ThunderHex inspired question

I mentor a team here in Israel and we have recently come across how expensive it is for us to order many common parts from the states (shipping and import taxes can double the price on anything we buy.)

A few days ago my team was planning an order for parts to build our first in-house designed drive train (simple WCD), and I came across the thunderhex line of products.

Now, the cost of bearings really add up with the taxes, and seeing as the thunderhex bearings are metric sized, I thought I would be able to find them locally, but I didn’t. Instead I came across 14mm ID bearings. Now, the thunderhex ones are supposed to be 13.75mm. When put to CAD, the difference between them looks very small. I was thinking of taking a regular hex and using a lathe to round the corners to a 14mm radius and using 14mm bearings in the same way.

Would that work properly in the same way that Thunderhex works? My basic math makes it seem like I still keep 80% of the contact surface with the bearing. I’m not sure if that is enough, or even what the adverse effects that too little contact surface would cause.


If you’re already thinking of turning a hex shaft down to fit into a round bearing, I’d recommend using a small enough bearing that you get 100% contact. The sheet metal on bearings is mostly meant to hold the bearing together when it is between assemblies, not to distribute the forces tangentially.
A less desirable alternative would be to add a spacer with a hex bore and an outer diameter that fits into a larger bearing. This would also need to be held in place with a shaft collar or similar.

The difference between what you’re planning and a thunderhex bearing is .25mm or .009 inches - the thickness of two or three human hairs or two sheets of paper. This difference is pretty small and works out to around 4% less contact area than a normal thunderhex.

Although bearings were intended to have 100% surface contact, plenty of teams have used thunderhex without failure and the same concept before thunderhex came out with even less contact area.

However, I only have anecdotal evidence and intuition - I don’t have any solid math or analysis to assure you it will work, but I think an offseason project is the perfect way to test it.

This would entirely defeat the purpose of ThunderHex. The idea with ThunderHex is to have a round bearing, yet still be able to drive hex components of the same size on both sides of the bearing. An alternative that achieves full bearing contact and still allows for hex components on both sides of the bearing is to take a 1/2" hex shaft and turn it down to 1/2" round for the width of the bearing, and then machine a 7/16" hex geometry on the other side of the bearing. This takes intensive machining resources for the shaft however and an odd size broach for the 7/16" hex components.

It’s important to understand the design criteria likely used to develop or decide on the size for ThunderHex bearings and shaft stock. In a typical West Coast Drive, bearings for a press fit in a 1.125" hole are used. When I looked at metric bearing sizes, I see that 13 mm ID ball bearings generally are a good size for a 1.125" press fit and 14 mm ID ball bearings are almost exclusively well sized for a 1.375" press fit. This indicates to me that somewhere in between 13 mm and 14 mm a transition in acceptable press fit holes for the bearing size occurs. A 13 mm hole would be unacceptable because ThunderHex with a 0.512" round bearing wouldn’t really work that well (barely any hex shape at that point), and I guess they felt 13.75 mm was an appropriate size and found that they could source bearings of that size which would press fit nicely in a 1.125" diameter hole at a reasonable price.

In addition, I’m fairly certain teams have used 3/4" hex with far less bearing contact and not had issues (Thunder Chickens --> ThunderHex???).

All this to say that I doubt it makes much difference from a loading stand point if you go with 14 mm bearings. As long as you are willing to use 1.375" bearing holes it should work out fine. It’s just different than what VexPro came up with to make ThunderHex.

I really haven’t figured out any real advantage to thunder-hex for teams that already have a lathe and can custom machine any shaft that they might need. Vex does sell 13.75 mm bearing to go with the thunder shaft however.

If you have a local source for hex shaft (preferably 0.500 Inch) for compatibility with Vex/AM parts, it would be trivial to machine that down to .500 round or any metric size smaller that you want to use for bearings.

We have an on-line metals warehouse nearby, and simply pickup all our hex shaft there.

Comment- 6061 is fine for most shafting. We normally use 7075 as its it’s yield strength is much better (73Ksi vs 40Ksi).

