Here's version 2, going back to 4 rollers per side, now using the "optimized" roller curvature and a flexure based capture mechanic for the roller axles. Blocky to allow the animation to run at a reasonable speed. You can see how the contact patch with the ball passes from one side of the wheel to the other smoothly.
I am assuming someone somewhere has tried omniwheels. This rolls freely in all directions and also twists with zero effort. Virtually no stiction unlike my OG MS Trackball Explorer, and also silent, unlike BTUs. The smallest commercial omniwheels I've seen are Rotacasters which are 35mm. These custom made ones are ~25mm diameter. I used pin type axles of diameter .6mm - simply rolling in PLA and the friction is remarkably low. Rather than try to figure out how to have grippy rollers, I took the approach of a grippy ball, this one being a quite large lacrosse ball. It's not quite perfectly round but good enough for this test and actually feels kind of nice on the fingers. It took a while to get these omniwheels to be as smooth as they are, mostly trying to figure out how to 3D print the tiny snap-in roller mounts so that they have absoluetly no side to side play. But the roll is a little bit "lumpy" probably due to not quite optimized roller profile handing from one roller to the next.
Hi, I've been considering a Viwoods device but I'm confused about the Google integration. I think that the device does not come with Google Services installed, but you can somehow use apps from the Play Store. I think the Play Store is somewhat dependent on Google Services (could be wrong) is it easy to install apps from the Play Store? I would like to use Google Drive on the device to sync documents, is it possible to log into a Google account from this device (i.e. from Google Play) and use that to connect to Drive?
No, not trying to print heels :-) I'm making a stand for my eInk device, and want to have the legs start as small round patches and then "swoop?" up to support the entire underside of the flat portion on top. Can this be done in Openscad?
OK this was really dumb. I was trying to get an M50 plug printed and noticed that the BOSL calls I made created the bolt upside down making it hard to print. Rather than rotate 180 I took a "shortcut" (saved a few characters) and did scale([1,1,-1]). After printing it, I drove over to install and it would not start. I stared at it stupidly for a while before even realizing it had left handed threads and then was like wow, you can't count on BOSL to make the correct kind of threads!? Only after I was trying to find the "handedness" flag for version 2 did I realize that my inversion of course changed the handedness of the threads. So I brought this ignominy on myself! GRRR Don't do it kids!
Did Google roll this out without a backup plan? It just doesn't work. To be more precise, I am trying to set up Google Messages for Web. This is something I have used and loved for a very long time. Suddenly the pairing steps are different than before, with no alternative/fail over to the previously working method. It used to be that the system would display a QR code that you could scan from the mobile device scanner and that would pair the devices. I have no idea why they would change this, it wasn't broken. But then one day the system won't connect and there's a new approach - choose the emoji that is appearing on your device. No emoji on the device, the device can only do QR codes. No alternative for pairing, it's just broken. I just dont understand how Google can let something like this happen. I trust that they are making the change to improve security or potentitially convenience but did they even test this behavior? I'm using Google Chrome latest on a Google Pixel 8 Pro also updated to the latest Android and Messages.