Part Numbering Systems
Just about every company I’ve seen struggles with something that seems basic and straightforward on the surface, but in actuality is incredibly complicated: how to number their parts.
In general, you have a choice between two kinds of numbering systems: smart and dumb. Let’s take a look at the pros and cons.
Smart numbering systems are built around a logic in which certain numbers provide information about the nature of the component. For example, the prefix could indicate what kind of component it is - structural, electronic, piping, cabling, COTS (Commercial Off The Shelf), etc. Or a string of numbers could indicate the component’s location within the product, or the product to which it belongs.
These smart numbers have their allure in that once you hear a part number, you have a sense of what it is, what it does, and/or where it is located in your product.
The problem is, all smart numbering systems break down at some point. If you try to separate components by function, you’ll find that there can be overlap, and some people will choose the wrong number. At one company, our part numbers differentiated between Avionics and Propulsion, but what about Avionics that drive a Propulsion system? Similarly, we had different numbers for Structure, but what about structural components that house or support avionics or propulsion? If you give the wrong number to a component, do you re-number it? Because after a component has been released or incorporated into a released product, that can be a headache.
If your part number indicates what product uses that component, what happens if you re-use that component on a subsequent product? If you leave the number the same, that can cause confusion. Do you change the number? Or do you give it a new number? If the latter, then you have identical parts with different numbers, so you have drawings and manufacturing process plans that are the same, and you probably want to designate the components as interchangeable (alternates or substitutes) in your PLM system. But they will be segregated in your inventory system…
If the part number indicates location, what happens when you want to use the component in multiple locations? (This is another very common problem, and why different instances of the same component in an assembly can have different reference designators.)
For these reasons, a lot of product development organizations opt for dumb numbers. If using a Product Lifecycle Management (PLM) system like Windchill, the Object Initialization Rules (OIR) automatically serve the user the Next Available Number (NAN - a lot of acronyms, I know) when a CAD document is created. If you’re not using a PLM system, maybe you’ll have some kind of shared document to keep track of part numbers, or perhaps someone will assign them.
There’s something safe about a part number that doesn’t mean anything. It forces people to look at the object’s name. And that can be its own disadvantage. Some people don’t bother giving their parts and assemblies a name, so it’s the same as the number.
Just like your company should have a numbering convention, it should also have a naming convention (usually nouns followed by adjectives going from general to specific - or vice versa). This is another system to develop, and enforce; far too often, this doesn’t happen.
In cases where parts haven’t been given good names, people have to pull open the object in either the original authoring application or a lightweight viewer in order to figure out what it is. And although engineers and designers working on a product develop a familiarity with its components and their part numbers, this can be a serious headache downstream people such as manufacturing, procurement, and inventory.
However, one common way in which dumb numbers take on meaning is when people reserve a block of numbers for their product or project.
One word of advice: figure out how many digits you think you need. Then add at least one. At one place, we came up with six digits, thinking a million numbers would last us forever. After a year, we looked at how many numbers we had burned including reserved blocks, and quickly extrapolated given expected growth and the eventual number of products, we might run through our six digits in 12 to 15 years. That seems like a long time, but let’s put that in perspective. At the time of this writing, 12 years ago:
The top movies of the year included X-Men: The Last Stand, Casino Royale, and Borat.
The top television shows included American Idol, Dancing with the Stars, and Grey’s Anatomy - all of which are still on the air.
Now a word about dash numbers (the suffixes at the end of a part number). Personally, I hate them. Dash numbers were intended to save people from creating additional drawings that are essentially the same as another one. Perfect examples of the use of dash numbers are components that can be made out of different materials or come in different colors.
Dash numbers can also be used to document a family of components on the same drawing. For example, a drawing view can call out a dimension with a name like “LENGTH” instead of a value. Then a drawing lists the value of LENGTH for different dash numbers.
The problem with dash numbers is that they end up being used for a variety of reasons, resulting in ambiguity. Dash numbers can also be used to represent:
Product variations (e.g., material, color) as previously mentioned.
Component families as previously mentioned, especially fasteners.
The left-hand and right-hand versions of a part. This is very common in aerospace, where components on either side of the vehicle are the mirror images of each other.
Composite parts. In some modelers for CAD systems, you can model the individual plies of a composite lay-up, and you may end up giving each ply its own dash number.
An assembly and its components. For example, the assembly might be the -001 and the components it contains are -101, -102, -103, and so on. Or vice versa. I see this a lot with electronics boxes, and also with structures that contain inserts that are consumed and inseparable when assembled. For example, a panel can be the -001, and when threaded inserts are added, the panel now becomes the -101.
Revisions of a component. The first iteration of a product is the -001, and the next is the -101, and the one after that the -201, and so on. This helps people identify a series of products as being related.
That’s a half dozen different ways in which I’ve seen dash numbers get used, and there are probably more. And therein lies the problem with dash numbers: you can’t tell from looking at one what it means.
All I’ve really done here is identify why part numbers can be such a headache. However, the best advice I can give is to come up with a system that you think works for you… and then be prepared to scrap it. It will probably take you at least three tries to find something that works, but it will probably still have holes and not accommodate the edge cases. But if you have something that works well 95% of the time, you should be happy.