Issue link: https://iconnect007.uberflip.com/i/378511
20 The PCB Design Magazine • September 2014 feature is why I propose a path that is far less expensive and risky than that advocated by Coates. The first step along this path was to clarify areas in the Gerber format specification that were sometimes misunderstood, and to remove elements in it that were outdated, rarely used, or superfluous. This has been carried out in re- cent years, so the current spec is clear, sharp and to the point—there are no useless bells and whistles in the Gerber format. The second step, completed earlier this year, was to introduce the Second Extension, or Ger- ber X2 format. Gerber X2 contains attributes that specify how the layers stack up, identifies via pads, indicates where the impedance con- trolled tracks are, and describes a host of other parameters that support the image data. With X2, what was missing in Gerber has now been added—in Coates's terminology, the attributes add intelligence to the format. The neat thing is that they do not affect the image, which means that existing workflows are not broken: X2 requires only minimal changes in working practices, and certainly none that would require approval, testing and all the rest. The fully X2- compatible CAD and CAM software will read entire Gerber X2 archives automatically, with all layers in place, while identifying the func- tion of each object. And even in combination with older software that does not support X2, the correct image is still produced. This means that even if users do not reap the full benefits of Gerber X2, they can happily move within the X2 world without problems. Ben Jordan of Al- tium concurs: "ODB++ is a good standard, but Gerber X2 does solve the problems while being backwardly compatible." More importantly, this means that no- body is forced to buy anything, and Gerber us- ers can decide in their own time if, how and when to adopt new X2-ready software to take their processes to the next level. For those inter- ested, there is a sample X2 job on the Ucamco download page. It shows the simplicity of the concept. Download it and try it on your own Gerber input software—in all probability you will be able to read in the images correctly, but your software will throw some warnings. This demonstrates the compatibility of X2 with non- supporting software. That's good. My impression was that Coates saw Gerber as an intrinsically error-prone image format whereas I maintain there are very few er- rors when transferring images in Gerber. So we both agree that Gerber is a very fine image for- mat. Where our opinions diverge is in how we proceed from this fact. Coates went on to state: "At a recent industry debate, I suggested that the best way forward is to use Gerber for the graphical data and another format for all the other information that Gerber cannot carry." Thus, he promotes the idea of intelligent, all-encompassing formats for carrying data, but excluding the graphical part. Why reject the ad- vantage of having all of that other information linked to the graphical objects as well, and vice- versa? The problem that needs solving is taking all of that fragmented data into a single coher- ent model comprising both the PCB bareboard and the assembled PCB. Actually, in no way do I reject the idea of linking all the other information to the graph- ics objects. On the contrary: It's clear that a PCB is more than a set of images, and all the data describing it must be transferred as a coherent whole. Here, too, we agree. Where we disagree is how we achieve this coherent whole. Coates believes that the wholesale adoption of ODB++ is a practical way forward. I do not. In another passage from his July rebuttal, Coates correctly analyses why ODB++ is not more widely used: "What limits the implementation of ODB++ more widely? … I would suggest that the reason is not technological; it is a combination of busi- ness and human factors. Firstly, it costs money to change a business process; tools have to be upgraded. […] Secondly, there is a perception that continuing to use the old method is not only free but also safe, whereas to use the new method is expensive and uncertain. The safe versus uncertain part is the human part." These are entirely rational and justified concerns, and clearly the vast majority of this industry feels that they outweigh the benefits of ODB++ (or of any new format that has been tried over the decades for that matter). Who am I to judge that the whole industry is wrong? That said, our industry must move on, and like Coates, I too would like to see CAD to CAM data transfer advance beyond current practices. This THE GREAT GERBER VS. ODB++ DEBATE continues