Upright bearing housing optimisation

[Cucumbers]

Hi all,

I'm currently doing some preliminary FEA analysis for next year's uprights, and I was wondering if anyone had some advice or good resources on optimising the thickness of the bearing bore wall.
Seems to me that in addition to hitting toe and camber stiffness goals, the bearing bore wall needs to be thick enough to prevent bearing failure that would be caused by distortion under load - something there doesn't seem to be a lot of discussion about, in spite of the seemingly hundreds of threads on upright design in general!

I've done a fair bit of research, but most bearing design manuals seem to assume that your housing will be massive enough to have effectively no deflection.
The approach I've settled on for the moment is ensuring that the housing does not exceed the manufacturer's tolerances for cylindricity or conicity (as usually measured statically, before bearing installation) under load.
The ANSYS setup includes a model of the bearing's outer ring with initial penetration into the bearing bore wall to account for both the stiffening effect and the preloaded hoop stress introduced by the interference fit.

As for the bearings themselves, we are planning on using cartridge-style double-row angular-contact ball bearing units, as commonly seen in OEM applications for small cars.

[BillCobb]

Much of your toe and camber compliance will be the bearing all by itself. Your bearing choice will need to be loose to prevent failure by the way it operates (1/4 to 1/2 flat back off). Use a different type of bearing and focus on the other elements of toe or camber compliance. Build a lab fixture to test your design.

A limp spindle is not a good idea either.

[Cucumbers]

G'day Bill,

I feel as though I've been given a zen koan to decipher!


Much of your toe and camber compliance will be the bearing all by itself. Your bearing choice will need to be loose to prevent failure by the way it operates (1/4 to 1/2 flat back off).
Why would it need to be loose? I was looking at using an OEM unit from a small car, so the preload would be set to something appropriate for this application (I would have thought). Will it need to be loose because the upright wouldn't be as stiff as an OEM application, where it is pressed into a steel forging?


Use a different type of bearing and focus on the other elements of toe or camber compliance. Build a lab fixture to test your design.
We were keen on this kind of bearing because it simplifies the manufacture of the upright: just one bore with a shoulder and a circlip groove. Additionally, the length of the inner races sets the preload, so there is no precise machining of a spacer or trying to tweak the torque of the stub-axle nut to get that right. As a team with minimal access to machining resources, both of these aspects are pretty attractive to us! It is (was?) also used in Taylor Race Engineering's FSAE setup ( https://taylor-race.com/sites/default/files/ACFD06.pdf ) and other small formula-style cars.
What kind of bearing setup do you prefer - a set of tapered roller bearings? Why is that - because it allows a stiffer overall setup through wider spacing? Are roller bearings less susceptible to failure through housing distortion than ball bearings?

Any thoughts on my general approach of attempting to limit housing distortion to the manufacturer's tolerance for out-of-roundness - reasonable?

I hope these questions aren't too basic. Thanks for your help!

[Z]

Billy,


...optimising the thickness of the bearing bore wall.

Grrrrr... Bad start there...

DO NOT "OPTIMISE"!!!

Just build something that works! Something that drives AT LEAST 30+ kms, etc.

But you know I was going to say that.
~o0o~

Anyway, IMHO your suggested "...cartridge-style double-row angular-contact ball bearing units" (DRACBBs) are a country mile ahead of the 69xx (light) or 68xx (extra-light) deep-groove-ball-bearings (DGBBs) that are so common in FS/FSAE. So you are ahead of the game there. But tapered-roller-bearings (TRBs) are another country mile further ahead. Stiffer, stronger, lighter+++.

And as I am sure you are aware, off-the-shelf (OTS) replacement TRBs for the two main types of defacto standard box-trailer-axles we have here in Oz are available VERY CHEAPLY (~$14.99?, for inner-TRB + outer-TRB + oil-seal + split-pin + tin-cap + fancy-packaging) at any of the super-cheap autoparts stores. Even way up in FNQ! :) And, (err..., sorry Bill), the Ford units are better than the Holden ones.

The larger of these TRBs (ie. the one fitted to inner-side of axle) has 35 mm ID (Ford), together with a large radius on its mounting side (= good). Two of these fitted between a suitable axle and upright would give you the stiffest, strongest bearing-assembly ever seen in FSAE, and within a few hundred grams of the lightest, floppiest units ever made. And you can use the smaller TRBs (22 mm ID) on another part of the car (steering-shaft?).

