Back in 2018 I started my business Servicing, repairing and customising mountain bike suspension - I have ridden bikes most of my life and been a mechanic since I left high school. I found suspension super interesting due to how technical it was, but also found the was very little education and knowledge available to people. It is a huge influence in how comfortable a bike feels, how fast it goes (if you’re in to racing) and how enjoyable it is to ride. Each person is so different and I wanted to be able to explore the best ways to make it enjoyable for the most people.
So when I opened I invested in tings like data aquisition - measuring suspension movement on the bike, and a damper dyno for quantifying the hydraulic parts of shock absorbers. This was a a huge learning curve but became super valuable - these things were/are not well understood and mostly a “black box” to people.
I had played a little bit with computers as a kid, and during my one year of university I enjoyed the engineering programming part (MATLAB) but hadn’t done anything further with computers and electronics. But trying to learn how bikes work led me to needing to aquire tools and learn how to understand the measurements I was getting, which turned in to a slippery slope! Wiring sensors properly forced me to learn some basics of electronics, and I wanted ways to visualise data in a way to gather new insights or share information with people, which soon led me towards learning python! I was also curious about CFD for maybe modelling hydraulics and aerodynamics (also a big F1 fan btw…) and playing around with the open source OPENFOAM is what first encouraged me to install Linux. From here it progressed in to learning more and more about programming, electronics, CAD and even homelabs (although I hadn’t learned that word yet). i was generating tons of data and using all kinds of services but began to realise how much I was depending on outside companies so wanted to learn even more about self hosting as much as I could. I cobbled together a very cheap old PC to install truenas, and turned my old desktop in to a Proxmox machine to run various things like home assistant, docker and ollama.
I’m now upgrading that in to a more purpose built server including a 3060 12gb for locally hosted AI models, and my new goal is to take the cobbled together patchwork of tools and python files I’ve made over the years and combine them in to a much more powerful and efficient package I can use to be more productive in the workshop.
This is where the R9700 would be extremely valuable -
I want to build software I can use for the whole workflow - manufacturers don’t supply much info about their bikes so currently I have to take measurements myself, which can be used to model the suspension and calculate settings for the bike set up. Each step is very clunky and not as reliable as I would like so I am working on an eco system for the whole process, examples of a few of the pieces are-
Physical measurement tools like how much the bike sags under weight
(this is a prototype simple distance measuring tool made with 3d printed parts and the spring from a retractable keychain and a magnetic rotation sensor)
Photogrammetry or 3d scanning to capture bike geometry
I scanned this bike frame I built myself using a mobile app but would much prefer to process this on local hardware!
Then I would use AI tools to combine all of the knowledge and data I have to make better, faster prdecitions about the best way to set up bikes for people! I have mountains of settings from everyone I’ve worked with, but putting it all together to extrapolate setups for other people is really slow so being able to access it quickly would be a huge game changer for me!