Estimated reading time: 7 minutes

Tim's Takeaways: Say ‘No’ to File Hoarding: Data Management Tips

According to the Internal Revenue’s online resources, you should keep your tax returns and associated financial documents for three years. This goes up to seven years if you’ve filed a claim for worthless securities or bad debt, but for most of us, three years is the max. Now, in case you are wondering why I’ve started out with such a dry and boring fact, it’s because it uncomfortably applies more to me than probably any other person in the world. Yes, I admit it, I’m a tax return hoarder.

I’ve tried to tell myself over the years that keeping those documents was an act of preserving history, and from time to time they actually have provided some entertaining moments. For instance, looking back on our incomes from 20 or 30 years ago, or finding out how little we used to pay for health insurance, has made for some interesting jaunts down memory lane. The truth, however, is keeping those old documents was just plain laziness on my part. It was a lot easier to collect those documents every year, bag ‘em up, and toss them in the back of the upstairs closet where I didn’t have to think about them anymore. Unfortunately, my tax hoarding habits have produced some real problems:

• The closet was running out of room
• Critters tend to love piles of paper like this
• My family was getting increasingly annoyed with the mess
• All of that paper had become a potentially dangerous fire threat that kept growing each year

To summarize, I had convinced myself that the massive amount of data that I had collected over the years was essential for one reason or another. In reality, my collection was nothing more than an unmanageable mess that served no useful purpose. Fortunately for me, this particular problem was resolved by a trip to the recycling center and $30 worth of paper shredding services. But there are a lot of different types of “collections” in life that need managing, and like my proliferating pile of paper publications, they all need their own eloquent solutions to keep from getting out of control. Take, for instance, the amount of data that is generated during the design of electronics.

PCB Design Data
The first thing to consider in our world of PCB design is just how much data there is that needs to be managed. From a casual overview it may not seem that extensive, but let’s break the average design down into its four separate pieces. This gives us the schematic, circuit simulation, PCB layout, and analysis, and that is just a generalization. Designs often have more pieces than that in them, especially when you consider the depth of system level design. But starting with the basic four pieces, we can break those up even further:

• Design databases and associated system files
• Library files and sub-library element files
• Read-me, template, and example files
• Reports, standards, specifications, special instructions, etc.

The main four pieces of design data will have their own output or interfacing files. For example, let’s examine the PCB layout data in greater detail. Circuit board manufacturing requires a lot of different design data including:

• Artwork, drill, apertures, and other fabrication files
• Drawings, component XY locations, artwork, and other assembly files
• Netlists, test point XY locations, and other test files

These items are just a sampling as the average design will produce much more data than what I’ve listed here. Of course, all this must be multiplied for design revisions, backups, or alternate part numbers and board configurations. As you can see, the amount of data in a PCB design can easily stack up and create plenty of unexpected problems along the way—just like my collection of tax documents had.

Pitfalls of Data Storage
First, this data requires storage space. Everyone knows that, with all the advances in technology, our storage systems are larger now than they’ve ever been before, which is a great thing. However, what is easy to forget is how our databases and file sizes are also getting larger, and the amount of data that we are saving increases all the time. A friend once warned me, as we were moving someone else’s “stuff” into a larger facility, “People are like fish, they grow to fit the tank.” Ironically, this rather obscure law of nature seems to fit PCB design data as well. As soon as the data storage is upgraded, someone just fills it back up again.

Another problem to contend with when dealing with excessive amounts of data is confusion. It is not unusual for PCB designers or design groups to encounter one or more of the following problems:

• Which design is the most current?
• Incorrect databases being selected for work by designers
• Essential design data stored on personal drives where it is inaccessible to others
• Open file permissions allowing for critical data to be altered, corrupted, or over-written

Interestingly enough, in some cases, just the opposite is happening; not enough data is being saved and managed. Here are some of the problems that we’ve heard reported:

• Original design history and data that is permanently lost
• Missing changes or revisions creating a disconnect between design and manufacturing
• Inability to produce traceability documentation and data for regulated or controlled products
• Having to recreate databases from artwork, or completely redesign a PCB instead of simply rolling a board revision

Obviously, there can be a lot of problems associated with the storage of data generated during the PCB design cycle. The best solution is to incorporate proper management processes to keep all the design data in check; we’ll look more into how to do that next.

Data Management
The first step of good data management is to decide who owns what. This pivotal point can actually end up becoming a sticking point for some design groups, as data ownership gets swatted back and forth between personnel like a shuttlecock in a badminton game. The important thing to remember here is that it isn’t nearly as important who owns the data as it is that the data is actually owned by someone. In other words, if the design group is going to own its own data, that’s fine. If, on the other hand the company’s IT department is going to own the data instead, that’s fine too. As long as someone has the responsibility for ensuring that the design data is available, secure, and part of the regular workflow of the design group, the actual ownership isn’t as important. So don’t let the little detail of ownership derail anyone from setting up proper data management processes.

Speaking of those processes, it is essential to have good data management processes in place. These processes need to clearly outline who is responsible for data ownership, as well as what each individual user’s data responsibilities are. Additionally, file security protocols should be in place, and the processes should specify generation and storage procedures for users to prevent data losses or corruption. The processes should also outline what data should be generated for specific vendors, and what the standard manufacturing file sets should contain. These data management processes should also detail the company’s archival and vault procedures, and where to go with additional questions.

Another important part of effective data management is to ensure that the right data is being created and collected in the first place. As we mentioned earlier, a lot of data can be generated by the PCB design process, and you don’t want to get swamped by every single sub-library or system file in your design directory. On the other hand, you don’t want to casually delete a seemly innocuous file that actually contains some key traceability data in it either. So how can this dilemma be managed efficiently? While it may seem like data hoarding is the right way to go, we also know that it can create its own mess of problems. Thankfully there appears to be an answer to CAD file organization in some new design data automation tools.

Many design teams are turning to database solutions for file generation such as the IPC-2581 file format generators. CAD systems using this format will automatically add the necessary fabrication and assembly files into an IPC-2581 formatted database, making output file generation for your manufacturers quick and simple. As Patrick Davis, a product management director for Cadence Design Systems, said in the September 2021 issue of Design007 Magazine, “I don’t need dumb data. I need something that allows us to actually communicate back and forth efficiently.” With tools like IPC-2581, you can be sure that all the data that is needed for manufacturing is included in one easy-to-access database.

Whether you use design automation like IPC-2581, or you come up with your own data management processes and procedures, the important thing is to keep at it and not be overwhelmed by design data. With the proper data management and a clear delineation of who is responsible for what, you can keep the piles of data under control. Until next time, friends, keep your file systems neat, clean, and organized; and of course, keep on designing.

This column originally appeared in the November 2021 issue of Design007 Magazine.