Or just buy OTS 30 ID x 55 OD TRBs (PN = ~32006), which are more than strong and stiff enough, and fit them to an OTS axle I have in mind...
~o0o~

Getting back to your DRACBBs.

I was in my local, friendly, town-mechanic's workshop a while back, and he was telling me about the front-wheel DRACBB he had replaced on a medium-size, FWD car. Seems the very gentle lady driver had hit a curb, quite gently, and there was "a horrible whining noise" coming from the front of the car (which is why the OEMs use these bearings. ;)) Two weeks after doing the replacement the lady was back, because the horrible whining noise was also back.

Assumption was that the new bearing was a dodgy one (Chinese junk!) so he replaced it with a more expensive brand-name part. And one week later the lady was back again, because, yep, the noise was back again. So this time he swapped the whole upright/bearing/axle-unit with a second-hand part, because he has these all over the place. And happy days! No more noise!

The assumption this time was that the upright and its bore had distorted during the curb-strike. I had a close look at it, but NO distortion or damage visible to the naked eye, or by using a steel-ruler. Friendly mechanic has no bore-micrometer, or even vernier-caliper, because he is not that sort of mechanic. So hard to say for sure. But conclusion is that it only take a few kilometers of running in a very slightly distorted bore for this type of bearing to, well..., let you know it is unhappy.

On the positive side, I know these DRACBBs will moan and whinge and whine for quite a long time before finally giving up. Contrarily, the 68/9xx DGBBs, so beloved by the rest of the FSAE world, will spit their balls out ASAP when put under any sort of adverse pressure.

And the TRBs? Well, they just go on forever. (<- Again, the reason why the OEMs tossed them.)
~o0o~

Also,
1. Is your bearing analysis for front-axle, rear-axle, or both ends of the car (ie. same design for driven and un-driven axles)?
2. What type of axle? OEM or DIY?
3. Will your uprights be billet-machined-aluminium (crap), or folded-and-welded-sheet-steel (best), or other (say, 3-D-printed-titanium, like the ones that failed on one of Monash's cars, albeit during early testing)?

[Edit:] Most importantly.
4. What is your mass target for upright+bearings+axle? 1 kg? 10 kg? More? Less? Inbetween?
5. What is your mass target for the whole car?
[End Edit]

Images of your ideas so far would help us help you.

Z

[Cucumbers]

Hi Z,

Sorry for the late reply, pointy end of the semester had me buried for a while there!


Billy,



Grrrrr... Bad start there...

DO NOT "OPTIMISE"!!!

Just build something that works! Something that drives AT LEAST 30+ kms, etc.

But you know I was going to say that.


Yes, I should have seen this one coming from you :D
When I say "optimize" I don't mean "to a safety factor of exactly 1, assuming we'll never hit a bump," I mean more along the lines of "figuring out how much weight I can reasonably shave off of the 1kg lump of steel that made up the bearing housing on the previous uprights..." We're not even in the ballpark of reasonable, so I want to know how to get there, basically.



Anyway, IMHO your suggested "...cartridge-style double-row angular-contact ball bearing units" (DRACBBs) are a country mile ahead of the 69xx (light) or 68xx (extra-light) deep-groove-ball-bearings (DGBBs) that are so common in FS/FSAE. So you are ahead of the game there. But tapered-roller-bearings (TRBs) are another country mile further ahead. Stiffer, stronger, lighter+++.

And as I am sure you are aware, off-the-shelf (OTS) replacement TRBs for the two main types of defacto standard box-trailer-axles we have here in Oz are available VERY CHEAPLY (~$14.99?, for inner-TRB + outer-TRB + oil-seal + split-pin + tin-cap + fancy-packaging) at any of the super-cheap autoparts stores. Even way up in FNQ! :) And, (err..., sorry Bill), the Ford units are better than the Holden ones.

The larger of these TRBs (ie. the one fitted to inner-side of axle) has 35 mm ID (Ford), together with a large radius on its mounting side (= good). Two of these fitted between a suitable axle and upright would give you the stiffest, strongest bearing-assembly ever seen in FSAE, and within a few hundred grams of the lightest, floppiest units ever made. And you can use the smaller TRBs (22 mm ID) on another part of the car (steering-shaft?).

Or just buy OTS 30 ID x 55 OD TRBs (PN = ~32006), which are more than strong and stiff enough, and fit them to an OTS axle I have in mind...


