Design for Manufacturability has been around for a while now. The idea is to ‘analyze’ the design data provided. However, does it provide value to the assembly manufacturing organization? Does it increase revenue or reduce cost? … let’s expose this aspect of the process engineering task.
Realistically all manufacturers are reviewing the design data. It’s a necessary evil. Most organizations have a New Product Introduction process where first articles are carefully ‘walked through’ production. In essence going through the process once, keeping notes on what changes are needed when the production run happens, updating the production documentation and appropriate processes should be enough to push it through. For low volume production, this should be enough…..in some cases even mid-volume.
Contrasting this is the prevailing truth. All 1st tier organizations have DFx groups reviewing all designs before manufacturing. Most tier-2 and tier-3 organizations already have implemented or are rushing to implement DFx strategies in their organization. There are more tier-4’s that have implemented some DFx capabilities than not. Why!
Being the curious cat, I had to ask.
It seems the reasons can be broken down into 2 main reasons, in priority.
1. DFx as a service
Assembly manufacturers know their business. They know what should work, and what may not work. As in-house assembly has been more and more outsourced, this knowledge is also more and more outsourced. Design houses, designers and OEMs rely on the knowledge and best practices of their assembly manufacturers to ensure that designs are manufacturable with reasonable yield and quality. (Almost all designs are manufacturable…it is a question of “at what cost?”). Why not make some money based on the value this knowledge provides? In fact, why not charge for reviewing the designs, even if that assembly manufacturer may not get the actual order to build. There is money to be made. If an automated DFx software system that has several hundred rules-based analyses runs a Manufacturing Risk Assessment and generates reports is purchased and implemented, and a knowledgeable process engineer or manufacturing engineer is put to the task of managing the Engineering Rules, reviewing the defects, generating the report and working with the customer, then all that’s left is to charge the customer. Is it really that easy?
2. DFx to reduce risk to production
Being able to identify production stopping issues before production is immensely valuable. Wasted product and raw material, wasted time troubleshooting process problems, rework and re-ordering of material delays could seriously affect manufacturing cost or worse miss delivery deadlines to customers, adversely affecting customer satisfaction and future business.
A Dfx software system that can run Manufacturing Risk Assessment on the following categories would allow process engineers and manufacturing engineers make the adjustments, or go back to the customer immediately (not days after receiving the design package).
Supply chain risk – analyzing the Approved Vendor list and ensuring that all the Vendor-MPNs for a specific Customer Part Number called out in the Bill of Material are physically alike. For those components that are not alike, exclude them from the ordering list for materials, right from the beginning. If they are accidentally ordered after production has started, that order would immediately stop SMT Machines, as the recognition and shape definitions would fail. A delay would occur. Costs occur. Time wasted.
Joint solderability risk – Component footprints that are not designed well so that heel fillet or toe fillet distances are too small will create a situation whereby not enough solder goes down. Is the stencil designed so that chip components have enough solder, but not too much so that they are ‘swimming’. Are there ‘shadowed’ joints behind taller or wider components on boards that that are wave soldered. These and many other solderablity issues should be analyzed before starting production, so they do not add to rework cost. A strong DFx software system would be vital.
Testability risk – What is the current coverage of test points to Nets on this board? Do we have enough probe points to ensure we are covering all the Nets on the board? These questions should be answered by the Dfx software system early on, so that design engineers can be consulted an updates to CAD can be made as needed. Even if it is too late to make updates to CAD, this information can be provided to the Test Engineers to ensure they are aware of the potential issues to testability.
Rework risk – Components placed too close run rework risk. A DFx software system running a good Manufacturing Risk Assessment should be configured such that large components that are prone to rework are checked to ensure they have enough access. For example, Chip components or SOICs placed too close to a BGA would hinder rework, or rework at a reasonable rate. If there are many parts to rework, these tight component placements may delay product delivery. If there are component placed under other components or shields, it would be critical to automate the identification of those using the Design data, and know how to plan the production accordingly.
The prevailing thought is; if Assembly Manufacturers implement a DFx software system that runs comprehensive Manufacturing Risk Assessments then we can confidently say “Early Risk Assessment Protects Production”.
DFx in Assembly Manufacturing, why bother? I believe the above provides some answers as to why. Looking forward to hearing your additional comments.