Front knuckle/hub redesign

[matteo92065]

We are designing and building a new front knuckle/hub for the 818.

Your input and suggestions are welcomed, right up to the point of finished product. Then just keep your big ideas to yourself. :roll eyes:

This is in response to the thread, http://thefactoryfiveforum.com/showthread.php?15977-Front-Ball-Joint-Discussion-Thread-and-Potential-Safety-Concern, and the current weight of the knuckle.
The knuckle with FFR piece weighs 19.5lbs and lower ball joint is being constantly "pulled apart". I'm not making a judgement on weather or not the ball joint will fail. We are taking this opportunity to design it so if it does fail, it will be in compression and only get wobbly or loose feeling. With the redesign, it should never be able to fall or pull apart.

The final product should be a self jigging, box shape made of .065" CrMo sheet. The Sheet will be cut so that anyone can bend it using bare hands (not bear hands). Then weld it all together.

The first step was to get the hard points into CAD. These hard points are; hub face, steering arm hole, lower ball joint, upper ball joint, and brake bosses.
We hung the knuckle up and 3D scanned it. I've never done this before, so I'm sure it can be improved on.

Hanging knuckle:
35445

Scanned into iPad:
35427 35425
That is SO COOL! :cool:

Mesh it into an STL file:
35426

Import and overlay new design:
35423

its getting late, my battery is going to die, more later...

[Mechie3]

Awesome! I have all of the solidworks FEA packages as well as Ansys (not as practiced with this software) in case you don't have access to those and want help running some simulations. Any potential for a design that doesn't need to be welded up?

Did you scan it in house?

[Santiago]

My first concern is going to be safety. Ruling out the ball-joint concern is a cool, but not at the expense of additional risks. So I'll ask if there has been any effort or thought towards the material selection. I'm no expert in this area, but it seems that 0.065" steel is very thin for the expected stresses a knuckle assembly will see, even if it is CrMo. What safety factor are you aiming to build into the design? For that matter, do you have a clear sense of the loads this part would routinely experience? I guess I take a look at some of the aftermarket race-oriented knuckle assemblies and they're beastly stiff even for light-weight cars (billet aluminum is a common material).

I'm also not fond of a weld-up option. From what I've seen, CrMo needs to be professionally welded (or by someone who really knows the proper pre-heat and post-heat methods...much of which is routinely not done at so-called "pro shops"), so I would worry about coming to market with a product that seems to invite a high level of liability. One of the nice things about FFR's approach is that they're really only adding a bracket to an otherwise robust OEM knuckle assembly. They're not counting on welds to hold the whole thing together.

Just throwing out some concerns here early on.

Best,
-j

[matteo92065]

Thanks!, I only have the basic FEA for Solidworks.

We are still thinking about a billet aluminum version, but oh man, that is one BIG piece of aluminum! Also thinking about a bolt together version of aluminum and CrMo.

We've built some a-arms using this technique, and it was pretty straight forward including the welding. They are super strong even before welding.

Yes, scanned in house.

[Boog]

Maybe since you are doing a redesign it would be a good idea to incorporate the 05-07 STI unit bearing so you don't need to worry as much about the split bearing tolerances.

[Bob_n_Cincy]

I second the idea of starting with the 05-07 STI knuckle.
35433


Do you have a lower ball joint in mind?

[Boog]

Also if possible please don't forget about the ABS sensor mount for those of us who want to run ABS.

[Mechie3]

Lots of f1 and FSAE teams make welded together uprights from thin material. It's all about proper design.

[Frank818]

Lots of f1 and FSAE teams make welded together uprights from thin material. It's all about proper design.

And proper welding.

[Santiago]

Triple request on the 05-07 STI starting point.

I have no doubt one can design a proper welded upright, just that this is a concern. FSAE are just plain tiny vehicles, F1 teams replace parts with tremendous frequency, and neither ever sees a pot-hole infested surface (the occasional off-track excursions don't compare to the regular abuse a city street puts on parts). So in both of those contexts a designer is able to get away with plenty that we should view with strong suspicion for our application. Like does what counts as "thin" for their application make sense for ours? What caught my attention was the specific suggestion that these would be made of 0.065" sheet - if this were a go-kart, I wouldn't even blink. The ever-classic error in our hobby is using race parts for street use and guys wonder why stuff "doesn't last" or is hard to live with.

I'm also open to exposing my own irrationalities - I have no problem welding up control arms (they are simple and their stresses can be decently isolated even though they're large at times). However, when I think of steering knuckles on a full size street-driven car I get nervous. They're tasked with multiple duties, so proper design requirements really escalate. In my mind (clouded as it may be), you really want to be thorough in both how the part is designed as well as manufactured. But maybe I'm playing the part of the nervous old-nanny...

