Recently a client approached me about building a welded stainless steel structure. The client want this project looked attractive even after long-term exposure to the elements. Below, I’ll use a series of images to show how we solved one of the more challenging aspects of optimizing the design. Solutions like this are usually an evolution from an expensive, time-consuming method to a much more cost- and time-effective method.
Here was the challenge. The design calls for a bent section of 4” OD stainless steel pipe:
to which we need to mount two lighting fixtures like these:
Here’s the mounting bracket the fixture manufacturer supplies: 
And here’s where this bracket needs to go in relation to the pipe, so that the lights can shine downward: 
It would be easy and fast if we could simply weld the manufacturer’s bracket to the pipe, but that’s not a good idea for a couple reasons:
- The manufacturer’s bracket is aluminum, and the pipe is stainless steel. The two won’t weld to one another by any of the usual methods.
- We’d like to be able to quickly remove the fixture for maintenance or replacement in the future.
Iteration #1
Instead of using the manufacturer’s bracket by itself, let’s make our own bracket out of 3/16″ stainless steel sheet:
The simplest way to use our bracket would be to attach it directly to the pipe using a couple of small welds.
Here’s what that would look like (minus the welds):
Simple, cheap, and quick, but not visually seamless enough, and not especially strong or stiff.
Iteration #2
Let’s make a piece to bridge the gaps between the pipe and our bracket, by machining a short length of stainless steel pipe into this shape: 
This shape vaguely resembles a saddle in the way it straddles the pipe, so let’s call it that. Let’s flip it over and weld in our bracket:
Here’s what that looks like with the manufacturer’s bracket for perspective:
Here’s what it looks like on the pipe:
Much more seamless, as well as both stronger and stiffer. But machining the saddle is less than ideal in a couple ways.
- Due to the complexity of the shape, it has to be machined on a computer-controlled CNC milling machine. These are becoming more common, but they’re not cheap.
- It uses more material than necessary. To make the machining process work well, we’ll need to use pipe with at least 1/4″ wall thickness, which is more than is necessary, especially when stainless costs roughly three times as much as regular steel.
- Communicating our needs to our contractor is relatively cumbersome, again because of the complexity of the shape of the part. We’d have to send a 3D file rather than a simple 2D drawing. Some manufacturers are better than others at successfully turning 3D files into machined parts.
Iteration #3
Instead of machining the saddle, we can make it out of sheet metal. To do that, we’d take a piece of 1/8″ stainless sheet and cut out this funky shape, using plasma, laser, or waterjet cutting:
Then we’d roll it into a circle and weld the seam:
This accomplishes much the same thing as the machined version, with several benefits:
- It’s cheaper. In my experience, the machinery and know-how needed to produce this sheet metal version are more common (and therefore cheaper) than that needed for the machined version. Sheet is also a bit cheaper than pipe, pound for pound.
- It’s more foolproof. To communicate our needs to a contractor, all we need to do is specify a material, a thickness, a 2D shape, and rolling instructions–simpler and more error-resistant than specifying a complex 3D shape.
Great, but one lingering problem remains. The saddle and our bracket are a little tricky to line up, so we could easily end up with this:
…instead of this, which is what we want:
To answer that question, I submitted drawings of these parts to Bigbluesaw.com, a waterjet cutting service that gives quotes instantly via their website. Here’s the quotes I got:
So unless we need more than 4 of these assemblies, castellation adds zero cost. At some quantities, adding castellation actually reduces cost slightly, probably because it uses slightly less of the thicker material.
What we’ve learned
It’s rare to get something for nothing, but that’s basically what we’ve done here. By moving from machining to sheet metal, we lowered costs without negatively affecting function or aesthetics. By adding castellation, we reduced the risk of expensive, time-consuming errors without increasing costs.
How’d we do that? It starts by having an intuitive knowledge of manufacturing processes, so we can make good, fast decisions about which process is most resource-efficient.
Once we’ve identified a process, we used 3D CAD tools to model our solution, allowing us to streamline our workflow. Using 3D CAD opens up options that might not otherwise be viable, such as, in this case, sheet metal; generating the flat patterns for our saddle would’ve been prohibitively time-consuming to do by hand, but was relatively quick and easy using 3D CAD.
A student research group competing in Solar Decathlon, the Dept. of Energy’s biennial solar home competition, was working on an innovative way to heat domestic hot water. Rather than simply allowing tubes full of water to be heated by the sun, the team wanted to see if simple fiberglass troughs could be used to concentrate sunlight.
To make this work, it was essential that all the troughs be positioned at the same angle, and held rigidly in a frame. The frame had to be lightweight, weatherproof, and cost-effective; one of the judging categories in the contest is affordability. The team was also on a tight schedule.
