Lightweight Front Hubs
 

Given I have access to a CNC machine I thought I might make some lightweight billet front spindles. Borrowing from the design of the Keisler drop spindles, I figured I'd design it around the MR2 Spyder hubs since they are easy to source and have an inbuilt speed sensor which might come in handy. I also need to credit the locost wiki site for providing 3D models of the stock hub which was a great starting point for the design as well as Leafy for providing me with some initial design feedback.

My goals are to make a nice light spindle and at the same time design it around the MR2 hubs, add in 2.5 degrees of camber and if possible hard mount my dynapro calipers. I don't particularly need a drop spindle however it would be easy modify the design to incorporate a 20mm drop... I was even wondering if I should add a second mount point for the hub so that I can try it. In terms of the steering arm connection, I'll make up some rod end connectors that screw into the threaded ends of the steering rack arms. This means there is less of a concern regarding angle of the ball joint and it will be lower profile too. I can also easily modify the pivot point with the use of spacers.

Design no1:
https://cimg3.ibsrv.net/gimg/www.mia...a6285655f6.jpg

This design was based on something I found online. The idea here is that everything could be CNC cut out of thick plates and bolted together. It would need the mating faces to be machined and maybe doweled as well as high grade bolts. I still like this design as it is so easy to make. If I continued down this path I'd thicken the base plate a lot to give it way more strength and it would also reduce the canter-levered suspension mounts. Maybe I could brace these with a similar idea to Concept no2.

Design no2:
https://cimg0.ibsrv.net/gimg/www.mia...898bacdd61.jpg

Just a quick sketch. The idea here was an evolution of design no1, with the a solid central billet 'T'ing onto the steering arm piece and tying the two suspension connections together.

Design no3:
https://cimg2.ibsrv.net/gimg/www.mia...ae2abf62ff.jpg

This was my starting point for what would be a solid billet design. This looked far too flimsy through the guts of the hub and would need a truss backing to reinforce the structure which moved me towards my last design conept no4.

Design No4:
https://cimg5.ibsrv.net/gimg/www.mia...42425bf817.jpg

This design looks good and is the lightest design so far weighing in at around 650 grams. It is a single billet at nearly 75mm thick that includes the Dynapro caliper mounts. Looking at the design, I think the central ring around the hub needs some additional material on the sides as well as some thicker ribs to the top ball joint. I'll also need to add some mounting points for some brake ducting. I need to get my hands on some FEA software so that I can see where the stress points are and I can evolve the design. This concept well require a whole stack of machining time cutting away the waste material, but the end result will be awesome.

Anyone have some comments / concerns before I start prototyping? Just a couple of notes is that my rapid 3D modelling doesn't include any radiused interior cuts. I will add these in as I further develop the model.

well i like the design with the removable steering arm, this allows for angle and ackerman change without a whole new knuckle.

yeah I liked the idea of bolting on the steeing arm as per designs no1 and 2. It makes it easy to machine and easy to change if required. The big concerns that the load on the top and bottom suspension flanges is massive and making that connection safe is my concern with this design. I designed No1 as a drop spindle, so bringing this back to stock ride height will give me a larger area to attach the top ball joint and maybe give it some additional bracing. Also adding in the additional 2.5 degrees of camber will reduce the cantilever forces on the top joint. I might re-visit this design tonight when I get a chance.

I would machine in the ball joint mounts, but have the steering arm bolted.

Can you make tabs for APracing calipers?

Sexy

.

I'm not actually sure. I've ordered a pair of OEM hubs from a wrecker so I can do my prototyping.