We've previously used TRBs, but the attraction of these cartridge-style parts is that the upright/bearing housing would require a single accurate bore, a shoulder, and a circlip groove. Compare to a set of TRBs, where you need 2 bores and shoulders, almost perfectly concentric, etc. Multiple machine setups. And then you have to worry about setting preload by either a very accurately ground shim, or by tightening the stub axle and wiggling the wheel until you've convinced yourself that you might have got it set to some compromise between friction and stiffness...
That probably seems like silly minutiae, but when you haven't got much access to machining resources and want to get a car done quickly, eliminating that extra hassle probably cuts a month off of production time...

The other thing with the trailer TRBs is that the smaller of the pair seems a bit slim to squeeze a stub-shaft through for a driven axle, and we'd like to use the same bearings on all four corners.

Very interested in your OTS axle idea!



Getting back to your DRACBBs.

I was in my local, friendly, town-mechanic's workshop a while back, and he was telling me about the front-wheel DRACBB he had replaced on a medium-size, FWD car. Seems the very gentle lady driver had hit a curb, quite gently, and there was "a horrible whining noise" coming from the front of the car (which is why the OEMs use these bearings. ;)) Two weeks after doing the replacement the lady was back, because the horrible whining noise was also back.

Assumption was that the new bearing was a dodgy one (Chinese junk!) so he replaced it with a more expensive brand-name part. And one week later the lady was back again, because, yep, the noise was back again. So this time he swapped the whole upright/bearing/axle-unit with a second-hand part, because he has these all over the place. And happy days! No more noise!

The assumption this time was that the upright and its bore had distorted during the curb-strike. I had a close look at it, but NO distortion or damage visible to the naked eye, or by using a steel-ruler. Friendly mechanic has no bore-micrometer, or even vernier-caliper, because he is not that sort of mechanic. So hard to say for sure. But conclusion is that it only take a few kilometers of running in a very slightly distorted bore for this type of bearing to, well..., let you know it is unhappy.

On the positive side, I know these DRACBBs will moan and whinge and whine for quite a long time before finally giving up. Contrarily, the 68/9xx DGBBs, so beloved by the rest of the FSAE world, will spit their balls out ASAP when put under any sort of adverse pressure.

And the TRBs? Well, they just go on forever. (<- Again, the reason why the OEMs tossed them.)


The mechanic story doesn't surprise me at all... Going off of the SKF bearing manual, this class of bearing needs a housing that is within 0.0065mm of perfectly round...



Also,
1. Is your bearing analysis for front-axle, rear-axle, or both ends of the car (ie. same design for driven and un-driven axles)?


Ideally we'd use the same bearing and hub on all four corners, just with different stub-axles front and rear.



2. What type of axle? OEM or DIY?


Ahh yes, you've hit on another big motivation here. Using an OEM wheel bearing from a small car means we can also make use of an OEM hub. This is a big advantage to us as we can't cut internal splines locally/affordably, and I consider one-piece hubs with integrated CV housings to be very complicated and expensive, without much benefit.



3. Will your uprights be billet-machined-aluminium (crap), or folded-and-welded-sheet-steel (best), or other (say, 3-D-printed-titanium, like the ones that failed on one of Monash's cars, albeit during early testing)?


As of yet undecided. I'm a fan of fabricated sheet steel, but we have less access to manual equipment (Uni OHS: Dangerous pointy bits!!!) than we do to CNC equipment (Uni OHS: Ahh, a nice locked box of safety...)



[Edit:] Most importantly.
4. What is your mass target for upright+bearings+axle? 1 kg? 10 kg? More? Less? Inbetween?


At the moment, unknown. Still in the early stages of investigating. All we know is our old designs are very overweight, very complicated, and very expensive. To be honest, I'm more focused on the "complicated and expensive" part of that equation, for the moment.



5. What is your mass target for the whole car?
[End Edit]


Again, still nailing down specifics, but around 200kg is likely what we'll shoot for (600cc 4cyl, space frame, 13" wheels).



Images of your ideas so far would help us help you.

Z

Nothing worth showing yet! But when I do have something, I'll share.

Sorry for the lack of detail... But we are very early in the process. This is one of the first things we are considering, because the upright/hub/axle assemblies have been a major snag in previous designs, wrt cost and manufacturability. Very open to any comments or suggestions.

