Designing The Wedding – Part 7

It’s better late than never–at least that’s what I tell myself. At this point, I’m glad I finally found time (and we don’t even have kids) to sit down to knock out the very last part of this longer-than-expected series: the centerpieces! Designing, and making the centerpieces was probably my single favorite portion of planning the wedding. 

I knew I wanted to make the centerpieces at my wedding even before I knew who I wanted to marry. I didn’t know what form the centerpiece would ultimately take at the outset since I planned on merging both Tiff’s and my tastes. I’m extremely proud of the product we (this was a team effort) ended up with. For those who haven’t seen yet, this is what they looked like: 

Being an engineer, I began the process with a set of product requirements:

  • Symbolic: The centerpiece had to represent Tiffany and I as a couple. 
  • Aesthetic: It had to look good and match the overall theme of our wedding
  • Whimsical / Interactive: I wanted the centerpiece to be somewhat unique and fun. 
  • Manufacturable: About 10 needed to be made 
  • Easy assembly: Obvious for somebody who had never seen it before, with limited instructions
  • Packaging: Simple disassembly and airplane-proof for guests to take home

With the requirements in mind, I began an iterative process of creating components, testing them, and refining them. 

The first component I created was the top lighted portion. The shape is called a trefoil knot and it has a few interesting (for nerds at least) properties which double as symbolic attributes. First, like a ring, it’s made of a single unbroken strand (like marriage). From various angles, it looks almost like a three dimensional infinity sign (representing an everlasting love). The trefoil knot is also nontrivial in the mathematical sense in that it cannot be untied without cutting (insert your own apt marriage analogy here). Another property which made it highly adaptable for lighting in this project is the fact that even though it appears as if the faces twist around, they actually keep the same orientation with respect to each of the other faces (I don’t know enough math terms to describe this properly). In practical terms, it means I could insert a single strip of LED rope and have all the lighting be visible without cutting it into separate sections. 

Didn’t expect to see parametric equations and trig functions? You were wrong 😉

The physical model for the knot was created by extruding a U channel along a curve defined with a set of parametric equations (shout out to IMSA math teachers! I knew this stuff would be useful). I tweaked the parameters for the shape to fit my print bed, fit exactly 1 meter of lighted rope, and have “lobes” that look good.

The first prototype print… was a PITA to clean up and prepare, but it did its job

As you can see, my first print (top) had severe stringing and had a lot of overkill in terms of support material which made it a bit of a nightmare to finish. Luckily, I’ve got a bit of experience in optimizing print settings which helped in the production process down the line ;). 

My first prototype trefoil knot used some WS2812 LED strips left over from projects I completed years ago. While it worked as a proof of concept, I didn’t like that discrete LEDs were clearly visible. 

Searching for a replacement, I found a newer LED strip product with super high density (96 LEDs per meter). Additionally, I started the project out by planning to control the LEDs by rolling my own arduino code and creating some basic power management circuitry as I had done in the past, but I was amazed to find integrated controllers (SP511E) which did everything I wanted to and more for cheaper than I could buy individual components (thank you AliExpress!). This awesome little controller has two different power inputs for either battery or wall wart, has an IR sensor for a remote control, has a ton of pre-programmed light patterns, AND has a sound reactive function! Above is a prototype video showing all of this in action. 

The centerpiece design I went with in an alternate universe

I toyed around with the idea of having a vase with flowers inside the trefoil knot for a little bit, but ultimately we decided that the lighting would look better raised. We also searched for and tried multiple color options on both black and white plastic before deciding on gold paint on white. 

A couple of many design iterations for the centerpiece base

I went through multiple design iterations for the base as well. I optimized the overall height, the number of and shape of the “layers”. Additionally, I had to ensure all the control components could be hidden from view to add to the “magic” factor. 

After finalizing the design and building the functional prototype seen above, we got to work in the production phase…

Not only did I need to print about a dozen trefoil knots (each print took ~18 hours), I needed to cut the same number of bases while minimizing plywood waste material. 

This took over our living space for quite a while… Not shown in any of these photos are the hours that Tiffany spent sanding to finish the tops before spray painting, the time I spent trimming and soldering the LED ropes, and the preparation of the floral elements.  

