PCB or of modules. The modules can be composed to a full printed circuit board afterward. Doing a block diagram basically means placing blocks representing the parts and assign the component properties to them as additional information. These schematic blocks have to be connected and the net properties have to be added. Finally, the engineer has a schematic either of certain modules containing components or of the whole circuit board. Now, the compatibility of the connections between the integrated circuits has to be checked. If there are problems between these circuits, the engineer needs to fix them by looking for new parts, different placing, or a wrong connection. Celus can automatically layout the plan and design schematics, optimize component placement and routing based on design rules, and generates a PCB layout.
5. Simulate the schematic
To be sure, that everything runs correctly a simulation of the first PCB draft is being done as the fifth phase. For this, the scope of simulation has to be defined. Then, the simulation files have to be created and not surprisingly, the simulation has to run. The hard part is then to evaluate the results of the simulation and to conclude which adaptions to the design have to be done.
6. Specify board stackup and technology
After having a detailed plan about the requirements for the PCB, the next phase is to plan the production of the board. The stackup is a cross-cut of the PCB one is planning to design. In this board stackup, it is defined how many layers are needed and the total thickness of the manufactured PCB. For this, the engineer compares different board stackup tradeoffs. With a larger search space, the chances of a better board stackup increase. On the one hand, this means some extra work, but on the other hand, it may reduce later stage changes due to optimizations. Also, one needs to compare different manufacturing technology tradeoffs. Introducing design rules, that are aligned to the knowledge database, can reduce the effort by helping the engineer to find the best solution. If the most appropriate producing technology has been selected, the electronics engineer knows how many layers he has to design for the printed circuit board design. This is what the PCB is mainly made of: soldermask (an insulating material to prevent shorts), copper (deducting traces), prepreg (holds the layers together), and the core (special fiberglass, called FR4).
7. Design process: place components
In this phase, the printed circuit board design really starts. Now, what has been done in the schematics has to be done for the final PCB is to place the components: A block diagram has to be created and all specifications added. The difference between the schematic and the board design is, that certain specifications of the shape and size of the board as well as the stackup have to be respected. As the first step, the components or modules that have been specified as schematic have to be placed. For this, the engineer compares different board placements and options and their tradeoffs. More space means it is much easier to place everything fitting on the board, but more space also means more manufacturing costs. So, it really is a demanding task to find the right tradeoff. This can be done completely automated by the Celus software.
8. Design process: route components
After placing all elements on the board, everything has to get connected as the next phase. In the routing process, the state-of-the-art strategy is trial-and-error based on a selected routing strategy. Sometimes it is even not possible to do a complete routing with a certain placing specification. Then, the engineer needs to redo parts and try again. The primary goal here is to prevent shorts as the copper, of which the routes are finally made of, is conductive. Celus can do this completely automatically.
9. Check again
In this phase once again, the electronics engineer verifies design rules and reviews the whole design. Additionally, he or she has to calculate the trace lengths, impedance, trade resistance, and much more. These values have to be compared to the target values — if they do not match one has to find adjustments to achieve the goal. Celus software can provide all needed outputs as a list of requirements.
10. Prepare to manufacture
Finally! The last phase! (Well, maybe) If everything functions correctly and properly fulfills all requirements, the PCB can be produced. For this, the engineer first needs to set up CAM configurations, then generate the needed manufacturing files and finally creates datasheets for the board. Celus provides this fully automatic.
About the PCB manufacturing
The process of manufacturing is also very complicated, but we will describe it here very simply. The commonly used material of a PCB is fiber glass, more specified FR4. On this, a copper foil is laminated. The copper traces are generated by a chemical etching process. To prevent the traces from oxidation and shorts, another insulating layer is added: the solder mask. The solder mask is mostly green colored and that is what we typically have in mind when thinking about PCBs. Finally, the layer of silk screen is added, which are the wight marks that can be seen on many boards. They do not have any technical purpose, but they make it a lot easier to understand the PCB and be able to orientate on it. Especially when it comes to error spotting, marks can help a lot.
As the last phase, the component or modules have to be molded on the board. A very popular methodology that is often used since the 1960s is the through-the-whole technology, where the components are soldered on the board through wires threading through the holes.
The whole process of multi-layer PCBs has to follow the same steps, but it is a whole lot more complicated, especially regarding the copper traces. Depending on the test results after manufacturing, the process can be started again from the beginning or some further point. So, if you are a very test-driven person, you would enjoy this process and through-hole technology. Until the printed circuit board is ready, it mostly takes a minimum of three iterations of all those steps in the development process. This is because you have various prototypes first and a pilot run to be absolutely sure that the series production will work with this design. One can assume: The higher the point where a new iteration has to start, the higher are the cost of changes, in both, money and time.
One could compare this to development cycles in software development. A decade ago the usual process for software to be developed was to use the waterfall model and the average lifecycle of the developed software took up to two years. The entire development team sat together and discussed the progress with their project team. Nowadays the majority of software developers use the agile model and iterate their processes one after another.