designer’s notebook
Is it Time to Design Printed Circuit Boards around the Bill of Materials?
Make every chip a layout unto itself.
Being a printed circuit board designer is not easy. Parts we used to take for granted have become really hard to come by. Geopolitical trade wars and a pandemic were serious triggers for the undersupply. We really didn’t need a Japanese chip factory to burn down to make things worse. A giant cargo hauler clogging up a vital shipping artery for a week was no help either.

The fear, uncertainty and doubt sown into the supply chain put the squeeze on purchasing managers who, in turn, did their best to secure as much material as possible. Ordering more inventory than their forecasted requirements is a typical kneejerk reaction for the big players. Some purchase orders may be defensive measures, an effort to block competitors that are caught shorthanded themselves.

Automakers are a vital sector of the US, German and Japanese economies. They have been busy lobbying their respective governments to pressure chipmakers, with the goal to create a sufficient supply of devices for the vehicles they want to build. Propping up that industry with their ruggedized devices leaves even less bandwidth for other industries.

The mainstream chips that remain have become a strategic and tactical pawn to be overbought to ensure a supply down the road. Scarcity breeds scarcity as the big players use their clout to absorb as many chips as possible. Things will remain messy and uncertain for the PCB design engineer who must accommodate the supply chain crisis.

Lead times are one of the key elements in forecasting (Table 1). What I’ve heard from contacts in the chip houses is they are facing a lot of uncertainty in this environment. If everyone is trying to get a larger allocation than they require, what is the true demand? It’s not like they can announce a new fab and, presto, more chips. It takes chips to make chips, and the companies that make equipment for the chipmakers are in the same boat we are. It’s a vicious cycle.

Table 1. Material Lead Times, in Weeks
Table
It’s not only lead times but also the confidence in those numbers as they continue to grow. We don’t see light at the end of the tunnel yet. When you don’t know what cards you hold, or even how many cards are in the deck, planning with any certainty is difficult. Do you commit to a higher price to ensure a supply six or even three months out? These are tactical problems, but they become technical problems of the highest urgency once we find a viable path forward.

Hurry up and wait? We need something better than that. Getting away from traditional program management in favor of agile program management might be a better solution for uncertain times like these. Maybe you’ve heard of agile methodology. It’s more of a software development thing but can be applied to hardware as well. Instead of projects that have concrete start and end dates, “agile” relies on getting everyone together, setting a goal around a set of parameters and executing on that plan in a sprint. Learn and repeat.

Design pre-use. In this case, the parameters are the line items that make it into the bill of materials. The full set of items has a lot of redundancy, and only those most likely to be available go into the product. Every chip is a layout unto itself, and a library of chip layouts are generated as you go. Some circuit elements may never be incorporated at the board level but are on hand if you need them.

Instead of cataloging design reuse modules, you’re designing sub-circuits for everything. Even if it is a one-off on the board, it can be created as a module to free it from the constraints of the reference designators. The idea is to keep your options open by having a sub-layout for everything you might need.

Modular schematics may not be as ecofriendly since each chip will use a full page or pages, leaving you with some white space. Some of us still work from paper schematics, especially during placement and as prop for the final design reviews. In the interim, people are proving out the sub-circuits to reduce the technical risk while the tactical risks are sorted out.

This is more about simulation than actual bench work, but it is possible to generate PCBs for the sub-circuits where more data are required. (Assuming you can get your hands on sample quantities of your projected components, that is.) Various functional pieces can be cobbled together on the test bench to get a better characterization prior to the final form-factor.

What you want is a collection of circuits you are confident will perform as required. Then, pick and choose the sub-circuits so you can complete the BoM and deliver the board within the budget and on schedule. The single breakout board becomes an interchangeable system of multiple printed circuit boards.

Multiple footprints for an at-risk component. To alleviate uncertainty, it may be possible to have a single footprint that supports two different components (FIGURE 1). It is not ideal, and it will want to be distilled to a single footprint on the next iteration. This is the key. You’re working with a dynamic universe of components and need to be able to pull a board together quickly.

U27 and U46
FIGURE 1. U27 and U46 are crystals of the same frequency but with different types of packaging. They are mutually exclusive assembly options, depending on the component available.
Interposers are another way to keep your options open. The concept is simple: one footprint on the bottom of the substrate and a different one on the top. The two are connected such that an old board can learn a new trick or at least accommodate a second-source component with a different footprint.

Rushing a board may not be an optimal solution, but rest assured, an agile methodology means doing one iteration after another. Fail quickly and move forward with what you know for the next time. Managing this risk is a team sport that requires solid communication. Interesting times call for more evolved solutions.

John Burkhert Jr. headshot
John Burkhert Jr.
is a career PCB designer experienced in military, telecom, consumer hardware and, lately, the automotive industry. Originally, he was an RF specialist but is compelled to flip the bit now and then to fill the need for high-speed digital design. He enjoys playing bass and racing bikes when he’s not writing about or performing PCB layout. His column is produced by Cadence Design Systems and runs monthly.