Manufacturing guy-at-large.

Filtering by Tag: mfg

Directions in automation adoption

Added on by Spencer Wright.

I've been spending a lot of time looking at industrial automation the past few days, and had an idle thought:

I've touched on this before, if only obliquely, when writing about's role in manufacturing logistics. Much attention is being paid to companies who want to simplify (or circumvent) some part of the product development value chain. Many of these are companies I admire, and think are doing really valuable things. Take Within, whose 3D design software generates structures that are driven directly from functional constraints (but can't, as far as I can tell, deal well with thin-walled structures). Or Willow Garage's PR2, the really slick research robot (that takes charmingly long - 20 minutes per bath towel - to fold laundry).

Each of these is an incredibly impressive feat, and one that follows an ambitious (and I would argue honorable) line of thinking:

If we can encode all of the information needed to complete a routine yet complex task, then we can use machines to automate the process, freeing up our minds to do other (presumably more important) things.

But consider an alternate proposal:

If we can get machines to mimic a series of behaviors that humans can plan and execute with relative ease, then we can decrease the amount of rote mechanical work that humans need to do.

This is the tact taken seriously by Baxter, the admittedly not-too-serious (but cool nonetheless) humanoid task robot built by Rethink Robotics. Baxter learns by physically training his movements, presumably by the technician who he's "collaborating" with:

Even the traditional robotics companies, like Kuka, are moving in the direction of using robots simply to execute the complex tasks that humans calculate and perform with ease. Here a Kuka robot is trained how to clean a permanent mold by a BMW employee:

Both of these robots' use cases share a key feature: There's still a human doing the "hard" planning and calculation about how the task will be completed. In each case the robot doesn't understand the physical constraints or goals per se. Baxter has some awareness of his surroundings for sure, but all he knows is that his arms hit something; he doesn't have the vision or awareness of why that happened or how to correct for it.

Similarly, the Kuka bot doesn't understand that he's cleaning a mold, or have the facilities to learn how to do better work. He's just repeating a toolpath that he knows a human told him to do. Which, in this case, is good enough - and a hell of a lot faster than waiting for a computer vision expert to give him the intelligence required to do better.

I'm not sure what the implications of this are for the companies working to automate the design and supply chain. But the philosophical difference is striking, and I must say that the more hands-on model is very compelling - and I expect it to be so for the foreseeable future.


Parenthetically: All of the industrial 3D printing market is currently driven off of this same model: An intelligent, experienced technician makes manual edits to 3D CAD data in order to get a part to print within its design constraints. Anyone who suggests that build optimization is "right around the corner" is, in my opinion, *not* to be trusted. We're in a world of basic research still, and an automated design-print-post process chain is many years away.


Added on by Spencer Wright.

I have about 50 of these in my inbox on right now:

Not all Chinese manufacturers are as charming as this one, but in general they're *much* more outwardly communicative than American suppliers. 

Public Radio updates

Added on by Spencer Wright.

My last week has been spent mostly trying to squeeze money out of our potentiometer knob. The one we had been using - a Kilo product - has terrible pricing, even at quantity ($2.38 @ 2000 units). So I remodeled the part and started poking around.

I got a bunch of quotes from some US machine shops I know that are closer to the $2 range. I also got a quote from Taiwan for about $.50 apiece (though I'll expect additional shipping/customs costs). 

Then, I put the part on At quantities of 2000, this should hit a good spot for a lot of US mill-turn shops - and so far the response has been pretty good. I'll let the quoting period run for a while still, but I expect to pay no more than $1 for these parts, and that's including clear anodizing and possibly shipping to NY.

These are all for quantities of 2000 units.

Meanwhile, Andy has been researching FM IC pricing; Zach has been looking into batteries and tactile switches; and Eden has been poking around at PCBA shops. 

My next focus will be on the lid and the speaker. The former is in a pretty good place right now, but the speaker wants some real love... updates soon.

Manufacturing logistics wishlist

Added on by Spencer Wright.

+1 for breakfast with a smart friend. Always good for taking a step back & seeing a bit of perspective.

I like mass market products a lot. But I want to really love the things that I make, and that's hard to do if you're trying to appeal to a mass audience. Compromises need to be made when you're designing for a diverse customer base; when the market is smaller, it allows the product developer to be a bit more choosy.

I'm excited that making short-run products is getting easier - but there's still a lot to be done. These are a couple of the things that have been bothering me.

Soup to Nuts design toolchain.

As a designer of traditionally manufactured parts, features are specified according to more or less knowable tool properties. When I place a drilled hole, the results will be predictable, regardless of my supplier's equipment or process.

The same is not the case with additive manufacturing, where build orientation and support structures matter a lot. As a designer of these parts, I want to be able to simulate varying configurations myself, so that I can specify the build procedure to my suppliers. Simplifying the CAD/CAM toolchain - ideally with solid/NURBS, T-Splines, topology optimization *and* build plate prep all in one application - will be crucial to lowering the barrier of entry to AM.

