Hackaday Blog Entry
The project got a blog entry on Hackaday, and this gave us some interesting feedback and useful information.
Reading through the comments, one thing that stood out was many people had an optics kit growing up, usually with molded plastic lenses. These are obviously meant for kids, the kids grow out of them, and the lenses get all scratched and foggy.
Heirloom Optics Set
Taking inspiration from the Heirloom Chemistry set, we're envisioning the optics bench as a quality product that you might use as a college student, perhaps use at home as a gentleman scientist, and then give to your teen-age children so they could learn optics.
We're trying to make a kit that has low-level learning experiments as well as high-end setups that can be used as part of an experimenter's kit.
(We're not planning on making an actual product, though it's sometimes useful to think in those terms to help set the scope and prioritize the tasks.)
We're also trying to come up with ways to scavenge the components so that people who are on a budget can build the kit for little or no money. For example, you can take apart an old scanner/printer and recover some glass front-surface mirrors and possibly a prism.
Inkscape doesn't do CAD files very well
It turns out that Inkscape has problems with CAD files. It doesn't understand the POLYLINE construct (common in DXF files), converts bezier curves to lines, and doesn't output paths very well. When we import the resulting files, we have to "unite lines" on our laser cutter to consolidate line segments into paths.
This isn't a problem for us (several team members use high-end CAD programs), but we need to recommend something for end users to use and as a basis for wiki descriptions. For example, we recommend tuning the magnet holes to compensate for inaccuracies in their laser, and have a descriptive narrative that walks the user through doing this.
So we're looking to QCad. Selecting and scaling a feature (magnet hole, for instance) seems to be simple, QCad is free, and runs on all platforms.
Our laser is Inaccurate
Our laser cutter got tuned by the resident laser expert, and this is causing some trouble with the holders.
(This is not a problem, we recommend that the end user make test cuts and modify local copies of their files as needed. It's not hard, there are only a few features, and they might not need to be modified.)
This got me to thinking about our laser: we're making assumptions based on perfect cuts, but we don't know how perfect our cuts actually are.
For example, the 6mm magnet holes are actually 5.9mm to make a strong friction fit. This works for our laser, but now I'm wondering if that's specific to our laser.
So I'm going to go over everything in the files and holders with a micrometer and my AWG drill set to determine what the correct scale differences should be.
The plan is to design files that assume a perfect cut, since that's the design most likely to be correct for the end user, then advise them to make test cuts and adjust as needed.