Well, although your questions are directed more towards the manufacturing community, I consider myself to be in a position to address these issues as a member of the design community. (I design circuit boards for Caterpillar)
During the design process, I provide a BOM to the planner who will be searching availability AND ALSO obsolescence (we have had more problems with obsolete part numbers than we ever expected, things are changing FAST, and we need to have products that last 15-20 YEARS), and in the meantime I am evaluating the engineer's component choices for the goal of combining reels and and helping him select "preferred" parts over more costly alternatives (not just price). When the component placement is well underway, I send data downstairs for preliminary DFM reviews while we are routing traces. Panelization and material yield is discussed before the first prototype is ever ordered. And of course we have a checklist that reminds us of many things we have learned (in other words, screwed up in the past) to prevent idiotic repeats. We also have a way for anyone to record comments and suggestions for future revisions.
As for test, we give them the data and have no idea what they do with it. Test fixture development is still a "black art" to me.
To summarize, your question "Is the manufacturing group being heard?" At least in our case, the answer is:
I would be interested to hear how successful the manufacturing community is at being able to give feedback on future designs at the different stages of the evolution of a PCB design from Bill of Materials, to schematic to final layout.
Do you only get to review the PCB layout once it is completed?
Do you get a chance to review Bill Of Materials for potential part issues and the usage of those parts in assembly?
Do test engineers provide input on test strategies during schematic or layout?
Is Design For Assembly analysis performed after component placement in layout or after routing?
Thanks for your input Jack. I'm glad that you are working with your manufacturing people to improve product quality.
You mention a number of times in your post about the data being distributed to various manufacturing groups. I would appreciate if you could elaborate on the types of data that you send them, whether different groups request different data sources and how parties communicate changes back to the design group.
mark_laing wrote: > > Thanks for your input Jack. I'm glad that you are working with your manufacturing people to improve product quality. > > You mention a number of times in your post about the data being distributed to various manufacturing groups. I would appreciate if you could elaborate on the types of data that you send them, whether different groups request different data sources and how parties communicate changes back to the design group. > > > Regards. >
Well, we probably don't have the most sophisticated process in place, but I'll tell you what we do:
As soon as the schematic is packaged we send the Mentor ASCII output Bill of Materials back to the engineer, who is supposed to forward it to the project manager to start looking for long lead-time buys. Sometimes component changes are made right then for parts that are more easily available. We are also starting to do an obsolescence study at this point because we have had trouble keeping parts in stock. Also at this point I make sure if I see goofy things (like a few through-hole parts on a "mostly SMT" design, or a few 0603s on and design that is mostly 0805s) that i give a heads-up to whoever entered the schematic. Quite often these are just oversights where the wrong part was selected, and these can be corrected before placement really starts. We are hoping that DxDesigner databook will help engineers do a better job of part selection with the parametric searching capability (transition process will start eventually,still using DA/BS/RE on UNIX)
When the placement is done we initiate a "preliminary DFM" where we give our assembly partner a chance to look at what's coming down the line, at this point it isn't much more than a picture of the board (no gerbers or anything) and if necessary we discuss the best panelization for yield and assembly processing steps. I guess I should have mentioned that before this phase I may have contected them to decide if there is so much Through-Hole stuff we need to wave solder. but we try not to, we have selective soldering machines if the TH parts are minimal.
When the design is done we send the entire package to the same people as the paragraph above for the formal DFM, and they have a fairly formal checklist to make sure they have tooling holes, adequate rails, fiducials, etc. but honestly, since we know pretty much what they are looking for we already have it taken care of by the time they see it. The times when this process BREAKS DOWN is when we revise an older design where these things weren't taken into account, and the assemblers want to make changes, where we quite often resist because we don't want to change a working, proven product too much. mostly fear based (grin)
Then we have a extensive check by the CAM guys at our manufacturing partner, and working with them is what inspired the article I posted a few weeks ago, so I don't want to repeat all that [http://frontdoor.biz/PCBportal/CAD2CAM.pdf]
Since this is a Mentor site I can't resist the urge to complain that Mentor never bothered to keep up with the times and offer a built-in netlist generator for IPC-D-356 format, like THE REST OF THE WORLD! (I know Expedition has it, but not BS/RE)
Finally, we are being asked to make a better package for the ICT test guys (which is not a formal process yet) but they mostly use our Mentor Neutral File to extract what they want, but they also ask for schematics as a PDF file in addition to the data we already have in the fab/assy packages. This is a "black-hole" department for me, because they just take the data and do whatever they do, I never hear any complaints about test coverage or percentages or suggested improvements even though I know I don't do a very good job of DFT.
As far as your question about how changes are communicated back? We don't hold as many formal meetings as we used to, I'm fine with email. We also have a place where any suggestions can be recorded for future revisions (add an extra test point here, move this away from that, etc) but it is not often used, unfortunately
hope that's what you meant,
Thanks for your further input Jack.
As I used to work for some test companies, Marconi Instruments, GenRad and then Teradyne, the interface between design and test engineering piqued my interest. I have always heard that complete schematic and layout data analysis helps test engineering to make sure that their feedback is being incorporated in to a design as part of the design process. Then once a design is complete, both schematic and layout information in electronic format, helps enormously for prototype runs and program debug. Then once in volume manufacturing, repair activities and the mean time to repair a defect can be shortened with access to intelligent data.
Maybe you could solicit some input from the test engineers you work with. The Mentor Neutral file is a reasonable choice but it does not provide data on the traces that are particularly important to convey the routing on the board for test engineering.
<!--[if gte mso 9]><xml> Normal 0 false false false MicrosoftInternetExplorer4 </xml><![endif]
><![if gte mso 9]>
Thank you Mark for bringing this up.
For many years I was on the manufacturing side of the fence
and now I’m BSD’s DMS Product Line Manager.
For too many years the design world seldom communicated with
the manufacturing world. Kind of the “white collar” guys don’t meet the “blue collar”
guys. They belong to different classes. However
a number of things that are still in WIP encourage and many time force these 2
groups to collaborate very intensively , as earl as possible and during the entire
process and product life cycle.
1st thing in my opinion is the economy. The pressure on price, competition,
and globalization is forcing the manufacturers to be “lean and mean”. Time
to market and cost reduction becomes a survivability task – you do it or
major catalyst is regulatory compliance. The cost of non-compliance to environment
regulation is high not only in money but also in public relations. The
users will distant themselves from makers that are not following the
regulation and the authorities will fine them heavily. The manufacturers
have reputation to protect. The compliance requires that the makers document
and be responsible that their product “as designed” is exactly “as build”
this calls for full traceability for all materials and processes used in building
an electronic product. And by-the-way, the brand name owner (the OEM) is
the legally responsible one in case of non-compliance and they can’t pass
the blame to their contractors.
3rd is the improvement in automation (software) that can much easily
bring these 2 world closer and collaborate
even in real-time when necessary. It is important that designer take into
account manufacturing constrains and processes as part of their design
practices using a FDM (Design For Manufacturing) tool will check and warn
if there is a violation of these manufacturing constrains. For test the is
DFT (Design for Test) that will help designer to comply with the test
procedures that are planed for the product – like designing test point –
allow enough space for test probes in areas test probes to reach the
traces or components pads. There are hundreds of rules and constrains that
need to be addressed and using automation is the only way to go.
So, we can see more and more traction between these groups,
yet not enough. These vertical “design <-> manufacturing” collaboration will
be strengthen more and more in the coming years as those 3 factors I mentions
will become even more important in the challenging times a head.