-
- News
- Books
Featured Books
- design007 Magazine
Latest Issues
Current IssueOpportunities and Challenges
In this issue, our expert contributors discuss the many opportunities and challenges in the PCB design community, and what can be done to grow the numbers of PCB designers—and design instructors.
Embedded Design Techniques
Our expert contributors provide the knowledge this month that designers need to be aware of to make intelligent, educated decisions about embedded design. Many design and manufacturing hurdles can trip up designers who are new to this technology.
Manufacturing Know-how
For this issue, we asked our expert contributors to share their thoughts on the absolute “must-know” aspects of fab, assembly and test that all designers should understand. In the end, we’re all in this together.
- Articles
- Columns
Search Console
- Links
- Events
||| MENU - design007 Magazine
Andy Kowalewski: Time for a Single Data File Format
June 6, 2012 |Estimated reading time: 4 minutes
Hi Andy,
I’d like to respond to Julian Coates’ latest letter to the editor, Let’s Have a Data Transfer Format Discussion.
I’ve been associated with IPC for many years and have lived through a number of well-meaning attempts to establish a global open standard for data transfer. Before my time there was IPC-D-350, starting in 1972 (!) and going through multiple revisions over the years to 1992. Then GenCAD and GenCAM came and went before raising their heads above the parapet.
While this was going on, Valor, with its ODB++ format and Genesis tool, was doing an excellent job of developing software that could hone board fabrication files for board fabricators. With a wide range of design rule checks, viewing and reporting options, ODB++ solved many problems before they got to the factory floor, and was (and of course continues to be) a boon for board fabricators worldwide. A great tool for a great need.
IPC saw the value of embracing the ODB++ format quite a while ago and worked hard to get Valor on board to establish a standard (GenCAM) that was acceptable to industry worldwide, as is IPC’s remit. That nearly worked, until Valor made a business decision that it would allow IPC to use its format, but reserved the right to change it at some future time. In other words, Valor wanted to keep the ODB++ standard to itself. I wasn’t in the high-level discussions, but as I see it Valor wanted to tweak its software for future technology changes (highly commendable of course), while IPC couldn’t accept a format that was to be globally accepted without the usual control of changing it by industry committee. IPC couldn’t have one of its own standards subject to a private company’s whims, no matter how well intentioned.
So the standoff continued, and GenCAM went back on the shelf to gather dust, as it were, waiting for something else to take its place.
We are now in an advanced stage of developing a standard that looks to be gaining traction, at last. This is called IPC-2581, and it comes from a convergence of GenCAM and ODB++ that is being carried out by a NEMI committee. The file format will be based on XML (a reasonably simple but powerful text format) that should be readily adaptable to future technology needs because it is very hierarchical. The file (and it is exactly that, one file) can be all things to all people, and designers, board fabricators and assemblers can take out the bits that suit their needs. I don’t doubt the file will be (maybe already can be) split up easily so that a company can protect its total design and only give each segment of the industry the bit it needs, in a globally suitable form.
There are now IPC-2581 viewers that can read in a file and re-establish a visual presentation of the data. That’s a big step forward in validating the file to make sure the data is good.
My take on all of this is pretty simple. I’m a board designer and I have lived with the Gerber files, drill files, BOM, etc. – a bunch of files that I have to send at the end of each design job. I need to separate the files into sets, some for board fabricators, some for board assemblers, some for stencil manufacturers. Some go overseas, some stay local and all need to have thought put into them as to who gets what, and in what form. Mistakes can happen easily; I’ve been embarrassed more than once to discover the wrong files were sent, or files were missing. The single, standard file format is an absolute blessing, right up there with food, clothing, shelter and quality red wine.
And I don’t think Valor has anything to worry about. Its software continues its role of deciphering a design and checking it for showstoppers before they become expensive rework or scrap on the shop floor, both for board fabrication and assembly, though I don’t know too much about its assembly side. It will (eventually I guess, if not already) quite cheerfully read an IPC-2581 file instead of an ODB++ file and carry on doing its excellent job. And I don’t doubt that the ODB++ format will last for many years because of its current momentum.
Now, the big IF: IPC-2581 needs CAD vendor support to generate a legitimate IPC-2581 file. And here’s the rub: It seems theoretically possible to totally or in large part reverse-engineer a design to translate it into another CAD system. Gulp, shock horror and gee whiz. This has been a stumbling block for many years, and let’s hope it isn’t again. I’d really – and I mean REALLY – hate to see the IPC-2581 effort die too from lack of support by the CAD vendors, who can really kick it along in double quick-time with output utilities.
Oh, this is a big one too: We need a name for this “All Things to All People IPC-2581” file. A name that is simple, catchy, has no double meanings globally, identifiable in many languages, and meaningful. Whoever establishes that will also have my eternal gratitude.
Cheers,
Andy KowalewskiAdvantagePCB Pty Ltd