Creating 3D Printable Ties

Back to overview

MyTie3d v0_7 Illustration
by Boris R.
on Sketchfab

I met Sketchfab community member Boris Rabinovich at 3D Printshow in New York where he showed me his 3D printed tie. He has since published all the parts here on Sketchfab, allowing you to print your own copy. Today, we’re talking about the creation process of such a complicated piece.

Hi Boris, can you introduce yourself?

I studied math at the St. Petersburg State University focusing on geometry and topology. After completing my masters I taught college level math for several years. In 1991 I was accepted to a Brandeis University graduate program in math. Across the railroad track from Brandeis campus was the HQ of PTC where I found my first job in the US. I started working for PTC ​ as a summer intern in the geometric algorithms group and continued in a variety of roles in the R&D for CAD and Data Management products for a satisfying 20+ years career. At the beginning of the financial crisis I went back to Brandeis for a part time degree in Finance. In 2010 I became the chief architect of what became PTC Creo – a redesign of Pro/Engineer. More on LinkedIn.

What is your connection with 3D Printing?​

​In 2013 Hod Lipson visited PTC with an inspiring talk about 3D printing, which excited me enough to jump orbit. I bought an Ultimaker Original and started on a self-guided apprenticeship of additive manufacturing.​

What is your inspiration for designing this tie?

The inspiration came from a vintage Armani tie design that appealed to me as a balanced patch structure: fairly regular but not boring.


How did you design the parts? What was your process and which software did you use? 

Technical inspiration came from the PTC Creo project to support configurable assemblies.​ I had quite a few variants in mind. Interchangeable blades: necktie and skinny , long and short. The blade could be printed in a solid color, reversible, or with an ornament. I designed 3 kinds of neck loops: short hooks, longer hooks, and a full loop with a ziptie on the back of the neck. And 3 variants of the knot: flat, super geeky with planetary gear, and one with a hole in the middle and a locking interface where a number of objects, 3d printed or not, could be inserted. I further wanted to reuse the same meso-material for bracelets, etc. All in all seemed like enough requirements to warrant use of PTC Creo Options Modeler. However, with each tile having unique shape, relationship between the tiles of the tie is too complex. ​T he blade consists of over 100 similar but unique tiles. Cost of tooling to manufacture so many unique parts with subtractive methods would be prohibitively expensive for a consumer product. With AM, the manufacturing complexity is gone while design and data management complexity remains.


The data was structured so that each key parameter would be defined once for the whole assembly and shared by all the parts. For that I created an empty part with the parameters (thickness, dimensions of the connectors, gap between the tiles), exported it out of Creo to a neutral file format, imported it back into another empty part and used that part as a starting template for all tiles. This allowed me to update all the tiles at once and gave me explicit control when to do the update.

Next I created a prototype tile where I modeled every possible connector (male/female, removable/fixed) and created a Publish Geometry feature for each connector type. Now to create a connector in a tile part all I had to do was to mark the desired location with a coordinate system, copy geometry from the prototype tile, and solidify it.

As a final preparation step I created the design pattern part, which defined shape of the whole tie as well as of each tile using sketched curves. To get the desired design I first imported the image and traced splines over it. Then I could move and adjust the curves as necessary. Individual tiles would reference the design pattern part in order to derive their own shape.


With this preparation in place, each tile became a lightweight part with only a few features, which regenerated quickly and reliably, for the most part. This structure served me well while I was working out the shape of the connectors needed to make the meso-material durable and flexible and to avoid tiles fusing together during printing. When making changes to the shape, however, I couldn’t find a reliable way to propagate changes from the design pattern part to the individual tiles. Before AM, massive late changes to the shape of the parts was unlikely due to high cost of re-tooling. Now, with zero tooling, the manufacturing complexity is gone, which puts additional pressure on the designer and the CAD tool. 

The knot part presented a data management challenge as all of its geometric complexity is hidden inside and I could not visually tell apart prints of different iterations. I had to introduce part numbering and attach QR code stickers to the prints.


For the first couple of months of this project I was ashamed by how hard it was to use PTC Creo. Over time I learned how to use it better, how to avoid some of the hidden traps, Creo 3.0 replaced Creo 2, and I began to enjoy some of Creo’s awesome capabilities. Still, changing roles from a software architect to an end-user has been a thoroughly humbling experience.

Can the parts be printed on an Ultimaker as is, or does the model require assembly after printing?

​The tie is too long to fit on the Ultimaker’s platform, so it is printed in 6 parts. The parts have to be assembled together, .

Got any future plans?

​The goal of MyTie3d project was to create a sellable 3D printed product – the goal I have so far failed to achieve. For example, the version now published on Sketchfab is not durable enough and is not easy enough to use to be sold. I have since redesigned all of the components but haven’t finished the knot part yet.

Most of the rest of my future plans are in software. AM technologies change design and manufacturing in many ways. Among them, accelerating innovation by allowing designers to prototype often. ​Daily prototyping is mostly done using prosumer level desktop printers utilizing FDM or SLA technologies. To reach its full potential, rapid prototyping should be fully integrated into the design process. I believe that a designer should be able to evaluate the printability, make corrections, setup the print, review the results of a simulation, initiate the print job, and follow its progress – all without leaving the CAD tool. That’s my plan.

Thanks Boris!

– Bart


About the author

Sketchfab Team

Sketchfab Team

Leave a Reply

Your email address will not be published. Required fields are marked *

    Related articles