Hey, check it. This is a post about CAD software for designing bikes. Lots of people use Solidworks or an AutoDesk solution, which is cool. I have beef with those models, so I use FreeCAD. It's fair to say that since version 0.17, a lot of people have criticized practical usage problems in the software. The most recent stable release is 0.21, and it is pretty nice. It still has the initial welcome first-run display of being difficult to get started, and it still has quite a few problems with filleting of solids with regard to swallowed edges. However, a LOT of quality of life things have been improved.
What follows is a video of a work in progress linkage fork I'm making for myself. It was made using a Python script for updating a "master sketch", which barfs out numerical values for various objects to reposition around. Even the shock coil is physically altered. A screenshot is stored for each positional update, and this stream of images is assembled into video afterward.
I know y'all are familiar with why most people use telescopic forks. I've used them too. I get why they work well enough, I've read the Tony Foale stuff and the James Parker stuff, I get why "we are forked" one way or another regarding the far-from-ideal task of making a fork. I just want choice of shock and the ability to hot-swap shocks. You don't have to like it, but it's a thing I'm doing with eyes open. I spent a great deal of kinematic analysis on getting the instant center to scrape along the ground line to explicitly avoid a high anti-dive value. This "don't use the anti-dive feature" work was by FAR the most difficult component to nail down. These designs so readily want to become very pro-dive or very anti-dive, real fast. I found it most useful to adjust the relative angle of the link pivot ends, rather than the driving the lengths of the links directly.
The axle path is almost straight, with a slightly rearward crescent path that deviates away from a straight path by 3.5mm maximum. Compared to the Trust Performance Message fork that has a 26.0mm maximum deviation, my design is rather mild and should feel intuitive.
One upshot about the curved fork uppers, the lack of conventional telescoping stanchions means I could get that crown shoulder down a bit, in order to, say, clear an old MTB frame that wasn't designed for long travel suspension crown clearance at the down tube. If one were into that sort of thing. (Which I am.)
I tried to get the shock as close to the center of steering axis as possible while also allowing finger / tool access. The components are thin, which will be permissible since I'm going to use a carbon fiber sandwich / core construction. Note that the non-suspended-mass yoke is quite thin, but broad. I've made 5mm thick carbon sandwich panels that seemingly do not bend. My thinking is that since I don't need to physically touch the fork very much, the structures do not need to be round.
The design is EXPLICITLY based on the 1980 Valentino Ribi motocross design that was raced with some success in a long travel off road application.
I really liked the smooth action of riding a Lefty that glided on needle roller bearings. I did not like interacting with Cannondale. I want the next best thing, which is really, really ugly.
Anyway, again, this post is mainly about FreeCAD and the sophisticated stuff it can do in its current state.
Say what you will, or don't.
It seems Firefox is having trouble with embedding the youtube link in this forum, here's a link if you're having trouble viewing.
What follows is a video of a work in progress linkage fork I'm making for myself. It was made using a Python script for updating a "master sketch", which barfs out numerical values for various objects to reposition around. Even the shock coil is physically altered. A screenshot is stored for each positional update, and this stream of images is assembled into video afterward.
I know y'all are familiar with why most people use telescopic forks. I've used them too. I get why they work well enough, I've read the Tony Foale stuff and the James Parker stuff, I get why "we are forked" one way or another regarding the far-from-ideal task of making a fork. I just want choice of shock and the ability to hot-swap shocks. You don't have to like it, but it's a thing I'm doing with eyes open. I spent a great deal of kinematic analysis on getting the instant center to scrape along the ground line to explicitly avoid a high anti-dive value. This "don't use the anti-dive feature" work was by FAR the most difficult component to nail down. These designs so readily want to become very pro-dive or very anti-dive, real fast. I found it most useful to adjust the relative angle of the link pivot ends, rather than the driving the lengths of the links directly.
The axle path is almost straight, with a slightly rearward crescent path that deviates away from a straight path by 3.5mm maximum. Compared to the Trust Performance Message fork that has a 26.0mm maximum deviation, my design is rather mild and should feel intuitive.
One upshot about the curved fork uppers, the lack of conventional telescoping stanchions means I could get that crown shoulder down a bit, in order to, say, clear an old MTB frame that wasn't designed for long travel suspension crown clearance at the down tube. If one were into that sort of thing. (Which I am.)
I tried to get the shock as close to the center of steering axis as possible while also allowing finger / tool access. The components are thin, which will be permissible since I'm going to use a carbon fiber sandwich / core construction. Note that the non-suspended-mass yoke is quite thin, but broad. I've made 5mm thick carbon sandwich panels that seemingly do not bend. My thinking is that since I don't need to physically touch the fork very much, the structures do not need to be round.
The design is EXPLICITLY based on the 1980 Valentino Ribi motocross design that was raced with some success in a long travel off road application.
I really liked the smooth action of riding a Lefty that glided on needle roller bearings. I did not like interacting with Cannondale. I want the next best thing, which is really, really ugly.
Anyway, again, this post is mainly about FreeCAD and the sophisticated stuff it can do in its current state.
Say what you will, or don't.
It seems Firefox is having trouble with embedding the youtube link in this forum, here's a link if you're having trouble viewing.
