[EricH]

Building versus buying... An age-old debate.

If a COTS mecanum wheel will fit your needs, you can buy it. However, if for some reason the COTS just won't work, build it. This is a classic engineering tradeoff. To say that COTS isn't as good--what you really mean is "Not as X as you'd like", and usually the reason is that the wheel is a compromise of multiple factors, and they all got shortchanged somewhat.

It's never been a requirement to build your own mecanums. However, some teams do it because they can. Back when 357 first applied them to a competition robot (2005), there were no COTS mecanum wheels that would work for the competition. So, they invested in the infrastructure to build their own. The result was Jester Drive--molded parts of various materials, assembled into one wheel. See the 2006 Behind the Design book. Then AndyMark realized that there was a market for these wheels.

With engineering in general, the build/buy debate is critical. Usually, if you can buy a solution that can be modified to work, that is the way you go. However, sometimes that isn't possible, or the potential solution will not work. Then you have to engineer your solution and build it, which tends to be more expensive because you need to test it--and if you did it wrong, it doesn't work right, so you have to redesign and rebuild.