ECAD / MCAD suggestions?

 Well, first, tools don't fix bad internal design processes.  I've been down that path before and if a company does not believe in a process it will never use a tool to guide it.

Fast forward.... I use OrCAD CIS->Allegro<->Pro E with a very comfortable process and meet the design targets just fine.  But I impose a strict set of rules on the flow that allows the process to work.   There are a LOT of hidden man-hours in my flow and the one in the video.  It's unrealistic that a company could bring on a new set of tools and get the results you see in the video without a lot of work.

Hey thanks for the response Redwire.

I'm really interested in your statement about the results still requiring a lot of work.  Would you care to elaborate?  The video looks really impressive and I'm just trying to get a sense of what 'hidden magic' they may have done to get it to work like that.  Sounds like you have a fair bit of experience with this sort of stuff so I thought I'd ask.

Marty

 I won't elaborate too much...but think about what goes into libraries...

PCB Physical: pcb footprint, thickness, material data (for RoHS etc...), solder footprint, annotations

Mech Physical: part has to be created differently to "see" pins, body, colors, thickness, weight, thermal data

Schematic: part symbol

Database: tie all of these together, check, costs, datasheets, company part number assignments, alternate vendor pn's, power data, thermal data, misc data (we use current, impedance, Q, esr, etc...)

Simulation:  deep black hole...there is logical, thermal, signal integrity, timing....

You still have the backend to consider which (in most companies) is another system (why use the engineering database?) to record the part numbers, boms, vendors, etc... A lot of companies still have issues with data transfer when it's been there for 20+ years but they (purchasing) refuse to integrate their MRP system with engineering's. :)

A lot of companies will offer a service to build you parts.  I know that Cadence does.  I would suggest contacting them (or EMA EDA) as well before you commit to a 'magic show' from someone else.  Try out the tools out of the box.  What do you get?  What will it cost?  What is the advantage to having it be magic (new process, longer design cycle, it works the first time)?  It will ultimately take a deep commitment from the entire company to do it right.  I started my own company seven years ago and clean slated the whole deal. It's working now but the head count is low.  Not sure how to scale it (i.e. make sure everyone commits to the system).  I have started up multiple startups with the tools and process and a year later they've all chucked it and it's become a rogue system with the first-to-eject winning the hearts of the management despite the fodder left behind and then the respin, respin, respin cycle begins... Sigh...

HTH

Thanks again redwire,

I see your point.  But I suppose ANY tool flow is going to require a significant investment in order to set up properly.  So that's probably why I'm really interested to hear any success/failures people have had.  I'd hate to make the investment only to find that it falls over at the end.  I'd also like to make sure that I'm not locking myself into any one vendor that I have to turn to each time I want a new part or mechanical model.  At the very least I'd want to make sure I can import stuff from other sources.  A good example is connectors.  A number of connector manufacturers offer some sort of 3D mcad model of their connector which presumably Pro./E can deal with.  But I want to make absolutely sure that I can link those models in with the PCB flow so the position of components is always in sync between my PCB and MCAD views. I've looked a little into the literature on NEXTRA (http://www.cadence.com/rl/Resources/conference_papers/4.9KrebsTMecadtron.pdf) but I'd like to get some grass roots experiences.  Is that what you use redwire?

Marty

Dr Marty said:

Is that what you use redwire?

No. See my first post.