Has anyone had success press-fitting Vex VersaHubs into 1.125 ID aluminum tube? If so, do you have any advice regarding the process? If possible, this would work nicely for a winch drum/spool or other roller applications.
Because it’s on back order currently.
We make our own hubs to insert, we typically need to ream the tube to get the fit right. You can make it work with a 1.125 step bit to get the fit right.
1.25 OD x 1.125 ID tube is typically 1.115 actual ID
I’ve done this with some screws added through the tube into a tapped hole in the side of the hub to help transfer torque.
+1, can confirm.
This seems better than just a press fit. Did you have to ream the tube out at all? Is the hole tapped before or after you press it?
We didn’t have to ream the tube out at all. If we did anything, it was just a deburring tool on the inside edge of the tube. The tube was sourced from onlinemetals.com.
We’d put the hub in the tube and match drill through both to ensure that the holes lined up. Afterward, we’d take the hub out and tap it, then enlarge the holes in the tube for clearance fit.
We used 10-32 button heads; this left a VERY small margin of error to where you had material all around the threads but didn’t have the head of the screw interfering with the inside face of the hub flange. If you go this route, I’d recommend using smaller head diameter screws, probably. #8 SHCS.
Here’s a photo of the finished product:
I’d personally suggest something like two screws for retention and then two roll pins for strength here, especially for something like a hanger. You can use bigger shear pins than you can screws because the roll pins don’t have a large head, which gives you a bit more strength.
How are you getting the roll pins out if something goes wrong?
Pin Punch?
I’m just trying to figure out the geometry. blind holes obviously won’t let you easily remove a roll pin, so yeah a hole from one side of the hub to the other. If your goal is a hex axle through the versa hub, a 1/4" roll pin through that hex axle is not wise. I suppose you could drill all the way through without the axle and put in two shorter pins from either side, but then you get to punch one of the pins out with another through the hex hole for the axle, which seems questionable. Hole all the way through that doesn’t intersect the hex axle hole?
It just seems more complicated than two more 10-32 screws.
Bring an extra spool.
Indeed. Right now, our plan is:
- chamfer/deburr the inside of the aluminum tube
- press the vex hub into the tube
- drill a hole appropriate for a 10-32 screw through the aluminum tube and into the vex hub
- tap 10-32 threads into the aluminum tube and vex hub
- apply loctite to the screw and screw it into the tapped hole
We plan to do 2 screws per hub.
Regarding 4 - does it matter if both the tube and the hub are threaded? should only the hub be threaded?
Thanks for the help!
It’s much easier if only the hub is threaded. It doesn’t really add strength to make the tube threaded, and this way you will not have to worry about possible issues with alignment or anything like that.
Drill the tap-size hole through both parts at once, then tap the hub while you drill the tube hole out for a clearance fit.
As for the roll-pin thing: It is harder to take apart, but your metal hub isn’t going to fail before your hex shaft does, especially if it’s an aluminum hex shaft. You can drive the pin into the center using a punch if you have a pin shorter than ~1/2", I guess?
If you’re not comfortable with the roll pin (which is reasonable), add 2 more screws instead. You really don’t want the entire weight of your robot relying on just two 10-32 screws, right?
Doesn’t the press fit / interference fit take the majority of the load? Apparently teams have used products like these alone to climb.
A good epoxy is perfect for this sort of situation. You’ll get a lot of adhesive area in shear, the mode in which it is strongest. It will be quick to build and assemble, and can be taken apart with heat (a heat gun will likely suffice) if needed. Or, build a spare.
Pro roll pin argument: Roll pins can be drilled and inserted at the same time, so you know there’s no gap between the pin and hole. Or very very little. So you know both pins will take up about the same load.
Screws, especially in clearance drilled holes, have a tendency to be different distances from the hole, so one screw may take up all or most of the load, which can lead to a zipper like failure of one screw after the other. This is the reason it’s discouraged to use bolts and screws in shear (side-to-side) loading situations. The gap between the hole and screw is why it’s so hard to get a bolted connection really rigid in the side-to-side direction. Pins and rivets are much better for shear loading. And much worse for tensile loading, which bolts are great at.
I agree. However, in a climber, I think the rules can be bent.
We only see load form one direction, so things won’t be shucking back and forth (how a lot of damage gets generated when bolting concentric parts). A single 10-32 screw like those ordered from McMaster (with 120ksi strength) has a shear load capacity of around 1200lbs. When bolted around the perimeter of a drum this would result in a FoS of 1200/155=7.7, a healthy FoS if I ever saw one.
If the single bolt is properly torque to a conservative tension, say 1000lb-force, it offers a tremendous torque-holding potential. CoF of al-to-al when clean is 1.05-1.35, we’ll call it 1.1. This would result in a FoS of 1000*1.1/155=7.1 per fastener (and torqued fasteners will add, unlike loose ones as you correctly point out).
Related - I would definitely trust my robot to be picked up by 2x properly-graded and installed 10-32 screws.
A proper interference fit, where the fit in both parts is precise, can be strong enough to lift a climber. However, a team that’s capable of doing that is a team that doesn’t need to be buying and retrofitting VersaHubs in order to make their winch drivable. This is clearly a team without the resources to make holes and ODs that precise, so bolts are going to be the primary method of torque transmission.