Minutes of Weekly Meeting, 2012-01-09

Meeting called to order: 11:06 AM EST

1. Roll Call

Brian Erickson
Carl Walker
Ian McIntosh
Brad Van Treuren
Richard Foster
Peter Horwood
Heiko Ehrenberg (joined 11:08)
Tim Pender (joined 11:15)
Harrison Miles (joined 11:22)

Adam Ley
Patrick Au

2. Review and approve previous minutes:

12/05/2011 minutes:

  • Draft circulated on 12/06/2011.
  • No corrections advised.
  • Brian moved to approve, seconded by Carl. No objections or abstentions.

12/12/2011 minutes:

  • Draft circulated on 12/12/2011.
  • No corrections advised.
  • Insufficient attendees for approval.

12/19/2011 minutes:

  • Draft circulated on 12/31/2011.
  • No corrections advised.
  • Carl moved to approve, seconded by Brian. No objections or abstentions.

3. Review old action items

  • Adam proposed we cover the following at the next meeting:
    • Establish consensus on goals and constraints
    • What are we trying to achieve?
    • What restrictions are we faced with?
      {Action to be retained}
  • All: do we feel SJTAG is requiring a new test language to obtain the information needed for diagnostics or is STAPL/SVF sufficient? see also Gunnar's presentation, in particular the new information he'd be looking for in a test language
    {Action to be retained}
  • Ian/Brad: Condense gateway comments and queries into a concise set of questions. - Ongoing
    {Brad moved to close this action}
  • All: Forward text file to Ian containing keywords from review of meeting minutes. - Ongoing.
    {Brad moved to close this action}
  • Carl/Brad: Get annotated keyword worksheets to Ian by Wednesday Close of Business. - Ongoing
    {Brad moved to close this action}
  • All: Consider how a keyword can be used to define the chain configuration for a given test step, and what that keyword might be.
    {Brian moved to close this action}
  • Harrison: Prepare slide showing matrix of industry sectors by volume/mix. - COMPLETE.

Long standing actions were reviewed for continued relevance. Decisions are noted above.

It was noted that in future it may be helpful to record the date new actions are created.

4. Discussion Topics

  1. Refresher on discussions from December 2011
    • {Deferred to another date}
  2. Extending the Primitives descriptions to address 2-wire topologies
    • {Deferred to another date}
  3. Harrison's industry sector matrix
    • [Ian] I think if Harrison has his slides available, we can maybe look at those today.
    • [Harrison] OK, let me find them. I'm just trying to remember all the context behind these.
    • {Harrison shared slides}
    • [Harrison] In here, there's the notion of New Product Introduction and the notion of Qualified Production. The data here was put together by people working in the industry, so it's first hand, which is not normally the case with this type of study. The one sector that isn't represented here is that there's no DoD.
    • [Harrison] The key things to note are the number of weeks to NPI and the number of weeks to Qualified Production. Automotive had the longest period to NPI, but that was reduced by 29% which is a massive change. Consumer/ Portable sector, and even Medical, had fairly aggressive times for NPI and Qualified Production.
    • [Heiko] Is Medical really the same as Consumer here?
    • [Harrison] Let me check the source data in case I got something wrong... No, that seems to be correct. They both have the same figures.
    • [Tim] NPI maybe doesn't include the design time.
    • [Harrison] Ah, and in Medical, the certification and FDA approvals; that can take a lot of time.
    • [Tim] Im also thinking that in Medical, they maybe don't have the same space restrictions, so may have bigger boards, maybe more warpage. [Harrison] OK, but this is maybe more the portable medical equipment: I know there's a $10M business there. Some wireless, and all non-wireless portable equipment. It doesn't include MRIs, etc.
    • [Tim] Things like portable ultrasound?
    • [Harrison] Yes.
    • [Tim] X-Ray, MRI?
    • [Harrison] Not considering those.
    • [Harrison] For the Consumer and Portable sector, I think this data represents the reality check that's going on for them.
    • [Harrison] This second slide has the value mix and the segments that are typically in each:
      • High Value, Low Mix : Automotive, EMS
      • High Value, High Mix : Silicon vendors, Consumer/Portable, Medical
      • Low Value, Low Mix : Small-CM(s) (regional, provincial)
      • Low Value, High Mix : Aerospace/DoD, CM(s)
      • Medium Value, Low Mix : Netcom
      • Medium Value, High Mix : Office/Large business systems
    • [Tim] Why would silicon vendors be here?
    • [Harrison] Well, they have to produce silicon in high volume *and* produce the reference design. First design invariably use the silicon vendor's reference design. If it's successful then they go into what's called 'value engineering', where they may use parts other than those recommended.
    • [Ian] I'd generally agree with the table that for the Mil/Aero sector, we're probably in the Low Volume, High Mix group: We'll often build several variants of a product, and maybe only two production units for some of those.
    • [Ian] But, if you're looking at boards rather than systems, we might have boards that are common across several products. And in radar, with the newer electronically scanning antennae, we may have 1000 or more identical daughter boards in the 'array' on each radar. Over a year's production of radars that gets into relatively high volumes, and a cent or two of cost per unit that normally wouldn't normally bother us starts to become much more significant.
    • [Harrison] That's an important point you raise: Times are changing.
    • {Slide of graphs shared}
    • [Harrison] At the top right, the chart shows the increase in pin count of devices. The number of pins is becoming startling even for hand held products.
    • [Harrison] At the lower left, we can see the number of solder joints in a board, with each solder joint being a possible defect opportunity.
    • [Ian] These are interesting, and there's couple things that come to mind. I know some of the other folks here use ASICs extensively, but we tend to use FPGAs more. With the more capable FPGAs, the pin count is going up but we rarely need all the pins: There's a lot of parallel IO we don't use, because we're using serial IO. On a 1200 pin device we might 400 or more unused pins.
    • [Ian] Then, if you get a solder defect on an unused pin, is it really a defect if it's having no real effect? The contract manufacturer sees the defect and wants to rework it, but is it worth the additional stress/life on the board?
    • [Harrison] Many fabless silicon vendors also use FPGAs for their initial development.
    • [Tim] The Altera tools may use spare pins to get additional grounds into the device.
    • [Tim] You get into the issue of how you decide if the solder defect is significant or not; that's hard for the CM.
    • [Ian] Agreed. The CM simply sees that the board isn't to drawing, and feels obliged to rework it.

5. Key Takeaway for today's meeting

  • [Ian] New actions should have the creation date recorded to aid identifying the context at a later date.

6. Schedule next meeting

Next Meeting:
16th January.

7. Any other business

  • The Newsletter is scheduled to be issued at the end of January.
  • The reaffirmation of officers should be conducted in next few weeks.

8. Review new action items


9. Adjourn

Brad moved to adjourn at 12:14 PM EST, seconded by Peter.

Respectfully submitted,
Ian McIntosh