Best,
-j

[STiPWRD]

Let me know if you need any help looking at the stresses, I've been doing structural analysis with Ansys for several years. I pretty much agree with what Santiago said. This is a critical component and a complete redesign like this really needs to be looked a closely.

[Scargo]

I think it's great that someone is looking at this. I have to ask if you are going in this direction because you have to work within the envelope of the 818 design that adapts a Subaru AWD front upright?
I started a thread called "Lighter, race-oriented Suspension Parts, Primarily Uprights" (http://thefactoryfiveforum.com/showthread.php?12802-Lighter-race-oriented-Suspension-Parts-Primarily-Uprights&highlight=spindle). You may have missed it, but then again, I am race oriented. I don't care about painting within the lines of what FFR has given us as far as the front suspension goes.

I was talking to a racer at the track last Saturday and mention my STi's geometry. He asked if I had checked my Ackerman... Troy Ackerman, of the Dallas Cowboys, I asked? :o
Now, I know that the comments about, and performance of, the "R" has been outstanding, but does anyone know what Ackerman it has with an AWD upright and how to set up toe properly for it on a RWD car?
This, and weight, and a host of other things that compromise the front suspension, just to stick with a WRX donor... bothers me.

As John (Santiago) said, "Dropping the upper arm mount would be the way to go, especially if you "only" have to drop it - as opposed to draw it inboard of the frame's current tubing, build a standoff to mount it outboard of the frame, or just ditch the supplied A-arm in favor of one that will mount where you want it and still work with the geometry of the lower."
I guess this post may be a moot point if you are committed to using the upper mount. If you offered an engineered, weld-on bracket kit then there might be a plethora of spindles out there that would be candidates (as we were talking about in my thread (http://thefactoryfiveforum.com/showthread.php?12802-Lighter-race-oriented-Suspension-Parts-Primarily-Uprights&highlight=spindle)). I'm hoping that before long someone will want to offer a sophisticated front suspension setup for the guy who doesn't mind modifying the front end a little. I'm not interested in pursuing anything towards offering parts to the 818 builders, but I am trying to go in that direction on my own.

[Buzz Skyline]

This is a terrific start. If you weren't doing it, I'd be working on a custom made piece for my own car. I'd rather buy them from someone like you than try to build them on my own.

I'm all about the fail safe design. Have you looked into the model of ball joint you will be designing it to accept?

I've been trying to make my 818 as frugal a build as possible, but I'd consider this a necessary performance and safety upgrade.

[Frank818]

This is just for the FRONT knuckles, right?
Which is the one with the safety concern.

[Mitch Wright]

Back when we were racing Trans Am cars all the Riley and Pratt-Miller front uprights were made from chrome-molly sheet welded together.

[wleehendrick]

We hung the knuckle up and 3D scanned it. I've never done this before, so I'm sure it can be improved on.


Very cool... I've been using a RevWare MicroScribe digitizer (http://www.revware.net/products/microscribe/) recently at work. It's basically a mini-CMM that has a USB connection to output 6DOF data of the tip of a stylus (I have it set-up to export to Excel) It fits in a medium Pelican case and I'm not far from you and would be happy to give you a hand if you feel you need to improve the accuracy of any of the critical dimensions of your model of the OEM part.

[Santiago]

Back when we were racing Trans Am cars all the Riley and Pratt-Miller front uprights were made from chrome-molly sheet welded together.

That's good to know - those are big cars. Do you know what thickness they were using? Might be nice to get some insight into how they designed them.

Glyn, didn't you contact Palatov about going your approach? Did they ever get back to you? Just curious.

Best,
-j

[sponaugle]

Glyn, didn't you contact Palatov about going your approach? Did they ever get back to you? Just curious.


Do you mean Palatov Motorsports in Portland?

Jeff

[MrDude_1]

Do you mean Palatov Motorsports in Portland?

Jeff
I am pretty sure he means him.
If you haven't kept up on his blog.. its here: http://dpcars.net/

and heres a page where hes going through some of his upright design:
http://dpcars.net/dp2/da.htm

edit:

also look here:
http://www.palatov.com/products/suspension.html

[sponaugle]

I am pretty sure he means him.
If you haven't kept up on his blog.. its here: http://dpcars.net/
and heres a page where hes going through some of his upright design:
http://dpcars.net/dp2/da.htm


Cool.. Yea a good friend of mine Adam is the one who owns the lime green car. It is a pretty sweet car with some really cool engineering behind it.