I've re-worked my first concept and made it a bit more solid. I think it still needs lots of work but it looks pretty damn epic! It's designed to fit into a 35mm thick billet so it's easy for me to machine. The top ball joint connector needs some work so that the bracing runs more to the actual mount point. The brake mount also needs some attention but overall I'm really liking the look of this design. All up around 850 grams.

https://cimg1.ibsrv.net/gimg/www.mia...a2b4b0dc1b.jpg

edit: with this design I can add some packing spacers to the top ball joint bracket to alter the camber a bit. I will also key the bottom steering arm where it connects to the main body of the spindle so that all the shear loads go through a a couple of locating rings rather than the bolts which are just applying a clamping load. This won't be packable due to the extremely high loads but replaceable steering arms could be handy for suspension tweaking.

Because they are custom I can do whatever brake mounting points are needed. I have dynapro calipers already though so I'll stick with those for now. I could make the caliper mounts bolt on so they can be interchanged but that reduces rigidity a lot.

Any idea on what kind of corner/bump forces we see in a Miata? How are these files saved? Can they be thrown into ABAQUS?

It's all modelled in Rhino3D so I can export as IGES, STEP and mesh formats. What do you need it in? PM me your email and I'll send you a few models.

If you take the weight of 250kg on the front tyre for the mass of the vehicle. We can acheive around 1.3G under braking plus dipping through a corner and hitting a kerb. So I'd guess at 10x - 20x the static load for dynamic loading which would be mostly vertical but with a bit of sideways loading also? There is also a secondary force generated axially by the caliper.

I can try with an IGES file, but you'd have to tell me what units you're working with so all forces can be scaled appropriately. ABAQUS doesn't use units, but will take measurements at face value. If I know everything is in reference to meters, that makes life easy. If mm or inch, then conversion has to be made for the numbers to reflect the units.
For instance, 1Pa becomes:
0.000001 N/mm^2
or 0.00014503773773 psi

It's all in mm. I can convert it if you need.

If it isn't too much to copy and scale everything in terms of meter, that make life a bit easier in terms of understanding results. Also, what material are you using exactly?

I'll have to see if my copy still runs at home, otherwise I might have to sneak something into work on off times.

Want. Subbed.

One complication with MR2 hubs is the flange face OD is slightly larger than the Miata's, so you may need to make custom rotor hats. Not the end of the world given your resources but good to note.

I would definitely make the caliper mounts modular. Don't design the whole thing around a basic budget caliper just because that's what you have right now. Future-proof your design.

Find a spherical bearing rod end available with the correct female metric threads to match the steering rack end (M10 I'm guessing off the top of my head?) and do the steering arm around that rod end.

Definitely want. Subbed

You really should confirm everything you got off that locust model. I used it when laying out the suspension arms and playing with them and They're not quite right. I had to tweak them to make the alignment match real life.

This is delightful. I am watching.

These are just concept models at this stage. They were modelled rapidly to test my ideas so a final 3D model will be built from scratch once I decide on the best design. Most of these concept were modelled within a few hours, so they are very rough around the edges but are fine for volume calcs and to get an idea of structure. The final model will be far more finessed, with nice fillets / chamfer etc all over it. I really like 3D organic forms so I might go all out and smooth oall surfaces to make it organic. Anyone who has done 3D prototyping / modelling will know how time consuming that is to do and a small design tweak can require a complete rework so you need to have a firm base design.

I'll need to confirm mounting dimensions, hub centerlines, hub bolt holes (once I have one), etc etc. I'll also make a prototype out of something easy to machine like chemiwood so I can test fit and double check all the geometry. The chemiwood is ideal as it's strength is just enough to test rigidity with your hands but still bolt things up to it. I'll be able to bend and break the prototype model I'll know where the weak point is on the design.

I need to decide if I'll go with Concept No1, or concept No4. I like the modular design of No1 as it's very easy for me to make. I can get the base structure waterjet cut first then load it up in my CNC to finish. I can also make different steering arms and brake mounts to fix alignment issues if I need to. No4 will require a massive billet and I need to get it right first go, with a single mistake requiring it to be remade from scratch. My CNC machine is a little hobby machine so it isn't suited to cutting from a single big billet, but I do have access to a workshop with a big machine.

Is this correct?