[Z]

Billy,

I have checked my old notes, and they have your 2014 car as ~260 kg overall mass, and DNSs for all dynamic events. So, to make "massive progress" in 2018 you only have to build a car that "completes all dynamic events". Even just completing Enduro will put you a long way ahead of your last car, and will set your team up for more success in later years.

So it is enough for now that you simply aim for an overall mass of, say, sub-250 kg, but with the OVER-RIDING GOAL that the car is ROCK-SOLID RELIABLE, and RUNNING long before December! Preferable first-drive around July.

So you can now break your 250 kg (maximum) overall mass down into:
1. Front-Axle (2 x 13" tyre, wheel, hub/bearing/axle, upright, brake, suspension-bits) = 50 kg.
2. Rear-Axle (as above) = 50 kg.
3. Engine = 50 kg.
4. Chassis/bodywork = 50 kg.
5. Everything else = 50 kg.

Unfortunately, you have chosen, or been given from the 2014 car (?), a ship-anchor for an engine, so you are over the mass budget there and will have to borrow some kgs from elsewhere. Perhaps from the "Everything else" section?

Fortunately, you have 25 kg to spend on each corner. Even with the heavyweight 13" wheels (again, I guess from the 2014 car?) this gives you a fairly large mass budget for the hub-bearing assemblies.

Incidentally, this is the correct way to start addressing these detail design issues, such as "Which wheel bearings?". Always best to start with the high-level objectives first, then work down to the details. (My apologies for almost missing this in my last post.)
~o0o~

To repeat my above comment, the over-riding goal of these hub-bearings is that they are RELIABLE and enable an EARLY BUILD FINISH.

So here are three options, in addition to the DRACBBs you have already considered.

1. A "Generation-3" hub-bearing assembly from the driven wheels of a modern small to MEDIUM sized car. These Gen-3 units have the wheel-mounting-flange, which is integral with the axle and bearing-inner-races, on their outboard-side. On the inboard-side they have an upright-mounting-flange, which is integral with the bearing-outer-races. Seals are built-in. No user serviceable parts.

So you only have to choose a unit with your preferred wheel-bolt-pattern, perhaps that suits the wheels you already have, and on the inboard-side there is a 3 or 4 bolt flange that bolts to your upright. So no worries about circularity of the bearing-housing! (BTW, the DRACBBs you are considering are "Gen-1".) Avoid buying units with built-in ABS sensors, because they co$t much more.

Also, I have in my "Museum of Prior Art" (= junkyard out back) some Gen-1 DRACBBs for a small, ~900 kg, car designed ~50 years ago, that are 40 mm ID x 80 mm OD. This size is nowadays fitted to large sedans of 1,800+ kg! Modern "small" cars have much smaller units. My point is that modern bearings are much closer to the "limit", and may fail you at the worst possible time.

To stress yet again, your over-riding priorities here are that the units are strong enough, and are at a reasonable price, and are available IMMEDIATELY.

* DO NOT order "optimal" units that need shipping and "... should be here at the end of the month". (What month!!!)

* And DO NOT FUSS OVER AN EXTRA ~KILO, since you should be able to get under that 25 kg/corner budget quite easily.
~o0o~

2. Find similar OTS units, but this time taken from a LARGE ATV/Quad-bike. Again, the "rightest" unit is the one that is right in front of you, so immediately available. And, of course, is STRONG enough and at the right price.
~o0o~

3. This next option is the funnest, because you get to climb around car wrecker's yards. Here are three from my MuseumOPA.

3a. Subaru 4WD "Touring Wagon" from ~1980s, front-axle. The axle itself is a solid bar about 20 cm long, 35 mm OD, and runs in 2 x 6207 DGBBs. (Note the second digit "2" = heavy-duty, while "3" = extra-heavy-duty, "0" = standard, and 8 & 9 as described before). These 72 mm OD, ~290 gm, DGBBs are strong enough for the front-heavy Suby, so good enough for you. Mine only failed because the seal failed -> let water in -> rusting/pitting -> loud whining noise. The axle has a Rzeppa-CV-outer-race integral with its inboard-side, and a spline onto which the brake-disc/wheel-mounting-flange is bolted on its outboard-side.

3b. Same Suby but rear-axle. This is again a solid bar about 20 cm long, ~30 - 35 mm OD (I would have to check), but this time running in 2 x TRBs (more details below). Inboard-side is splined for a Lobro-CV (= 6 balls in angled runways), and outboard side is splined for the brake-drum/wheel-mounting-flange.

