BOSTON As an IPC task group ramps work on the latest revision of IPC-2581, the PCB manufacturing data transfer format, Mentor Graphics is continuing work on its own format, ODB++.
PCD&F spoke this week with Julian Coates, director of business development at Mentor's Valor division, on the company’s latest plans for the format, and on whether it would support the IPC effort as well.
PCDF: What is the status of ODB++?
JC: We have very solid plans to enhance ODB++ and build the value around it even further; our strategy for improved design-to-manufacturing integration is based around ODB++ as the recommended best-practice method for transferring a PCB product-model from design into fabrication, assembly and test. It is a high-level format that carries all product-level intelligence for manufacturing; in other words it is directed toward design/manufacturing exchange, as distinct from the lower-level task of driving the manufacturing machines themselves as was the role of formats like Gerber and Excellon. Mentor’s strategy for ODB++ is consistent with that of Valor over the last 15 years - a lot of investment went into it, in terms of format, tools, support and partnering; Mentor believes it makes most sense for our customer base to continue to build on those investments. Mentor has no intention to implement IPC-2581 currently, but in principle that can change according to what will bring most value to our customers. Our current strategy is entirely built around ODB++, to enhance the use of it, and bring new innovations that express the technologies coming onto the market. We are working continually on upgrading the format and the tools to go with it, both on the design and manufacturing sides. Overall, we plan to leverage the format to bring more value to the end-users. That’s our strategy and it’s entirely built around ODB++. The ODB++ format is open, published, and will remain so as the basis for partnering.
We recently put ODB++ into the context of Mentor’s OpenDoor Program (http://www.mentor.com/company/partner_programs/opendoor/). We’re just getting that kicked off now. OpenDoor is our general partnering program. Even competitors like Zuken, Cadence and DownStream are welcome to partner with us in the context of OpenDoor.
Mentor, like Valor before, owns the format. Whether that means it is proprietary or not depends on your standpoint I guess; I would emphasize that it is open and available to the whole industry; it is not a format that is kept behind closed doors. But, even though Mentor ultimately controls what is written in the format specification, what goes in ODB++ is all determined by what the market requires. It’s been in use since 1995, and all that time we have been listening to our customer-base and making improvements to the format and the tools that use it. All the enhancements are the results of the requirements of our customers. We have customers getting involved in embedded components, embedded actives, and other new types of PCB constructs and manufacturing processes; addressing those developments is what drives the further development of ODB++.
PCDF: Historically, Valor wrote the translators for ODB++ itself. Is that continuing?
JC: Yes, we’re going to continue to write them. For example, we recently issued a new version of the new ODB++ output for Cadence Allegro. In general, our policy is there’s no change to that and we hope to maintain and grow such partnerships with third-parties.
PCD&F: Some have noted that the IPC standard allows all companies an equal say, while for ODB++, Mentor gets the final word. What’s Mentor’s response to that?
JC: Yes, we do hold the key to the format, as did Valor before; nothing has changed there. What we want to emphasize is, we deliberately design ODB++ to be in the best interest of our users: the designers and manufacturers. Mentor has 50% market share of the PCB design market, and Valor and Frontline have something like a 60% share of the front-end tools in PCB manufacturing. When we come at it, we come from a very broad view, representing the needs of a very large segment of the industry. Indeed, some would argue that our approach is more democratic than that of the IPC, if you use the yardstick of direct end-user representation.
So yes, Mentor owns it, but we hold the key on behalf of a very large number of users, measured in the thousands. The IPC also has its customer-base, its membership; in the end it is a matter of which organization can be relied upon by the industry to bring the most value to the most number of users, sustainably.
PCD&F: What is the ODB++ update process? Is it like traditional software development?
JC: ODB++ data are only good data if there are tools in the chain that can use it, the same goes for any such format. You need a combination of good software tools and good data. We enhance the data format in synchronicity with how we intend to bring functions to our tools that will drive value. We listen to all the inputs. In the end, it’s a combination of format, tools and support in order to bring the market a complete solution. It’s not defining a format independent of software tools themselves; that leads nowhere, in terms of value for the end-users.
If ODB++ doesn’t address a software need, it’s because a tool isn’t offered for the user to take advantage. If you take a look at IPC-2581, it’s a very good format, but the challenge is for the software companies to make use of it. Just having the CAD output won’t help you; the data aren’t useful to the CAD vendor; its main point of use and place where the direct value is realized is on the manufacturer side. If the manufacturer doesn’t make the investment in software tools to read it, then the value chain breaks down. I should note that the overwhelming majority of tools-vendors already have very robust industrial-grade ODB++ interfaces that their users and supply chain partners are getting value from For many years already.
PCD&F: Is the XML version still offered?
JC: The XML went end-of-life a few years ago in the days of Valor, after being handed over to the IPC to be used as the basis for IPC-2581. The main format, the 1995 version, is what we are building on and sustaining because that’s the one that has the most use in the industry. We decided to focus our resources on one format – the mainstream ODB++. In terms of bringing more value to the users, it offers the best return for all stakeholders.
PCDF: What is Mentor’s position on IPC-2581?
JC: Indeed, on the face of it, they compete as intelligent formats but Mentor is not looking for a competition here. In your article, IPC-2581 was positioned as developed from GenCAM. In fact, we handed the ODB++ XML-variant format to the committee and, with some further extensions learnt from the GenCAM direction, it became the foundation for IPC-2581. The establishment of a separate format, plus control by the IPC, effectively created a format that stands as a potential alternative to ODB++.
PCD&F: So to clarify, is Mentor planning to participate in the IPC task group?
JC: It’s not as though we’re staying away for doctrinal reasons, but I want to emphasize our strategy is to put our resources where they bring the most value. We handed over the ODB++(X) specification to the IPC; and we see that the process of developing it goes on there. Mentor will be happy to stay in touch with the IPC as we develop our offerings, with a view to ensuring that we follow a strategy of bringing maximum value to the end-users.
PCD&F: At the end of the day, designers just want a format that works. The format represents more of a manufacturing issue.
JC: That’s always been the case. How can the manufacturer convince the designer to produce better data, regardless of the format, when the designer isn’t the one who benefits directly from producing it? That’s something we have been thinking a lot about lately. In the end, the combination of tools, data and support have to translate into advantages of time, cost and quality across the design-to-manufacturing flow.