Working in close collaboration with student researchers, BICEP staff designed an aluminum frame using state-of-the art 3D modeling tools, allowing the students to see exactly what the system would look like when assembled. Because time was limited, the design was carefully optimized to minimize completion time. To reduce the possibility of error, every step of the fabrication process was planned in advance.
The result was a frame that perfectly fit the students’ research needs, delivered in only 3 days, allowing their research to continue without a hitch.
This was originally posted at Green Light Distrikt, then picked up by the New England Clean Energy Council blog.
Clean energy is a tough business. As entrepreneur Eric Smith put it, we’re really good at leveling mountains and burning coal. So competing with coal and natural gas on price will require every ounce of innovation we can muster.
But unlike information technology, energy technology is often expensive to prototype. For some energy innovations, the prototyping stage is so fraught with expense and uncertainty that it becomes a barrier, preventing good ideas from achieving commercial success.
For the past few years, I’ve studied clean energy prototyping. Much of my work has been with startups, including several in the Boston area. I’ve helped companies design and build prototypes, and interviewed founders on the challenges they faced during the prototyping process. Many of the same themes keep cropping up, so I’ve put together a review of some of the most common mistakes to avoid:
The “chicken and egg” problem:
This is the fundamental challenge of prototyping: A startup needs funds to finance prototype construction–-but raising those funds without having a functional prototype to show to investors presents a daunting challenge. This is especially important in energy, where the cost of an early prototype can easily exceed $100,000. There isn’t really any way around this, except to make prototyping as cost-effective as possible, so keep reading.
Don’t allow small problems to become big ones:
Startups, justifiably, are sometimes so busy accomplishing the impossible that they make expensive, time-consuming peripheral oversights. If your startup builds, for example, solar panels, then you might be justifiably less concerned with, say, the racks the panels sit on. But, really, your startup is just as dependent upon a viable racking technology as it is upon a viable panel technology–if the racks don’t work, no one’s going to be very interested. Moreover, off-the-shelf solutions are often somewhat inadequate, because, after all, this is a new technology you’re developing.
Example: One startup I worked with was so busy worrying about large, prominent prototype components that several smaller parts weren’t ordered until the rest of the prototype was nearly complete. These components had been designed by an intern, and the designs were never deemed significant enough to be reviewed by a more experienced engineer. But because they needed the parts now, the drawings were rushed out, and the parts manufactured based on the existing drawings. In the end, the parts were expensive to produce and had design flaws, causing a ripple of further delays lasting several weeks.
Don’t assume contractors are saints:
Contractors (like machine shops) exist, primarily, to make money. They make more money on rush jobs, so your delays and scheduling errors work to their advantage. They make more money on change orders, so it’s in their interest for you to make last-minute design changes. They make money whether or not you correctly specify exactly what you need, so they don’t care if you specify something that won’t actually work for you.
Example: One startup I worked with sent a machine shop a drawing for a part, but verbally requested a small change that would significantly speed up assembly later on. Although this change wasn’t documented in the drawing, the shop assured the engineers that it would be made. Of course, it wasn’t, causing precious days of delays.
Solution: Have your act together. Triple-check your drawings. Make sure part X is going to work with part Y. Set your tolerances so that you can call your contractors out if they make a mistake–it happens more than you might think.
Use materials & processes like a ninja–light on your feet, but deadly effective
To maximize the benefits of prototyping within tight restrictions on time and money, you’ll have to squeeze every drop of value out of every ounce of material you buy and every dollar you spend on labor.
Example: One startup I worked with was unaware that stainless steel could be welded to non-stainless (“mild”) steel, so they built one part entirely out of stainless, rather than selectively using the two metals as needed. Stainless steel costs between 2 and 6 times as much as mild steel–and it’s also more expensive to drill holes in.
Lastly: Get help if you need it. And you probably do.
I’ve long been impressed by the caliber of people in the clean energy startup scene, especially here in Boston; these folks are well-educated, passionate, and intelligent.
But intelligence and education are poor substitutes for technology and experience. Prototyping is cheaper, faster, and more effective when at least one person on the team has experience with the constraints, materials, and processes involved.
Moreover, prototyping is fundamentally a project management exercise, but some startups ignore this fact and put engineers in charge who have no project management experience or training. This is a great way to teach those engineers project management, but not necessarily a great way to run a project.
If your team doesn’t have expertise in prototyping, then do us all a favor and seek it out, because we really need good ideas to succeed, especially in clean energy.
.jpg)
Here’s what our new castellated saddle and bracket look like on the pipe: 
And with the manufacturer’s bracket for perspective:
