Design for Manufacturing

The majority of new part and assembly introductions happen during research and development. Because the components are new, very few particulars are known, other than the information documented on datasheets or in trade press announcements.

An engineer may have read a magazine or journal article featuring a new design with application notes that highlight a new part. Or perhaps demand for the new part may have come from viewing a competitor's product or having discovered the component at a recent tradeshow. Questions concerning cost and sample or volume availability remain unanswered until the research for the part selection process is complete.

Think design, early on
Manufacturers, distributors, and logistics are standard considerations when it comes to part availability and suitability. But what happens if a component has been designed and prototyped by your company — to your form, fit, and functional requirements — and the circuit design, materials, or processes required to build the product are so unique that no other manufacturer has ever seen, much less produced, anything like it?

Unless you can produce the product in-house to meet your projected volume requirements, you have already put yourself behind the virtual eight ball. If you wait until the prototype is built and working before identifying a potential outside manufacturer, your eventual supplier likely will have to climb a pretty steep and costly learning curve to meet the delivery schedule for your finished product.

With custom materials and dimensions, your company may have to underwrite the supplier's tooling. If your product requires unique testing or inspection methods, the supplier may have to purchase or rent test equipment on an expedited schedule. This could add more cost, as well as delays if the equipment is unavailable or has an extended lead time. Training production or test personnel will also add cost and delays.

Working with engineering
If a company wants to minimize both cost and delays, potential suppliers for proprietary parts and assemblies must be identified as early as possible during the R&D process — as soon as engineering knows and can articulate the particulars of the build and test requirements. The nondisclosure agreements should be comprehensive and put in place ASAP. Potential suppliers should undergo a thorough capability assessment at their premises, including a historical business analysis.

Once the nondisclosure agreements are signed, the supplier or manufacturer should be invited into the design meetings, where the manufacturing particulars can be reviewed and modified to meet the supplier's core competencies and equipment capabilities. The goal of these meetings is to design the product to match the manufacturer's most efficient methods of operation.

The term for this is “design for manufacturing” (DFM). If the manufacturer has input into the build process, the supplier will be up to speed by the time the product is ready for mass production — with no surprises that can slow down the production cycle.

Hard lessons
Recently, I was part of a design team that developed a board with a 0.004-inch trace width and 0.004-inch trace-to-trace spacing. It wasn't as easy as calling a PCB fab house and sending it the zipped drawings and Gerber files. The first three vendors selected from the requests for quotes said they did not have the technology for such tight trace spacing. That caught us by surprise. We had to go around the block several more times, and we ended up at an out-of-state fabrication house. We had to pay the extra cost for the special service requirements, and we were crossing our fingers that there would be no ugly setbacks due to product quality issues.

In hindsight, as part of our early production planning, we should have identified several fab houses that could handle the 0.004-inch narrow trace. For our first six boards, we paid $659 each for a one-week turn. Our second round of six cost us $354 for the same turn time. The difference was that we had the time to search for another PCB fab facility. This one had been doing narrow traces for more than a year, and it had developed cost-saving techniques that it was able to pass on to its customers.

This was a hard lesson in DFM. In this case, the DFM acronym could have stood for “design forfeiting money.” We paid good money that could have covered the cost of two prototype builds had we gone to the correct manufacturer in the first place.

Bring the potential suppliers and fab house into the early product design stages, and save your company a lot of money, time, and grief.

Related posts:

17 comments on “Design for Manufacturing

  1. Nemos
    February 14, 2013

    It sounds wise and also a truly professional way acting as a developer, the question should be why a high-tech company doesn't follow this advice at the very beginning?

  2. Brian Fuller
    February 14, 2013

    The late great UCLA basketball coach John Wooden taught his players the extreme value of preparation by teaching them, every year, how to first put their socks on correctly and then lace their basketball shoes correctly.

    What Douglas outlines here, as a process, echoes that. But as we know, the best-laid plans go sideways when time-to-market pressures get too intense.

    If only we had the firmness of Capt. James T. Kirk as he rocketed into some chaos: “Steady! Steady!”


  3. t.alex
    February 14, 2013


    Because of pressure. Pressure to bring in products in time. Pressure for more sales. Pressure from management who could not see the good points of such early planning.


  4. Ariella
    February 14, 2013

    @Brian Yet another valuable lesson we can learn from Star Trek. Another one would be not to bring any Tribbles onboard.

  5. _hm
    February 15, 2013

    Yes, main reason is presssure. Also, design changes many time to achieve better product specifications. It looks like real time engineering and sometime it add significant cost too.

  6. Brian Fuller
    February 15, 2013

    I've never been able to hear or read that word “Tribbles” without laughing aloud!

  7. Ariella
    February 16, 2013

    @Brian that was definitely one of the best episodes. 

  8. prabhakar_deosthali
    February 17, 2013

    Design for manufacturing paves the way for a successful product for a successful design. Many a times it may happen that the design department may use a totally new technology in terms of components or the process. In such case the design department must have the ability to lay out the manufacturing process very clearly at the design time so that the procurement department will be able to identify the suitable manufacturers -or if not available develop new suppliers who can set up the required processes .


  9. dalexander
    February 17, 2013

    @Prabhakar…you nailed it. Engineering should understand all manufacturing processes, both internal and external to the company. Design for Assembly and Design for Test, coming up next.

  10. Brian Fuller
    February 17, 2013

    Here's a question that feeds off @Douglas and @prabhakar… organizationally, who “owns” the problem of inculcating that DFx mentality into engineering? 

    In a highly specialized design chain, this can be a challenge, no?


  11. dalexander
    February 17, 2013

    This is a shared responsibility that involves Purchasing, Materials Planning, Sustaining Engineering, Component Engineering, Project and Program Managers, PCB CAD, Mechanical Engineering, Test Engineering, and people who are responsible for the company's supply chain. Because of the number of people that are impacted, this is why the best approach is a process called “Concurrent Engineering.” I wrote on this recently as a fundamental good development practice where participants representing the various departments become involved at the earliest stages of development and continue to participate until the New Product Introduction process is complete and Engineering as released the product to Manufacturing via an Engineering Change Order. Usually, Revision control numbering format will indicate if the product has been released. Changing from version A1,2, 3 etc. for product still in engineering to the Numerical designator in first position as in 1A,B,C for released goods, serves to keep the record straight as to quick ID of product status. short answer…use Concurrent Engineering practices and DFM, DFA, and DFT will be covered.

  12. dalexander
    February 17, 2013

    Possibly the genetic pool from whence come Furbees.

  13. prabhakar_deosthali
    February 18, 2013

      I would prefer to  say  “Who owns the solution?” instead of “who owns the problem.

    Because if the solution ( to manufacturing) is thought right at the time of design then the problem may not exist at all.

    So it is the joint responsibility of R & D and engineering to create a product designed for manufacturing.

  14. t.alex
    February 18, 2013



    This is an interesting question. We can say everybody has to own this problem and solve it and it usually turns out nobody is pushing for it from R&D to production. From the management team, they has to definitely own and push for this problem at least.

    February 18, 2013

    DFM for commodity items is critical but owning a unique manufacturing capability can be a real competitive advantage.  The best companies innovate in all areas and oftentimes unique solutions that nobody else can design or build are key to success.

  16. Brian Fuller
    February 18, 2013

    This is an interesting sub-thread with good inputs from @t.alex, and @prabhakar and of course, Douglas. BTW, here's the concurrent engineering piece Douglas referenced in his comment: {doclink 255727}.


  17. Brian Fuller
    February 18, 2013

    Here's Douglas' piece with a proper link: The Art of Concurrent Engineering.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.