I have been doing some research into yoyo design, and have been having fun working on a few designs. I have many years of CAD experience, and was exploring what kind of dimensional tolerancing others use.
I was shocked to read that multiple sources indicated that some dimensions are held as tight as +/- .0002" to +/- .0004". I would consider +/- .005 a normal “tight” tolerance, and that what is suggested is 10X as tight! Are yoyo halves really expensive to produce? Why so tight?
Anyone have insight into this, or has tried to produce a less-tightly toleranced yoyo? Thanks for the help.
My guess is parts like the bearing seat, or rim fitment on multi material has to be very dimensionally sound to eliminate vibe. A looser bearing post will introduce vibe, and a bearing post thats too wide just wont work. I imagine you can get away with looser tolerances if you’re fine to have some vibe
Yep I was just having a discussion about this. In metric I had a shop produce some halves with +/- 0.05 and found out after the fact that this tolerance only really works for responsives (good thing mine is responsive). Talking with @Lotaxi who machines yo-yos he suggested that 0.01 mm is the minimum tolerance for unresponsives to ensure a good fit with minimum vibe.
It might be that I’m just in that particular niche of machining, but ±.005" is a pretty standard tolerance to me. Not exactly wide, but routine enough that I don’t really need to think about it to stay in spec. A .010" window is a mile wide. ±.0025" is where I’d start to consider something on the tighter end of things, but a .005" window still not too hard to hit.
I get paranoid and meticulous about everything when I get under ±.0015" or so. Holding total runout along an 8 inch shaft under .003" (±.0015" deflection from a cylindrical datum in any direction at any point) is scary, for example, where holding the diameter of a feature to ±.003" is trivial.
You only really need the tight tolerance (under ±.001") when it comes to coaxiality of the cup and profile and the fit of the bearing on the seat post.
For the cup/profile coaxiality, you obviously need everything to be balanced properly at a few thousand RPM, so that’s where the ±.0002" comes in. It’s not super hard to hit if you’re setting your parts up right and you have a robust fixturing solution, but it’s the kind of thing you’re actually going to plan for in fixturing and pay close attention to as you’re loading parts. I make custom fixturing for every design I cut.
As for the bearing post, you need to hit a locational transition fit (Class LT1) to keep things properly stable. A running fit (RC) or a looser location fit (LC) are going to allow the bearing to slip and shift, causing a state where your halves might become eccentric and thereby become an asymmetric mass vibrator. I think a lot of the Chinese manufacturers just take the guesswork out and use a light drive force fit (FN1).
Those are the only 2 truly critical tolerances in a monometal yoyo, as far as I’m concerned. Bimetals or multi-piece designs are going to require proper fit calculations for all the relevant components, but they’re a separate matter. Keep yourself within the range of an FN1 or FN2 force fit for a bimetal ring and things should stay pretty happy as long as the coaxiality tolerance described above is observed.
Mass is important, but you can have a yoyo come out a little too heavy or a little too light and still have it play properly on the string. The characteristics might be a little different than envisioned by the design, but it’ll function properly. Same with diameter and width.
Kinda yes and no? I’d say they’re more expensive to set up and measure than they are to run properly. Once your machine is set, the CNC will keep things happy as long as you’re paying attention to how you’re loading the machine.
Keep the 2 main tolerances intact for unresponsives or you’ll have bad outcomes.
Responsives don’t need to stay vibrationally stable for long periods of time. They need to be stable enough that they’re not uncomfortable to play at the end of the string and they need to return to the hand when called. They are also usually considerably thinner width than unresponsives are, which makes them MUCH more modally stable along the axis of rotation and therefore more tolerant to sloppier fits.
I’m always around to answer questions, if you’ve got them.
