HAW Amberg-Weiden front "upright" design

[Big Bird]

I just wanted to draw attention to some neat design from HAW Amberg-Weiden – Running Snail Racing, and their novel front “upright”. I really like the load path from contact patch through to a-arms.

https://www.youtube.com/watch?...6yTkw&feature=relmfu (https://www.youtube.com/watch?v=9rbw-_6yTkw&feature=relmfu)

The drawback – high bearing speeds – but less of an issue in FSAE than other motorsport.

The below is pasted from the centrelock wheels thread. The flying snails have gone and built what I was too lazy to even sketch:
“I generally don't like centrelocks, as the loads from the contact patch all need to be transmitted through the interface of the wheel centre and whatever it is attached to (axle, hub, whatever). The closer to the axle centreline, the higher the contact forces. In the case of a centrelock, we are taking tyre forces from a large diameter, transferring them as moments to a small diameter axle (and therefore high contact forces) and then feeding the forces into the uprights via bearings. I just think the larger the diameter where wheel loads are transferred into the hub, the lower those forces and therefore the less material required.
We need a draw function on these forums so I can explain myself, and I couldn't be bothered drawing something and scanning it and then hosting it somewhere.”

Neat work, I hope it was well received.

[Carrington]

Didn't Western Washington have a wheel-mounted rotor setup one year?

[Markus]

Regarding the quote on the first post:

While the hub bearing chosen is a big limiting factor on tire to a-arms load path I don't really see a difference between a centerlock and multilug fitment. You still have to path the loads through the bearing so I can't see you getting away with that much less material on either of the options.

Simply put there's only a difference where the path diameter reduction is done: wheel center vs hub.

I do have to say I like Amberg-Weidens design. I wonder how they made the rear? (Will look for pics after work...)

[Big Bird]

What I was trying to say, although struggling to do so in the first post, is that a centre lock hub constrains you to some sort of axle/hub arrangement at the axle line of the wheel. So force generated the contact patch gets transferred to the central axle and then is spread back out to the a arms via the upright. It's a fairly circuitous load path and you get some pretty major stress concentrations at the centre of the wheel/bub.

A multi-lug wheel "centre" offers a little more design freedom to space those lugs right out and run some form of large diameter hollow hub and large diameter bearing, like what these guys have done. In my mind I had half a picture of a hollow ring shaped upright with bearing inside it and a large diameter hollow hub bolting almost directly to the outer rim. These guys have reversed that and put the bearing on the outside of the upright and come up with a much cleaner design than my half - formed thoughts.

The load path in this case is much cleaner and more direct than a traditional wheel and hub, as we are transferring a force generated at the out side of the wheel to a-arms also near the outside of the wheel, without the intermediate step of passing through the central axle - with the high stresses that come from this.

I've probably explained that no better this time, so apologies if it all sounds like nonsense.

[Z]

Geoff,

It looks a lot like the Citroen DS front "upright". Especially the later (1960s) 5-stud type, rather than the earlier (1950s) centrelock type.

Except the HAW upright looks a lot flimsier. Understandable, given relative loads.

And its steer-arm is a lot shorter. Err, maybe too short...

And it has that funny little vertical stick in the middle.

And the disc brake is "inside", rather than "inboard" like on the DS.

And I guess it has a horribly expensive thin-ring ball bearing. Or 2? Type???

Z

[mech5496]

Markus the rear used a conventional setup. Z there ARE some thin-section ball bearings in there, which are surprisingly small. I have some high quality closeups of their setup, although with the wheels on; I would prefer some with the wheels off!

[Goost]

We used to joke about using slewing rings in an upright setup, didn't look into it hard when we found the diameters on the order of meters across, but if they make them this small there may be some potential there...
Very impressive work Amberg-Weiden!

[EPMPaul]

There's a photo of the setup without the wheel at the following address.

here (http://www.formulastudent.com/learn-to-win-blog/learn-to-win/learn-to-win---the-judges'-advice-blog/2012/07/24/fs2012---design-round-up)

[mmw2753]

Originally posted by EPMPaul:
There's a photo of the setup without the wheel at the following address.

http://www.formulastudent.com/...-to-win---the-judges (http://www.formulastudent.com/learn-to-win-blog/learn-to-win/learn-to-win---the-judges)'-advice-blog/2012/07/24/fs2012---design-round-up

working link http://tinyurl.com/ccwayrj

[Do-it-Guy]

According to an interview given by one of the Amberg-Weiden drivers to the livestream commentators at Silverstone, the bearing is a miniature version of a medical CT-Scanner bearing, specifically built by the manufacturer company for the Amberg-Weiden car.