Jeff

[Santiago]

Yep, Dennis is a stud. I was looking back at some of his old blogs, he's been at this for over a decade. Great stuff. I heard there's a D2 in my area, but I haven't seen it at the tracks yet. Damn shame...I soooo want a ride in one. =)

[Scargo]

Yes, it is that Dennis Palatov.
I was last in communication with him in early May. He recommended the GTS uprights.
35455
I filled out a worksheet that he provided for developing custom suspension geometry based on his parts. He said he was buried with work and then he never got back to me.
I will reconnect with him if others are interested. Or, perhaps one of you has more pull with him, especially if you know him personally?
I told him up front that I thought there would be interest in the 818 racing community if he were to tell us how to adapt it. I got all the geometry specs from Jim Schenck. I could share that, too.

As a side note: When reviewing my emails with Dennis I said, "Seems the one you recommend has reverse Ackerman geometry which, I take it, is more desirable in high-speed cornering on pavement. The stock Subaru front has about 0 Ackerman."

Glyn

[Santiago]

Sorry, I don't have much pull with him as I don't know him personally. I'm just a fan/follower of his blogs. Somebody else here might know him though.

I'd love to see the full specs Jim sent you though. My current plans involve some mild modifications to the front suspension geometry, so that background info would help regardless of what comes of new spindles. Might also be nice for Mateo and company to verify their measurements as well.

Best,
-j

[matteo92065]

Worked on scanning and getting parts into the CAD model today. Yes, I would like to use a bolt on hub like the 05-07 STI. But that is a 114 B.C., so we are also looking at 05-10 Scion TC rear hub. It is small and has 5x100 B.C. By the way, the bolt on STI hub weighs 7.5lbs. Any-who... The two hubs are almost identical in the way they mount and function.

We will recalculate the Ackerman angle to fit the shorter, narrower 818.

Ball joint selection... I'd really like to just flip the aluminum arms with the existing Subaru ball joint. That way the joint would be in compression. I'm working on trying to make that work right now. I don't want to get into making a new a-arm to match the knuckle.

The thickness of the material... I mentioned .06 sheet. I'm not set on that, just a starting point. The a-arms we've made in the past were actually .08". By starting with a slightly thinner material, I can double and maybe triple up layers in high stress areas.

FEA and Safety Factors... Modeling heavy braking is straight forward enough, but to simulate loading while turning?!
When modeling you have to choose constraints and forces/pressures. What forces do you want to use? Maybe we start with cornering at 2g's and the car weighs 2200lbs, and all the weight is on the outside front tire, and then double that force? Do you design for a heavy hit into a curb? Do you plan on 8" deep pothole at 90mph while drifting? I assume engineers in this field do have starting points, but I don't know them.
At the least I am going to model a similar production car knuckle (370Z has similar design) and see where it fails. Then match it. Again, I'm open to suggestions on this.

Too much thinking for tonight.

[Bob_n_Cincy]

Hey Matt
The 08+ impreza uses the bolt on bearing and is 5 x 100
https://www.rockauto.com/catalog/moreinfo.php?pk=1572284&cc=3307193

For modeling you need to use a 4" deep pot hole at 80mph. If you want to sell them in Ohio or Michigan.
Keep up the good work
Bob

[Santiago]

For modeling you need to us a 4" deep pot hole at 80mph. If you want to sell them in Ohio or Michigan.


Nice. But for Chicago you have to go deeper. =)

I think some (many?) folks would prefer the 114 bolt circle because it opens up a wider range of wheels. Something to consider.

Also, the link that Mr Dude sent has some great info on what D.Palatov had in mind when designing his uprights. A good read for sure. Back then he was looking at a 4G load on a 2,000 lb vehicle. He also indicated how he set up his FEA work. I seem to recall he put stresses on it in three directions (one for braking, one for steering, and one for the hub load during cornering). I thought the idea was that he was trying to "twist" the model to see how it performed. Take a look.

Best,
-j

[Scargo]

Here is another nod for 5x114. Seems like that opens the door to play with the big boys.

Here are the front suspension numbers from Jim Schenck, in reference to the worksheet below, and a comment he made:

"My thoughts on Ackerman are somewhat downforce dependent, if you are running the full aero package then I would stick with some ackerman in the steering. My experience is that mid engined cars in general get more and more tail happy as the speed goes up so in order to get them to be balanced in the higher speed corners you often have to make them push at lower speeds. Having the ackerman helps get rid of some of that low speed push without much penalty at higher speed. If you add downforce then you also help this out because the car can be low speed balanced mechanically and high speed balanced with aero, but having downforce on the front helps the inside tire to do more of the work, making the steering angle more important and hence the ackerman helps over a wider speed range."

Numbers for a 4.5 inch ride height (in inches):
Lower arm pivot to ground: 9.25
Bellcrank pivot from ground: 21.19 (upper shock mount)
Pushrod pivot from ground: 9.375 (lower shock mount)

Lower arm pivot from centerline: 14.0625
Wheel center from centerline: 30.125
Bellcrank pivot from centerline: 18.23 (upper shock mount)
Pushrod pivot from centerline: 24.9375 (lower shock mount)

Shock compressed length: 11.10
Shock extended length: 15.15

35472
Note that this is a worksheet from Dennis Palatov for a push-rod style of front suspension. Dennis said to just substitute the top and bottom shock mounting points for the pushrod reference points. That's what Jim is making reference to in parenthesis as he refers to the worksheet.

