RW TECHNOLOGY

Design Review

Independent review of hardware designs to identify issues early and reduce cost and risk.

• Concept
  • Rough triage of what stage the design is at vs. the improvement needed to be a good candidate to move ahead with.
    • Initial -- aka 'back of the napkin'

      If 10X faster, cheaper, or better then a good candidate to move forward with.

    • Fully designed

      If 2X faster, cheaper, or better; then a good candidate to move forward with.

    • In production

      Fractional percentage faster, better, cheaper; potentially a candidate to move forward with.

  • Component selection.
    • Critical components

      These are the parts that make your product different. These are the parts your design lives or dies on. Second source is probably not an option. Know them inside and out, be fully connected with the manufacturer, supplier and all else.

    • Everything else

      Have at least a second source and a backup plan for.

  • DFX

    Needs to be at this stage or it isn't really DFX, just a patch.

    • Test

      Which signals need to be brought out? What software will need to be implemented to test them. Can it be part of the production product image, will a special test image be needed? Security considerations?

    • Manufacture

      There are many levels. Board, board assembly, unit assembly. Just passing the layout tools DFM tests is far from DFM. The board assembly house needs to put the board together, the product manufacturer, need to assembly the product.

    • Repair

      It is often said, this is a consumer product, they won't send it in for repair, we won't need repair. High return rates are common though, and it will probably need to be repackaged and confirmed functional. The manufacturer will have yield fallout that will want to be made saleable.

    • Security

      Do you have locked images or hardware. Which I/Os need to be locked. (Just depopulating connectors isn't significant security.) When during manufacture will they be locked? Will Repair need a repair image? Per design locking, or per unit locking?

  • Cost down

    First you'll just want to make the design work, but keeping an eye to future cost down, may save a major redesign schedule.

• Schematic

  Are the schematic components too specific to a particular part. These are what the BOM will be generated from. Do you need to be stuck with these 0.1% parts, when a 10% would have done?

  Documentation for manufacture and repair?

• Layout
  • PDN

    Are the power planes too chewed up?

  • SI

    Do high speed signals cross planes? Are there sufficient return current paths, do they push crosstalk onto other signals? Do the signals simulate okay? Good looking signals with crisp transitions will likely be a problem for EMI and crosstalk.

  • EMI

    Sufficient decoupling, where are the return current paths running. 3D EM simulation.

  • Thermal

    Useful data is usually not available until you create it for your design operation. The data that is available typically suggests everything will be operating at sputtering temperatures.

• Cost Down

  As noted above this should have been part of the design concept, but life happens and now you need to figure out the best options remaining to you.

• DFX

  As with many things this should have been part of the design concept, but life happens and now you are here. What are the options? Can a redesign be avoided, or at least minimized?

  • Design for Test
  • Design for Manufacture
  • Design for Repair