My understanding is the ThunderHex was designed for teams who don’t have access to a lathe (or don’t want to invest the time on that project).


It’s still really awesome for teams with lathes though! We have three lathes in house, and we still used thunderhex for most of our torque shafts last year. We plan to use it for even more going forward.

It’s great stuff, we love it.

I’d be cautious recommending this, what is “most shafting” to the average FRC team? Length between bearings, load applied, whether or not the shaft is cantilevered, etc… all will greatly affect what is required.

For normal FRC designs such as Gearbox shafts, intake rollers etc. with some reasonable designed bearing supports, torque loads only, and shorter shaft distances 6061-T6 is more than adequate. We do typically use 7075-T6 since it is readily available.

7075-T6 is not a solution to poor engineering design; with long unsupported cantilevered loads. Aluminum Alloys have only moderate Fatigue strength and shouldn’t be used in this manner. For abusive situations like this a Steel hex shaft would be advisable.

Yield Strength 40KSI
Fatigue Limit 14KSI

Yield Strength 73KSI
Fatigue Limit 23KSI

4140 Heat Treated
Yield Strength 130KSI
Fatigue Limit 94KSI

This is still way too general/suspect advice. I can think of hundreds of gearboxes teams have made with short, well supported, only torque loaded shafts that would immediately fail if they were 6061.

With the prevalence of thunderhex and regular hex from VEXpro, there is really no reason to be using 6061 shafting, even if the application can handle the loading.

I thought aluminum doesn’t really have a fatigue limit. You load it in a reversing stress, it will eventually fail. (Granted maybe long past an FRC bots life time.) One reason air frames have a service life with inspections.

Maybe some one from VEX could comment, but didn’t they switch a few years ago from 6061 to 7075 shafting because of fatigue issues?

One reason to use thunder hex even with a lathe available is you can replace a failed shaft in the field by cut to length. A general efficiency in production rule is don’t custom make when an off the shelf product is available. OK two reasons.

Although the thunder hex bearings are sort of a odd size, I like round bearings better than hex bearings since you are not putting a stress riser in a brittle bearing race.

VEXpro Hex shafting has always been 7075-T6 Aluminum. The 1/2" tube axle on the other hand (217-2762) is 6061-T6 due to lower loads (double supported with a bolt through it) and never being used to transfer torque.


VEXpro has only offered 7075 hex.

The maximum torque capability for a 0.5" hex shaft is = 0.0262 * Yield Strength.

A 6061-T6 shaft can theoretically take 1048 inlbs (or 87 ftlbs)
7075-T6 shaft is likewise is 1912 inlbs (or 159 ftlbs)

You would be far more likely to break some FRC gearbox teeth before you ever failed these shafts in torsion.

As someone who has ran machines in an ME stress lab, you’d be shocked how many rotations a two foot bar stock can make before it ultimately fails! (Tinius Olsen Machine)

My mistake. Thanks for the correction.

See day 36 of JVN’s 2010 build journal for an example of a well supported aluminum shaft that failed due to pure torsional load.

The 1912 in lbs you quote is only valid if this torque is gradually applied. The shaft can fail earlier than that with a quickly changing load.

To add another data point, my old team used 6061 in drive shafts back in 2011 and we had several failures, most often at the center shaft between the gearbox and the power transmission. It is really not too difficult in a drivetrain to fail 6061 hex shaft and thus I really don’t think there’s much use for it when 7075 is so readily available.

To answer the OP’s question, as to why not just use 14mm bearings with hex shaft turned to 14mm max diameter a la ThunderHex. That’s a perfectly viable solution and you’ll probably find success if you give it a try. The reason that regular 14mm metric ID bearings aren’t used with the ThunderHex system is that these bearings aren’t readily available in the FRC psuedo standard 1.125" OD for obvious reasons. I’m guessing the designers at Vex realized they had to make custom bearings anyway, and thus decided to go with a slightly more optimal but less standard size for the ThunderHex ID.