Arguably the most important element of the design is how easy it was to assemble and disassemble… it doesn’t matter how good it looks if nobody else can put it together and take it apart again. I knew I would not have time on the date of the wedding itself to do either, so I put a lot of work up front to make everything as intuitive as possible including: optimizing cable routing, creating sub-assemblies, and writing work instructions. 

Thanks so much for sticking with me for this series for almost a year now. It’s been a ton of fun for me, and I hope it’s been at least somewhat interesting for you :). I want to take this opportunity to once again thank my wife, family, friends, and vendors who all made our wedding special and unforgettable!!

For you readers that have made it this far… Here’s a little thank you video I put together that shows the centerpieces in action at our wedding! 

Bailey Slow Feeder

TL;DR: I printed a slow feeder bowl for Bailey out of food safe filament and coated it with an FDA-compliant food safe resin. It works, but I think she hates it and me for making it.

I’ve wanted to experiment with 3D prints out of PETG for a while, and finally found some time to do so during this long Thanksgiving weekend. PETG (polyethylene terephthalate glycol-modified) is a 3D-printable plastic with numerous advantageous properties:

  • High strength
  • High density
  • High temperature
  • UV resistant
  • Food safe

The drawback is that it is a bit trickier than PLA (the most typical home 3D printed filament) to print. I invested an hour’s worth of time to adjust settings and complete two quick test prints before deciding I dialed my printer in enough to start a real project.

Our dog Bailey is a voraciously fast eater and I recently learned that various slow feeder bowls existed. However, most of the products on the market seem designed for larger dogs, and I wanted to make something that would fit in our existing holder. Spoiler image below:

Bailey’s custom slow feeder bowl fit perfectly!

The design of the bowl was straightforward. I measured the dimensions of our existing metal bowl, added another mm of thickness of the bowl for strength, added an extruded “B” in the middle to act as an obstacle, and finally made some cuts in the B to allow Bailey to access all the nooks and crannies. I took care to fillet any sharp edges away to ensure safety:

Left: Top view, Right: Iso view. I cut the “B” into an arch so Bailey could reach all the food and made sure to fillet all sharp edges.

The quality of my first real PETG print exceeded my expectations… Based on troubles I’ve read about people having, I expected some blobs/zits or stringing issues, but surprisingly, I didn’t have any real problems at all. The ease of support material separation was shocking too–99.8+% of my support came off in a single piece, and the remaining two pieces were easily removed with pliers:

Top: 3D print in various stages of completion. Bottom: Removal of support material. I was surprised at the great quality, high strength, and ease of support material for my first PETG print.

After the print was complete, I coated the bowl with this neat FDA 21 CFR 175.300 compliant resin I bought a while ago but hadn’t tried out before. The biggest pain point with the coating process was the 48-hour cure time.  Luckily this was a long weekend, haha.

Right: Top view of bowl post resin. Left-Top: Showing off shininess of the bowl. Left-Bottom: Water beading up on the print after washing. Not pictured: 48 hours of waiting, and the popsicle stick, old tofu container, and paintbrush all sacrificed to make this happen.

After washing the bowl with soap and water, I tested it out with Bailey by putting in a few training treats. She did not look very happy…

She eventually came around to eating, but clearly wasn’t happy:

I hope she doesn’t hate me forever because of this…

Does she look most annoyed, confused, angry, or disappointed?

Baymax Cord Lock

TL;DR: A cord lock for Tiff’s hat broke… so I made a replacement shaped like Baymax since I had white material installed in my printer and I was too lazy to change it, haha.

Tiff got a great sun hat from a friend’s beach birthday party last year. I use it almost every day when walking Bailey. I noticed that the cord lock was starting to break, so I decided to make something useful while scratching my maker itch now that the need for PPE has declined.

Since I currently have black TPU (an elastic material not really suited for this application) installed in one printer, white PLA installed in my other, and a severe lazy streak, I needed to design something white and ovoid.

White? Check. Ovoid? Check. Fun? Check Check Check.

With proper source material in place, knocking out the design was straightforward for me. I decided to use the sculpting tools in Fusion 360. Sculpting is great for quickly making organic shapes that don’t require a lot of exact dimensions. Fusion makes it super easy to combine sculpted forms with parametrically defined features as well. I split Baymax’s body into two parts, one main body and a removable front plate to install the spring and legs.