Forces on acting on the spindle:

Top ball joint - direction of force is inline with the arm, angling either forwards or backwards under brakes (x / y direction). There is no shear (z direction) other than MoI of the arm mass and suspension bushings and there is no rotational loading due to the ball joint.

Lower ball joint - Forces here act in all directions (x / y / z). The entire weight of the corner of the car goes through this point, plus the majority of the cornering load.

Steering arm - Forces are only towards and away the steering rack (y). There is no rotational loading other than the offset due to the pivot point of the rod end being offset from the arm.

Hub bolts - Forces here are in tension and compression only. All shear loading should be taken by the press fit of the hub into the spindle.

Caliper mounts - All forces here act axially around the centerline of the hub. The top caliper mount will be in compression under braking, whilst the bottom mount will be in tension. There needs to be enough sideways rigidity to stop any vibrations of the caliper.

not quite. For loading I like the standard 2g braking, 3g cornering, 4-8g bump. That way it should have enough margin of safety to even work on a crazy non-dot aero machine, at least for a while.

Upper ball joint will see some of the cornering load I'd give it 800 lb pull and 300lb push, and it will see the braking load but its very strong in that direction compared to what its going to experience so I'd probably not even FEA it.

lower ball joint will see a lot of the cornering load, give it 1.5k lb push and 800k lb pull, again for the braking. I'd give it 4k lb in the bump direction.

steering arm will see the steering thrust forces, which if you've ever driven on 275 hoosiers with a manual rack you'll know they can be enough to make your wrist sore for a week. I cant even come up with a number here, just its more than you're expect.

hub connection. Bolts should only ever be loaded in tension, just make sure your torque spec on them is high enough to ensure that. The mr2 hub itself shouldnt be press fit, it should be a slip fit into that hole for alignment only and should take the majority of its force in friction between the spindle and the hub, again make sure your torque spec produces a sufficient amount of clamping force to generate enough friction.

you forgot the hardest stress to account for, and thats the brake mounts. Those fuckers are the hardest thing to make happy in the whole damn spindle when playing with the parts in fea.

It's still only as much force as you can apply via a steering wheel though x leverage. Say I'm super buff and can hold 200 lbs through the wheel... whats the steering ratio? 4:1 plus a bit of extra leverage. so max of 1000lbs.

You can get a pack of ARP SS bolts that hold up to 170,000 PSI for around $40... something like that should be suitable I would guess. M10 x 1.25 is the thread. I'm not changing the hub mounting style so how they are mounted from factory should be ok... subject to big Hoosiers and aero of course.

The braking force isn't that massive given you have 4 of them working at once. I think the main thing here is to make the mounts rigid enough to stop the caliper moving around under load. I've found some references to high end spindles and the braking mounts aren't overly large, in fact some are on slim stalks. I figure as long as my mounts are larger and beefier than what is currently on the car I'm all good.

Power steering...

If you manage to put more force into the arm by moving the steering wheel than the ground imparts into the arm from cracks and bumps then you probably just had the tire stuck against something and reefed the fuck out of the steering wheel with PS.

Don't forget the force when bouncing against a barrier.
Think of the light ones that "just" bend a steering arm or a control arm.

Having the hub survive these incidents is a big plus. :o

I'm slowly working on my design. Not sure if it's still the right direction or not but I may as well try it and see how it goes. I figure I'm just going to start machining stuff and see how it turns out. I always find it easier to judge weak points in a design when I have it my hand. I'll probably start with the billet UCA given it's easy to make first and I can run it on my existing hub.

https://cimg6.ibsrv.net/gimg/www.mia...dee3c91d09.jpg

Here are a few key design features.
1. +3 degrees camber built into the spindle
2. MR2 hubs
3. Lightweight - approx 700 grams
4. Ducted 2.5" vent through hub to cool bearings and rotors (see below)
5. Spherical bearing for UCA connection
6. Billet UCA with 5/8 rod ends. Weight of the arm approx 550 grams
7. Rod end for steering arm
8. Solid mounts for radial caliper

https://cimg9.ibsrv.net/gimg/www.mia...c7b4f338c0.jpg

All up it looks a bit too light to me. I think I'm going to add a bit more meat by increasing the radius of the hub as well as the wall thickness of the inner and outer rings. I think the steering arm connection looks too large, so I might drop the height of that a bit.

