Design for Manufacturing 101: What Is DFM, and How It Changes With Your Mix (Part 1 of 3)
The Integrate Intelligently Podcast — with Jeff Brickler and Scott Brickler (CADTALK). Design for Manufacturing, Part 1 of 3.
Jeff: Welcome back to The Integrate Intelligently Podcast. I'm your host, Jeff Brickler, and after a brief pause Scott and I are back. Scott is the founder and CEO of CADTALK Software. He and I are brothers, and we've been working together here for ten years.
It's the middle of summer as we're recording. How have things been?
Scott: Great. The weather is finally cooperating and we've had some nice sunny days. Hot, but sunny.
Jeff: My summer has been a bit chaotic — quite a bit of travel. I was gone a couple of weeks in June, and it looks like I'll be gone a couple of weeks in July. Glad to be back recording.
Today we're talking about DFM — what it is and what we know about it. Think of this as DFM 101. Another acronym, of course. Scott, define it for us.
What DFM Means
Scott: DFM stands for design for manufacturing. It's the act of making sure your design is conducive to how you're going to make it. You take the manufacturing process into account during design.
Not that engineers don't typically do that, but some companies are better at it than others. You usually see it where there's high production and more volume, because if you're making hundreds or tens of thousands of parts, every small improvement in the manufacturing process scales well.
You also see it as you scale down, because you're trying to reduce cost, deal with material availability, and reduce lead time. So there are benefits at low production too, but it shows up most in high production.
A lot of it is figuring out how to optimize for manufacturing during the design stage. In design, the first thing you think about is how to make this thing work. Then you think about how to make it easy to make. That second part is the DFM.
Jeff: I remember when we were kids out in the garage with Dad, working on the car, and he'd always curse the engineers who designed the vehicle. Why did they put this bolt here?
To your point, in those high production environments they probably were optimizing for manufacturing. They weren't optimizing for you and me lying on the ground in a puddle of grease trying to reach a bolt on a transmission.
There are the three design principles — form, fit and function. Does it have the right form, does it fit, does it function? And then maybe another avenue is manufacturability. The design may make sense for all of those things and still not make sense for the aftermarket person working on it.
Scott: They're not designing for maintenance, they're designing for manufacturability. A good example: when you're putting an engine in on an assembly line, it's probably suspended by a jack that's easy to step under and get around. The conditions in the factory during manufacturing aren't the conditions during maintenance.
Engineering Is a Game of Optimizations
Scott: That brings up a good point about engineering in general. Everything from an engineering perspective is a game of optimizations. You're constantly figuring out what you're optimizing for.
I tell our engineers this. They ask what they should do here, and I say it depends what you're optimizing for, because it changes your decision tree on everything. Are you trying to make it cheaper? Faster? Easier to change the code or the design?
Typically you can only optimize effectively for a few things, and there are always trade-offs. In code, faster code tends to be harder to read, so you have conflicting optimizations. Same with physical things — sometimes you can make something cheaper to manufacture but harder to work on.
So DFM is optimizing on the constraint of how we're making it, which could make it more expensive from a material standpoint, or harder to work on, like your maintenance example. When you optimize for one thing, something else usually gets worse.
Jeff: Firsthand experience there. I remember working on the transmission on your F-250, and it was definitely not optimized to be worked on in a garage, in the winter, on the ground.
So engineering is about optimization, and designing for manufacturability is one optimization an engineer has to keep in mind.
Scott: And you're balancing it. From a manufacturing company's perspective, designing it to work is table stakes — it has to do the job. Then, as a company, what do you care about most? If you're providing value to shareholders in a large organization, you optimize for competitive advantage and lower cost. Cost comes through material cost and through manufacturability: the easier it is to manufacture, the less labor goes into the process. So they optimize for that, at the expense of the other side.
Different companies have different philosophies about what to optimize for. Mercedes-Benz optimizes heavily on maintenance. I remember Dad telling us there used to be a little tool kit that came with a Mercedes or a BMW — I can't remember which — and you could take the entire car apart with it. It was designed for maintenance.
But generally, people optimize for what they're trying to do: make a product at a price point the market will buy, with a decent lead time. Optimizing for manufacturability serves that well.
DFM When Every Product Is Different
Jeff: I understand DFM for products you make all the time. You optimize the material choice — instead of plywood you might use oriented strand board or fiberboard for cabinets. Is it as durable? No. But it's cheaper, you can make more of them, you keep material costs down, and more people can have them.
That makes sense in mass production. But what about project-based manufacturing, or configure-to-order, where you have low volume and high mix? Lots of different choices, and different products that are slightly different each time. How does DFM work there, and how would engineers design for it?
Scott: It's usually about standardization. You're optimizing around standardization to reduce lead time. You have variability in the end product, but consistency in the raw materials.
At the engineer-to-order packaging machine company we've talked about before, one of the things we did was standardize on certain material thicknesses for fabrication. We said we're only going to make the machine out of seven gauge stainless and three-eighths stainless.
At a part-by-part level, some of those pieces were probably over-engineered. But it gave us consistency, because we knew that the fewer material switches on the laser, the higher the laser utilization and the faster parts got through laser operations. That improved throughput, increased the likelihood that all the parts arrived together for assembly, and brought lead time down. We optimized for lead time reduction.
There's a book about this called Quick Response Manufacturing, which argues that in high mix, low volume you do well optimizing on lead time reduction. And you can design for that by driving consistency into the process — using the same hardware, two or three sizes of nuts and bolts instead of many, the same raw materials so you can fabricate them together.
So you drive consistency earlier in the process, because you don't have consistency in the product itself. The parts may change, but how you make them is very consistent.
Jeff: You don't say it as much anymore, but before I even joined CADTALK you used to use this rhyming phrase: consistency drives efficiency.
Scott: It's true.
Jeff: And that's what DFM is — driving consistency where you can, and getting that information to engineers during design so they make better choices. Better material choices, because material type also affects how you route parts through the shop, and which lasers run better than others, and how long those run times are. Even with high mix, you can drive efficiency through the materials and machines you have.
Scott: In mass production you can push more cost basis into your design for manufacturing, because you're making so many that any one part can be optimized for batching. You're making a thousand crankshafts, so you drive the cost of that crankshaft down as far as possible.
If you don't have the volume and you have more mix, you won't get the consistency you want at the part level. So you want consistency at the raw material stage. If I'm making five different parts but they all come out of the same bar stock, I can put the bar stock on the saw, cut all the different parts from it, and reduce setup.
You're still optimizing, but earlier in the build cycle. So you have to know your mix and what you're making. The DFM principles change based on those variables.
Next Time
Jeff: This is a good introduction — it was meant as a primer. A preview of our next episode: we'll dive deeper into design for manufacturing, beyond what it is and into how to use it. What if you don't have a manufacturing engineer? What if you have other challenges? What does ideal look like, and how might you implement this in your own company?
Thanks for joining today. If you want to download our white paper, look in the show notes. Subscribe to the CADTALK Integrate Intelligently Podcast wherever you get podcasts — YouTube, Spotify, Apple Podcasts. Follow or connect with Scott and me on LinkedIn: Scott Brickler at CADTALK, and Jeff Brickler at CADTALK. Be warned, if you look for Jeff Brickler you might find more than one — we have a couple in the family. You want the one at CADTALK.
Anything else before we close?
Scott: Great topic. Looking forward to talking about it more in the next episodes.