CT-Scanners aren't spinning very fast (~300 rpm), but this much smaller version should be able to handle the (quite slow) speed of an FSAE car.

I don't know if they can answer the "Does it really make you faster?"-question by the judges with "yes", but i love the very cool looking design! Can't wait to have a close look at it on Sunday at Hockenheim!

[Richard Pare]

If I were judging, I would have failed them for their "only 18% more" cost claim - BS of the highest order!

[Charles Kaneb]

Why?

The bearing itself might cost more than a conventional pair, and by more than 18%, but if the main segment of that upright is a standard aluminum extrusion of some sort it'll take a lot less machine time and cost than an all-machined upright. If it's an aluminum tube of some standard size with the other parts bolted or welded to it it might be even cheaper.

[Dunk Mckay]

If it's 18% more with the custom bearings than the standard setup costs on the rear (minus drivetrain of course) then their rears cost way too much.

[mech5496]

Added to what Charles says, we all forget the lack of a hub and the wheelcenter, so this might add up to the comparison.

[Richard Pare]

Geeze, guys - go out and actually try pricing ANY bearing that is anywhere near that ID. Even a production Timken 6" ID taper roller bearing will run nearly $1000. Special made, low production bearings can easily run multiples of that.

[jamesfhammond]

Originally posted by Richard Pare:
Geeze, guys - go out and actually try pricing ANY bearing that is anywhere near that ID. Even a production Timken 6" ID taper roller bearing will run nearly $1000. Special made, low production bearings can easily run multiples of that.

In the scenario of making 1,000 cars per year, it's pretty realistic to think the fictional company would be ordering 10,000 bearings per year, bringing the price down drastically.

Assuming 2 per front corner, that is 4,000. The other 6,000 would be spares to sell to customers. I don't know how often these bearings would need to be serviced/replaced during a race season.

[Richard Pare]

And just what price do you think a bearing that size will be, even at a 10,000 unit production level?

What production level do you think Timken does for that $1000 bearing?

Further, a company would have to be nuts to purchase 1 1/2 times their needs on an item that most likely will not need replacing very often.

Remember, most bearings used for these types of cars will cost well less than $100 each, even bought 1 at a time.

[Z]

I' m with Richard on this one. Any "bought" bearings of that size would be far too expensive for "the amateur weekend autocrosser".

Furthermore, I don't particularly like this sort of design because of issues to do with excessive wear due to flex of the races. Hence that "vertical stick" in the middle, which might have been added after the drivers felt a "cogging" vibration as individual balls rolled through the tight points. Also in the long run the seals will leak, the races rust, and more $$$!

If this design was forced on me, say by the marketing dept(!), then I would build my own. I'd buy the balls (cheap), but machine the races out of appropriate stainless steel (so can use no seals, just shields and grease) which are then welded to hollow section inner "upright" and outer "hub" sheet fabrications (so like two hollow doughnuts). Likewise homemade nylon/delrin ball "cages". The loads on individual balls should be low enough to allow this. See Chinese bicycles and their pressed metal races.
~~~o0o~~~

This would also allow for a quick release "centrelock" fastening.....

Originally posted by Big Bird:
I generally don't like centrelocks...
The closer to the axle centreline, the higher the contact forces...
I just think the larger the diameter where wheel loads are transferred into the hub, the lower those forces and therefore the less material required.

Geoff,

A "centrelock" nut is a toroidal, or ring-shaped, structure. It does NOT have to be "close to the axle centreline", having a small diameter, and hence high contact forces.

The HAW wheel could be attached to its rotating hub with a single, large diameter (~equal to the bearings) centrelock, err... ring-nut. This ring-nut would clamp the wheel against the hub at a large radius, and thus with low contact forces.

Furthermore, the nut would require only a single loop of thread to carry the loads. This is good because, as every small boy knows, the first thread of any bolted connection carries by far the majority of the load (because bolt stretches, and nut compresses), making most of the other threads hitchhikers (dead weight). This could be a beefy thread, say 3+mm high, yet still have high clamping force because the large diameter gives it a shallow wedging angle. I'd suggest an asymmetric "sawtooth" profile.

Z

[Moke]

You should see the custom bearing sitting on my boss's desk, $12 from China with a MOQ 100units and 2 week lead time, $120 from local suppler 2 month lead time.