Questions:
Should I include a small spindle drop or keep the factory hub location?
Running the vent through the hub I think is a good idea. It should route air right past the bearing which will be nice to keep the temps down in the bearing itself. I might need to add a small shroud out to the rotor surface. Any issues with this?

Keep up the good work!

Sub'd

Love the work, thanks for sharing the journey!

What kind of wheel/rotor offset are we looking at here for these? Or, how much farther does the hub face protrude compared to stock?

I run 6UL 10" rims with 245 Hoosiers with 11.75 Wilwood rotors and dyapro radial mounts.

Placement is exactly the same as a stock hub. The MR2 hub is shorter by a small distance. From memory I think the outer surface of the spindle is around 5 - 10mm further out to make up for the shorter hub.

I think if I wanted to run wider track I'd lengthen both the control arms rather than push the hub out further but that would mean making a billet LCA as well which is a whole stack harder.

Machining the spindle will be fun. It will need to be made from a 75mm thick billet. I'm going to need to machine up a tool holder for my CNC machine to mount up a 10mm bit. I'll need the length.

I'll get a new control arm scan done this weekend and send it over along with the upright scan I have. Sorry, that totally fell off my radar this past few weeks, been very busy with work.

.

I'll measure the bearing at some point. Now I have one I'll pull it apart and have a look. There is a whole range of timken hubs based on these bolt centers so there are more options for stud number and spacing. Ideally for a race spindle you'd use two offset bearings but thats a whole level of extra work.

The tie rod can be moved down by changing the spacer amount. If I make it so that the rod end center matches the stock position and then it can be packed downwards if required.

Two offset thin section angular contact ceramic hybrid ball bearings, hollow titanium spindle, etc?

Joking, I know what that would cost. Nice looking piece so far, wish I could run it for my class.

.

Abs mounting?

The MR2 hubs have a speed sensor. I don't run ABS anyway.

10 hours on the CNC and out comes this.
https://cimg2.ibsrv.net/gimg/www.mia...4da994e6a1.jpg

Machined from 19.05 mm 6061 ally.

Spherical goes in here with a bolted plate to stop it popping out. There is a small shoulder at the base of the hole.

https://cimg3.ibsrv.net/gimg/www.mia...19fb5683f2.jpg

It feels nice and light but still very rigid. I'm not sure if it needs more meat around the outside of the spherical. The holes might be a weak point.

https://cimg7.ibsrv.net/gimg/www.mia...a8b832c030.jpg

I still need to grab some 5/8" rod ends and a spherical before I can test fit it.

Beautiful!

I forgot if this got asked - is that your personal CNC equipment or are you borrowing time on someone else's? Just because I'd *like* to step up to having the ability to do stuff but relatively few people do.

I used to own this machine but sold it years ago. It's been sitting unused in a workshop so last year I asked if I could borrow it back for a while.

Was there any update to this? Now that I've got my AP Racing Pro5000Rs installed, the OEM spindle stiffness appears woeful at 1.4G longitudinal and 1200kg (2650lb).....

If you are still on OEM hubs, that is the source of most of the flex in an NA/NB. Just adding big billet hubs eliminates probably 90% of pad knockback and reduces flex enough to change pyrometer temps across the tread.

I’m on E30 hubs, so perhaps still something from that... changing pyrometer values in switching from OEM is crazy though! Would be scary seeing that on a high speed camera.

Theres no perceptible knockback with this setup, but it does have stiff knockback springs. From my pad taper I am suspecting the lugs are being deflected by pure braking torque. But I need to do the FEA to see what’s going on in detail.