3c. VW Beetle "IRS" from ~1969+. Note all Beetles have Independent Rear Suspension, but the earlier ones are called "Swing-Axles", and the later ones "IRS"s or "double-jointed" axles. The IRS axle is again a solid bar ~20 cm long, 30 mm OD, has an integral flange on the inboard-side for a Lobro-CV, and a longish spline on outboard-side for the brake-drum/wheel-mounting-flange. The bearings are a single 6206 DGBB (= 62 mm OD, ~200 gm) on inboard-side, and similar sized Cylindrical-Roller-Bearing on outboard-side. The CRB sits at the centreplane of the wheel, so carries most of the Fz load. See one of the Carroll Smith books to see that this is how real racecars also did it back in those days. But I would fit 2 x 30 mm ID TRBs, as explained below.

Key point here is that these Beetle axles are readily available, BRAND NEW, from many suppliers all around the world.
~~~o0o~~~

Some more comments on how to put it all together.

4. At this stage you should also be thinking about the driveshafts and CVs you will need. It would be easiest if you could use the matching shafts/CVs that go with any of the above options. But likely is that they will be either too long or too short for your preferred track-width. No big problem, because the shafts can be cut and rewelded. There should be someone up your way who does that.
~o0o~

5.
... TRBs ... need 2 bores and shoulders, almost perfectly concentric, etc. Multiple machine setups. And then you have to worry about setting preload by ... a very accurately ground shim...
That probably seems like silly minutiae, but when you haven't got much access to machining resources and want to get a car done quickly...

You are OVER-COMPLICATING IT!!! :)

After numerous laps of the island in that Suby-4WD, one of its rear-wheels started to show excessive play. I was in a remote part of Oz, with no local Suby dealer, and no workshop manual, so I just opened her up. The TRBs looked good, with no excessive pitting. So I used my trusty 4" angle-grinder to shorten the spacer a bit, with a steel square to check "accuracy". Reassemble hub, spin to check for right preload, disassemble and grind off a bit more. Repeat until "optimal". All up, it took about half an hour. And that bearing was good for another couple laps of the island, and is still good now!

Main point here is that I recall taking about a millimetre off the spacer. That's a lot! And the car was driveable before that millimetre came off.

Think about the bearing-nut on a box-trailer-axle. It has a thread pitch of ~1.5 mm, and is then typically locked by the split-pin every twelfth of a circle. So your FINEST preload adjustment is 1/8 mm, or 0.125 mm. That is a heavy cut when facing-off the end of a spacer on a manual lathe. In short, you DO NOT NEED "micronic" accuracy.

Also, "correct preload" is such that when you spin the torqued-up hub with your finger, it rotates less than one extra revolution after you release your finger. If it stops immediately, then too tight. If it keeps spinning "freely" for many revs, then too loose.

All this is easiest shown with a few 30 second lessons in the workshop, with practice inbetween. Then pass knowledge onto younger team members, and build a winning advantage into your team culture.
~o0o~

6.
... I consider [typical FSAE] one-piece hubs with integrated CV housings to be very complicated and expensive, without much benefit.

Very true!
~o0o~

7.
I'm a fan of fabricated sheet steel, but we have less access to manual equipment (Uni OHS: Dangerous pointy bits!!!) than we do to CNC equipment (Uni OHS: Ahh, a nice locked box of safety...)

Since you are building a spaceframe car, you will necessarily be doing a lot of "sheet-steel fabrication", albeit with the sheet-steel pre-rolled for you into tubes. It is to your team's great advantage to keep improving your skills with this process, and then keep using this same process for as many other parts of the car as possible.

In round numbers, ~98% of the "manufactured" parts of the car can be made from folded and welded sheet-steel. Stronger, stiffer, lighter, cheaper, faster. Only very few parts need to be "machined", and these can be done on a manual lathe.

Sadly, the more FSAE cars I see, the more I am convinced that it is the teams' obsession with CNC/billet-machining that is holding them back. For example, the multitude of "billet-machined rear-bulkheads". Utter junk!
~o0o~

Enough for now.

Z

[BillCobb]

When will we start seeing a 3D Printed car ? And I want to see the driver done this way too, extra points for creativity.

[Claude Rouelle]

Bill, University of South Dakota had their own 3D titanium uprights 10 years ago.

[Z]

Bill and Claude,

From my first post above.