Surface finish specs.

When designing a part - especially those used in assemblies - surface finish is often critical. With additive manufacturing, surface finish may vary widely, depending on build orientation and support structures. In some cases it may be possible to reorient a part in order to improve finish.

But today, these determinations are preformed by machine operators whose knowledge of the process is experiential and not publicly available - making it difficult for designers to know precisely what they'll get. Machine manufacturers and job shops should work together to develop design guidelines and detailed surface finish specs, alleviating this uncertainty.

Full service plastics prototyping.

I love Shapeways. But they *need* to offer secondary services, i.e. tapping and tolerancing. 

Advanced Manufacturing 3PSCM.

Again, I love Shapeways. But they *need* to offer assembly, and custom packaging, and small parts that are traditionally manufactured (i.e. bolts). 

3D printing is a cool technology. But so is stamping, and you don't see sheet metal shops selling useful products directly to consumers. Without a supply chain management offering, I believe that Shapeways will be confined to just selling parts - not products.

Industrial supply catalog APIs.

I love McMaster-Carr. But they need an API that can talk to both my e-commerce and the API of my manufacturer. When my customers purchase an assembly from me, my supply chain manager (whether it's Shapeways or somebody else) would automatically place orders through McMaster-Carr and my manufacturing partner(s) - with all parts being shipped to the SCM and assembled just-in-time for shipment.

A word on MFG and Alibaba.

I'm all for services that provide me access to a wide range of manufacturers. But it is critical to the design process that that service be completely transparent. Like it or not, most designers are *not* manufacturing agnostic, and speaking directly to a manufacturer - and preferably to the machine operator who will actually be setting up & running your part - is key to producing successful designs.

Too much effort is being spent trying to disrupt procurement in ways that adds an opaque layer between me and my supplier, and the end result is that I learn little from the process. All I want is for you to put suppliers in front of me and then get out of the way.

Everyone in the supply chain should have a blog.

Why aren't these people talking about their processes? Why aren't they sharing the non-NDA work they've done? Why aren't they showing me the capabilities of the new machines they have? I would gladly pay a premium for a shop that's actively showing me the engineering feats they're accomplishing. Get bloggy - the openness is quite becoming.


I hope it is, at least :)

This is a poorly formatted quote.

Added on by Spencer Wright.

A week or two ago I posted a really great quote that I received on MFG. This one is on the opposite end of the spectrum.

Screen Shot 2014-01-07 at 4.24.33 PM.png

Note that although I specified DDP terms, the supplier (a middleman for overseas outsourcing) has quoted FOB China. 

The price is also totally noncompetitive. I hope to purchase these parts from a US supplier for under $3 apiece at this quantity - and I don't want to be waiting 6 weeks for a boat to show up, either. At larger quantities (1000s) I'd hope to be paying more like $1.75 to my door.

MFG RFQ View rates

Added on by Spencer Wright.

For traditionally manufactured parts, MFG has a large enough supplier network that you almost always get a handful of good quotes. My rack end, for instance, went out to over a thousand suppliers. About 75 viewed the RFQ, and to date I've received 7 bids. Another dozen or so suppliers are apparently "preparing quotes." 

When you shrink the pool down, though, the rates matter a lot more. There are a *lot* of swiss turning shops in the US, but very few that can produce DMLS titanium parts. So when I posted my seatmast topper, I tracked the results with a lot more interest.

The RFQ went out to 110 suppliers. 8 viewed it, and I have yet to receive a single quote.

Incidentally, there is at least one supplier on this list who did *not* view the RFQ but whom I found separately via Google.  

The thing that really strikes me here is MFG's total incapacity to move my project forward - quotes or not. In optimal situations, MFG is an effective marketplace to compare quotes for manufactured parts. What it's not good at, however, is the broader function of connecting designers with manufacturers. Anything I want to learn about the DMLS process - information that inevitably is going to come directly from the engineers and operations managers that are quoting and building my parts - is totally missing from the MFG experience. Instead, I'm left to do leather-to-the-ground work the same way I always have: an afternoon spent with Google + a phone line.

There's definitely a place in the world for a manufacturing marketplace, but I'm pretty sure this could be better.

Regardless, I'm beginning to develop a few old-fashioned leads on good DMLS job shops, and I hope to have some real - and reasonable - pricing in the next few days. 

File conversion woes

Added on by Spencer Wright.

This is why dimensioned PDF drawings are so extensively used in procurement. The top photo is from a supplier from; the bottom one is the STEP that I originally uploaded to MFG.

Screen Shot 2014-01-04 at 9.27.05 AM.png

This supplier has obviously downloaded my STEP and performed some translation or conversion on it, and in the process has deleted a few faces (you can see the difference in the foreground of the part, and in the areas he's highlighted in red). This is clearly an inexperienced supplier, and one that I would ultimately have a *really* hard time choosing. I'm not sure what he did or how he did it, but the fact that he made this mistake is an indicator that we'd have issues down the road.

