ThunderHex Bearing Explained

Technical Technical Discussion
1

Hello all,

In the VEXpro unveiling thread, some individuals brought up some questions regarding ThunderHex and its usefulness in FRC robots. This post is to give everyone some insight as to why we developed ThunderHex.

Believe it or not, there was a time in FRC before the 1/2" hex bearing existed. Keyed shafts ruled the FRC playing field and standard round bearings dominated.

Some teams, mine included, started using standard hex stock found online instead of keyed shafts. For 3/8" hex and 3/4" hex, there were round bearings that would fit nicely on those shafts if they were turned down on a lathe ever so slightly. Metric 10mm ID and 20mm ID worked perfectly in those situations. The team I ran used 3/4" hex live axles with 20mm ID bearings on our drive train. We used the turned down shoulders to act as stops and needed little to no spacers in the design. These round bearings were smooth running with little to no clearance between the turned down shaft and the inner race. Live axles ran smooth and true.

Only teams with lathes (some with mills) could use these types of bearings.

Then came the hex bearing, which opened up live (cantilevered) axle usage to more teams by allowing them to use hex stock as is (that is, if it wasn’t too big and needed to be filed down by some team freshman).

Hex bearings are nice, but are extremely low precision on the inner race. Most all ball bearings have a ground inside diameter for holding tight tolerances to make sure the bearing runs with little to no runout. When putting a hex hole in a bearing, it is very imprecise compared to the ground round hole. Add to that, the relatively thin cross section of the bearing makes the broaching of such bearing very susceptible to angular misalignment with respect to the bearing OD. As a result you get wobbly bearings. Also, the clearance required to get the 6 faces of the hex fitting nicely in the hex hole makes it so the shaft does not fit tightly in the bearing.

Nowadays, teams with access to a lathe and mill avoid hex bearings by using a round shaft and making 7/16 hex out of the live axle ends allowing them to use 1/2" round bearings; a standard in industry. The problem with this is that it requires some mill work.

As the WCD style of cantilevered live axle is prevalent in FRC today, many more teams want to use this style of drive. This is where the ThunderHex comes in. Teams without a mill can now do the exact same thing teams that have this access can do: use a round bearing with hex stock. Teams with a lathe can now use standard hex stock and turn down some portion of the stock to round off the hex corners to get the diameter of the ThunderHex bearing. If these teams want a press fit, then they can leave the shaft diameter a little big. If they want clearance, then they can do that too.

VEXpro has taken it one step further and given FRC teams the ThunderHex shaft for teams that do not have a mill or lathe. These teams can simply cut off the Thunderhex stock to length, buy a 1/4-20 (or 1/4-28) tap and tap the ends of the ThunderHex shaft and poof, you have a simple way to implement WCD axles without a mill and without a lathe.

Below are some pictures to compare what I am writing about.

The first picture is a 1/2" round shaft milled at the ends for 7/16" hex. The tube is a 2 x 1 aluminum tube common for WCD style chassis. This implementation requires some level of mill work and a mill operator skilled enough to manufacture the hex.

The second picture is a 1/2" hex stock with part of it turned down to fit a ThunderHex bearing. The one end acts as a stop for the shaft and is a full 1/2" hex for those of you worried about the rounded ThunderHex shape reducing wheel torque. This requires a lathe to turn down the standard hex to ThunderHex.

The third picture is ThunderHex stock used with the ThunderHex bearings. There is no mill or lathe work required, which makes a WCD style drive with round bearings accessible to teams whether or not they have a lathe or mill.

Sure, if your team resources allow you to mill a hex and you want to use 1/2" round bearings, then ThunderHex isn’t for you. However, here are some interesting facts as compared to 7/16" hex.

Even with the hole, the 1/2" ThunderHex will see approximately 62% of the bending stress that the 7/16" hex will see given the same bending moment.

Even with the hole, the 1/2" ThunderHex will see approximately 62% of the torsional stress that the 7/16" hex will see given the same torque load.

For every lb of 7/16 shaft, ThunderHex will weight 1.093 lbs.

I hope this sheds some light on why we developed ThunderHex.

Paul

image (2).png
image (2).png537×860 39.4 KB



image (2).png
image (2).png537×860 39.4 KB

2

I am very glad you made this available. ThunderHex bearings partnered with turned down 1/2" shafts should be really useful for us. Other than sounding amazing, is there any story to the name?

3

Paul, thanks for developing another great product.

4

This is an exciting product.

I’m sorry I didn’t think of it. :slight_smile:

I’m looking forward to trying it out.

5

What a great alternative for teams without machine resources. Thank you for the clarification, this is something my team can definitely look into. Thanks for all you do for the FRC community.