Sculpt and boolean tools in Fusion 360 made designing and cutting the parts up for 3d printing a breeze!

Since the part was very small, I initially had some troubles with Cura deciding some areas (primarily the cut out for feet to retract into the body for cord installation) were so thin that I must not have wanted material there. I solved this problem by reshaping the Baymax body a bit and scaling the parts up by roughly 15%.

Functional? Check.

Overall, I’m pretty happy with the results… despite it looking slightly terrifying, IMO… like Baymax lost a fight. Maybe I should have gone with some sort of squid ¯\_(ツ)_/¯.

As always, hope everybody is staying safe and healthy!

UPDATE: 7/25/2020:

Looking at the Baymax cord lock I made last week depressed me because it looks like his body is getting pierced by some sort of tentacled foe. I decided to replace it by designing a Blooper (the squid thing from Mario games), since it is also white, but looks natural with long arms:

I think Blooper looks better than Baymax cause the strings are tentacles XD

I used all the same tools I used for Baymax to make Blooper, but it was much faster the second time around. While I like this cord lock looks better, but Tiff doesn’t like it because of all the legs, hahaha ¯\_(ツ)_/¯.

COVID: Endgame

TL;DR: Since the acute need for PPE has diminished, I am no longer producing parts on regular basis. However, I do have a reserve of face shields and earsavers remaining, and am more than happy to ramp up production if you or anybody you know need equipment.

Over the past 8 weeks, I personally manufactured about 1000 face shields and 1000 ear savers on my two 3D printers, delivering a quantity of about 880 of each to healthcare friends and friends of friends in places all over the country including: LA, SF, OC, Oakland, Tennessee, Oregon, South Carolina, Georgia, and New York. Furthermore, two local groups I work with have distributed over 75,000 and 22,000 face shields and other units of PPE, respectively.

However, it appears that more and more hospitals are getting their supply chains back in order, and the shortfalls do not seem as desperate as they were a few weeks ago.

This is what ~50 lbs of empty filament spools looks like

I feel this was a huge accomplishment, and I could not have done it without the support of everybody who chipped in for expenses—it was incredibly generous of you. I plan to donate the remaining funds to the charity Good360 in a few weeks if the need remains low and seems unlikely to ramp up in the short-to-medium term.

I hope everybody has a wonderful Memorial Day Weekend, and stays as happy and healthy as possible. I sincerely hope enough of us remain vigilant and change our habits enough to ensure the gains and sacrifices we’ve made the past few weeks are not wasted. I pray that the worst of this situation is truly over for us. However, if there’s one silver lining to this, I know that if the need for more PPE arises again, we’ll be able to ramp back up much faster next time.

Quick update — now with earsavers!

TL;DR: In addition to face shields, I’m now producing NIH-approved earsavers. Let me know if you need some!

About two weeks ago, I upgraded my old cloggy 0.4mm nozzle to a great 0.8mm nozzle courtesy of Micro Swiss ( Making this switch greatly increased my printing capacity—when you go from a smaller nozzle to a larger one, the volume of material you can deposit increases by r^2–you reduce both the travel count within each layer, and increase the layer height at which you can print at. This leads to a huge boost in printing speed, with the drawback of losing details. However, for what I’m mass-producing right now, loss in detail is a very minor concern, so cutting my print time nearly in half on one printer is well worth the trade off.

While I continue manufacturing and delivering NIH-approved face shields on one printer, I’ve dedicated my other to the production of NIH-approved earsavers ( for the next week or two. These popular devices are great for relieving pressure off the ears of healthcare workers who need to wear surgical masks for hours on end during their shifts. By the end of this week, I will have delivered over 350 of them (including shipments to South Carolina, Tennessee, Oregon, and NorCal!)

Here’s a snapshot of my life for the past few weeks:

Left: 75x frames and 260x earsavers ready to be delivered this weekend
Right-Top: I’ve chewed through quite a bit of material… each spool is 2 kg >.<
Right-Bottom: 100x frames and 400x shields delivered last week

Let me know if you or any of your healthcare worker friends need any face shields or earsavers! I’m happy to ship them out.

Again, hope everybody stays safe and healthy out there!