(To his credit: the photo came from a message he wrote me saying that he "noticed some missing surfaces on the part file," and asking me to fix them. So he knew that there was a problem, but didn't understand what it was and wasn't able to troubleshoot it himself.)

In traditional manufacturing, 3D part files are created and edited in a program like Inventor or SolidWorks. The parts are then brought into a separate environment in the same application and drawn and annotated in multiple 2D views on a "paperspace." The resulting drawing file (.IDW for Inventor) is a dynamic representation of the original part; if you modify the part file, the drawing will update automatically. 

You *never* submit drawing files directly to a manufacturer. Instead, you export PDFs of the dimensioned drawings, and *optionally* include STEP files (which are essentially cross-platform 3D files) as a courtesy. The STEPs can be used to help the manufacturer set up their CNC machines, but they're for reference only; the PDFs (with all their dimensions and annotations) are what you're buying.

"Organic" shapes - like those that 3D printing is so well equipped to make - don't fit into this process well. Complex surfaces are *really* difficult to define clearly and completely in two dimensions, and so most 3D printed parts are built from solid files. In this case I submitted a STEP, which manufacturers will convert to an STL and then run through a slicer and feed into their machines.

The problem is that STEP files aren't immutable, and the supplier in this case has apparently deleted a feature from the part. In this case the result was obvious, but there are a lot of features that he could modify or delete that would be a lot more difficult for him to detect, and my QC job would be accordingly tricky.

This process should be better. The PDF workflow is inconvenient, but at least it's an effective barrier to issues like this one. 

Also, we need more, and more *good*, DMLS suppliers. 

This is a well documented quote.

Added on by Spencer Wright.

Most quotes on MFG - whether from the US or overseas - leave a lot to be desired. This one is pretty good, though.

Note, that doesn't mean that I'm going to act on this quote. But I like the comments a lot - they show a lot of foresight and care on the part of the supplier.


Added on by Spencer Wright.

I've been really lagging on my rack ends project, but I wanted to post a quick update.

A few months ago I posted drawings and a STEP file to MFG is a sourcing platform for manufactured parts of all types, with suppliers and buyers from around the world. When I was building robot doors, I used MFG to source probably $50k worth of parts, mostly from contract machine shops around the US - but also from mainland China. MFG is a great place to get a *lot* of quotes quickly, and I've developed a few long-term relationships with great suppliers from it. 

My current project is a small one - there are certainly hundreds of job shops around the country who could handle it. I'm hoping to spend $2-3 per part in quantities in the 100s, though it remains to be seen whether that's possible. 

I got about twenty quotes back, though I scratched a few outliers off the bat. The remaining sixteen are below. Note that the 1st quantity is for 1000, 2nd quantity is for 100.

As it happens, I've received samples from a few of the cheapest suppliers on this list... they were pretty rough, with a lot of hand-finishing (presumably to cover up low quality machining).

Note also that I requested all pricing to be DDP, which means that the supplier is supposed to include all duties and shipping to my doorstep. From my past experience, a lot of Chinese suppliers either don't understand or don't care to quote these terms, though a few of the quotes here actually specify EXW, meaning that the supplier boxes up the goods and then leaves the rest for me to take care of. For a box of stainless steel in Guangdong, the cost of shipping could easily add 25-50% to some of these cheaper quotes.

For a part like this one, I could probably pick one of the lower bids and go with it. The part is simple enough; the worst I would deal with would probably be a bit of slag on some of the drilled holes. I want to be careful about material selection, though - I need the part to be 304 stainless, and the risk of a supplier using a different grade is nontrivial. 

The fun thing to do is to order the small quantity from one of the lowest bidders, and see how it turns out. For less than the cost of a used PS3, you can get a batch of parts that's probably ready to sell or use - *and* you might learn something interesting.

In reality, I'm more likely to buy the parts from a supplier I know (and found on MFG a few years ago). The quote I got from him was under $4 per, which is competitive with many of the Chinese suppliers here. He quotes EXW, but shipping UPS for these parts won't be more than $50. I would also follow up with all of the reasonable US firms here and see what their responsiveness was - especially if I can find someone local-ish.

This particular part is low priority right now, but I'll probably follow up on this early in January. Expect updates.

A weird translation

Added on by Spencer Wright.

From a message on that I received the other day (emphasis mine): 

Hi Spencer,
Have a nice day. We are very interested in you project.
Could you advise your precious about our quote?
We sincerely hope to build a long-term cooperation relationship with you. We have been specialized in this area for several decades of years. And we have gotten ISO certificate. We always provide high-quality parts with reasonable price to our clients,  you could check our good rating commented by our clients. Hope starting work with you soon. Look forward to hearing from you soon.
Any doubt, please feel free to contact us at anytime.
Many thanks&Kind regards

I can't for the life of me tell what they mean by this. Who is my precious? 

For anyone interested in these types of mistranslations, you should really check out Language Log's "Lost in Translation" posts.