6

Thanks a lot for this. This is a really tremendous idea. I’ve hated sloppy hex bearing fits, and this seems like it should solve the problem.

We plan on replacing our hex bearings on our bearing blocks with these, and just turning down our hex shaft.

The 13.72mm bearing seems to be the perfect size. You still get plenty of contact area on the bearing (http://content.vexrobotics.com/vexpro/pdf/VEXpro-ThunderHex-Drawing-20141216.PDF), and it’s still a hex, so you can throw a gear/sprocket on it, and not worry about the shaft breaking.

The clearance hole also lets us tap these for 1/4"-20’s, but doesn’t really weaken the shaft much.

This is probably the most excited I’ve been for an uncommonly sized bearing.

7

Paul, assuming your post was prompted by my earlier post in the reveal thread, I will admit I did not take any time whatsoever to compare strength of a 7/16" vex to strength of the ThunderHex profile. Clearly, you have. I didn’t intend to diminish the product in any way. I get how it works, and this will be a great advancement for many teams.

You may even make a customer out of me on this stuff. :slight_smile:

Can anyone comment on using washers and bolts into the ends of tapped axles to secure wheels and sprockets? The one time we did it in 2011, we had some of them unscrew over time. I’m assuming a nylon patch screw wouldn’t.

8

If you’re already making snap ring grooves, why switch?

9

That is how we have been doing it for many years now and yes a nylock bolt is the way to go. Depending on what wheel you are using a medium strength locker is another good option. The only problem with the standard thread lockers is that they will weaken some plastics including polycarb.

10

We did this last year, and it worked well. We tapped the hole using the lathe to keep the part/tap lined up nicely, and we put on some red locktite (the stronger one), and used a normal 1/4-20 cap screw. We chose this because it was difficult to slide a snap ring over the hex shaft.

11

We surprisingly had no problems with securing wheels like this on our WCD last year with a little bit of blue loctite. It might of helped that we tried to keep the shafts length tolerances in the negative direction so the screws would take up slop to prevent any ratcheting action. Snap rings are still significantly faster to manufacture.

12

Any idea when the Thunderhex bearings will be available?

13

I would not recommend using the high strength locktite unless you don’t want to be able to service it. On an Allen head bolt you will probably strip the bolt or the wrench trying to remove it w/o heating up the locktite with a torch which can destroy the items around it as well as the fact that you can’t use a torch in the pits.

Yes cutting a grove is quicker than tapping but you need a lathe to do that while with the pre drilled hole in the Thunderhex you can just tap and go. Plus I find bolting easier than using retaining rings since they can shoot off into never never land at the worst possible time.

I do recommend that you make you shafts +0/-.00x in length so that the bolt pulls every thing tight if that is a concern. Having them a tiny bit long isn’t really a problem though.

14

I’d actually recommend not using the bolt to pull everything tight because that will unnecessarily load the bearings.

15

Paul,

How does the internal bearing race handle the concentrated loads from the rounded hex tips? Does this shorten the bearing lifetime, increase friction, ?

Thanks!

16

If you can’t wait for ThunderHex, and really want to use round bore bearings with hex shaft, you can always use a trick I used to use on 100.

Vex Pro sells metric roller bearings (217-2741) that have an OD smaller than the standard R8 size, and a bore larger than the distance across a hex’s flats. We machined inserts for the bearing bore with a 17mm press fit OD and a 1/2" hex ID. If you were to machine a 1.125" OD 26mm ID adapter and press the VP bearing into it, you could basically get a R8 hex bearing where your actual bearing has a round, nonbroached ID.

Even though we put them in applications were they saw a lot of abuse (worm gearboxes), I can’t make any recommendations about them being stronger/weaker than existing hex bearings. Because the bearing’s race is thinner, I suspect the -2741 bearing is weaker than a hex broached bearing and almost certainly weaker than ThunderHex. YMMV.

That said, it was a cool trick that ended up being pretty useful in some designs. And, if you have a lathe and hex broach, it’s a lot cheaper than ThunderHex.

17

Neat idea!
Alternatively, you can make a rotary broach holder for the lathe and buy a 3/8" hex rotary broach from littlemachineshop forless than the cost of a 3/8" broach.

18

I would wager 99.9% of teams can’t just casually make a rotary broach holder.

19

And if you can, why would you not just make the broach as well?

20

For the sake of discussion, opportunity cost…a rotary broach holder is over 10x more expensive than a rotary broach. Doesn’t make sense to invest that much effort for a $65 tool.

At the same time I don’t see the value in making the holder, as it will be worse quality than a COTS one and not work as well, while requiring significant time investment.