3. Will your uprights be billet-machined-aluminium (crap), or folded-and-welded-sheet-steel (best), or other (say, 3D-printed-titanium, like the ones that failed on one of Monash's cars, albeit during early testing)?

IIRC, those 3D-Ti uprights were part of a sponsorship deal with the CSIRO, Australia's premier research organization!

Go Aussie innovation!!! :D

Z

[Cucumbers]

Hi Z,

Thanks for your comments! I'm totally with you on pretty much all of this.



[...]

Incidentally, this is the correct way to start addressing these detail design issues, such as "Which wheel bearings?". Always best to start with the high-level objectives first, then work down to the details. (My apologies for almost missing this in my last post.)


Yes, this is what we'll be doing (and I've read all the threads!). If this thread seems like I'm hyper focused on a particular small issue, that's because 1) it's a detail that has caused problems for us in the past, and 2) because there was an assignment that gave me the chance to think about the bearing/hub combination in a fair bit of detail. So I was mostly thinking in an abstract, technical sense about how you tackle this problem. We'll be doing big-picture design before things start in earnest, I promise!




To repeat my above comment, the over-riding goal of these hub-bearings is that they are RELIABLE and enable an EARLY BUILD FINISH.

So here are three options, in addition to the DRACBBs you have already considered.

[...]

* DO NOT order "optimal" units that need shipping and "... should be here at the end of the month". (What month!!!)

[...]

Key point here is that these Beetle axles are readily available, BRAND NEW, from many suppliers all around the world.


I'm with you on this, in fact, it's what drove me to pick out the DRACBB cartridge bearings: I walked into Repco and asked for a wheel bearing from the lightest commonly available car I could think of, and then walked into a dealership parts department and did the same for the hub. It's just that now it seems like DRACBBs are fussier than I'd like, although it's still not clear to me exactly how much momentary out-of-round conditions will shorten their life, given that the bearing manuals and your mechanic story talk about a permanently distorted bore. The fact they're cheap and easy to come by, and start to make noise long before they fail makes me tempted to carry on with this plan, but I'll definitely have to go and crawl around the wreckers to see if there's anything better on offer...



You are OVER-COMPLICATING IT!!! :)

After numerous laps of the island in that Suby-4WD, one of its rear-wheels started to show excessive play. I was in a remote part of Oz, with no local Suby dealer, and no workshop manual, so I just opened her up. The TRBs looked good, with no excessive pitting. So I used my trusty 4" angle-grinder to shorten the spacer a bit, with a steel square to check "accuracy". Reassemble hub, spin to check for right preload, disassemble and grind off a bit more. Repeat until "optimal". All up, it took about half an hour. And that bearing was good for another couple laps of the island, and is still good now!

Main point here is that I recall taking about a millimetre off the spacer. That's a lot! And the car was driveable before that millimetre came off.

Think about the bearing-nut on a box-trailer-axle. It has a thread pitch of ~1.5 mm, and is then typically locked by the split-pin every twelfth of a circle. So your FINEST preload adjustment is 1/8 mm, or 0.125 mm. That is a heavy cut when facing-off the end of a spacer on a manual lathe. In short, you DO NOT NEED "micronic" accuracy.

Also, "correct preload" is such that when you spin the torqued-up hub with your finger, it rotates less than one extra revolution after you release your finger. If it stops immediately, then too tight. If it keeps spinning "freely" for many revs, then too loose.

All this is easiest shown with a few 30 second lessons in the workshop, with practice inbetween. Then pass knowledge onto younger team members, and build a winning advantage into your team culture.


This is good perspective! There's obviously a big difference between what bearings require in the textbook and how they fare in real world conditions - which tough to figure out when you haven't done much real world stuff yourself!

Cheers!

[Z]

Billy,

Oops! The units I described above with flanges both inboard and outboard are called "Generation 3".

Z

Some images.

https://www.timken.com/wp-content/uploads/2017/07/Gen-3-Hub-340x257.jpg

http://www.gmb.jp/assets/templates/palette-en/img/product/axle/wheel_img05_1.gif

http://www.gmb.jp/assets/templates/palette-en/img/product/axle/wheel_img07_1.gif

[JSR]

For a build I'm doing I will use units from Lotus Elise S2. Attaches to the upright with 3xM12 bolts, and has a 4x100 bolt pattern.
If you want I can send you a CAD-model of them.