[Mechie3]

Unfortunately modeling "4" deep pothole at 80mph" isn't as easy as checking off boxes in a simulation.

[JAubin]

The spindles from Palatov look really nice and well designed. They're intended for a 2k lb. car @ 600 hp, so they seem like they'd be a decent fit from a spec. standpoint, and if the target from his blog was met, they're around $1500 a pair which seems like a really good deal for a race level component. I wonder if it would be a possibility to design some parts to adapt these to the current suspension in the 818. They're already 5X114,.

[Buzz Skyline]

Ball joint selection... I'd really like to just flip the aluminum arms with the existing Subaru ball joint. That way the joint would be in compression. I'm working on trying to make that work right now. I don't want to get into making a new a-arm to match the knuckle.

FFR sent me a cutaway of the ball joint currently in use on the Subaru LCA's. Unfortunately, flipping them to compression, while leading to a more fail-safe design, might not be a good idea.

35473

As you can see, it's assembled by dropping the ball in rod-first, then capped with a sheet metal piece. In compression it would press on the fairly fragile cap, which could be even worse than using them in tension. There are lots of cars that use load-bearing ball joints in compression. You could probably pick one from a light car that wouldn't be much bigger than the Impreza unit.

Howe ball joints let you mix and match components, so you could potentially find a compression socket and pair it with a ball that has teh right taper to fit the flipped Subaru LCA.

https://www.howeracing.com/c-560-howe-precision-ball-joints.aspx

[RM1SepEx]

or you could rebore the Subaru LCA bushing for another stronger joint's taper

[Buzz Skyline]

or you could rebore the Subaru LCA bushing for another stronger joint's taper

Yeah, I was considering that too. The taper drills are pricy, but it might make sense for a business to pick them up. Summit has some that would probably work

http://www.summitracing.com/search?SortBy=BestKeywordMatch&SortOrder=Ascending&keyword=tapered%20reamer

[matteo92065]

Thanks for that information on the ball joint!
How about the lower ball joint from the Nissan G35? It looks like it uses the same tapered shim as I've seen on Subaru LCA ball joints. Nissan uses that ball joint in compression, so that's good.
If the tapper does not match Subaru's, then we will make our own tapper shims, or "ball joint seats".
Or I need someone to suggest a suitable replacement.

For the hub, I'm now leaning towards the STI bolt on. And just drilling out for 5X100 studs when needed.
I'm still working out the ABS needs.

[sponaugle]

FFR sent me a cutaway of the ball joint currently in use on the Subaru LCA's. Unfortunately, flipping them to compression, while leading to a more fail-safe design, might not be a good idea.
35473
As you can see, it's assembled by dropping the ball in rod-first, then capped with a sheet metal piece. In compression it would press on the fairly fragile cap, which could be even worse than using them in tension. There are lots of cars that use load-bearing ball joints in compression. You could probably pick one from a light car that wouldn't be much bigger than the Impreza unit.


Interesting. Since the joint is dropped in from the top, the failure mode we are talking about is the joint pulling thru the bottom of the holder... that is physically blocked by holder, so the failure occurs by way of that holder deforming and letting the ball come all the way thru?

This picture below is upside down from my description, but the image on the right is what we are talking about, correct?

https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcTRgT62dSR38nc-TfuzH5CLoZ2lXs5S6xlPx7-kaldn87auulbD

Jeff

[Jaime]

FFR sent me a cutaway of the ball joint currently in use on the Subaru LCA's.
35473

It looks to me like this cutaway makes the whole ball joint safety question go away. In a WRX, the ball joint is essentially unloaded. In the 818, it's loaded in tension. We were assuming that Subaru didn't use a tension ball joint, but that looks like a tension unit to me.

[sponaugle]

It looks to me like this cutaway makes the whole ball joint safety question go away. In a WRX, the ball joint is essentially unloaded. In the 818, it's loaded in tension. We were assuming that Subaru didn't use a tension ball joint, but that looks like a tension unit to me.

Yes, that is what I was asking as well. Since the joint unit goes into the housing from the top (as used in the car), it would seem difficult to pull thru. I can't tell the exact dimensions but the ball will not pull thru the lower part of the housing, correct?

[Frank818]

For modeling you need to use a 4" deep pot hole at 80mph. If you want to sell them in Ohio or Michigan.

And around here in QC for modeling it would be a series of 10 consecutive 4" deep potholes at 70mph, filled with 2" high fast-compression bumps. :)

[Buzz Skyline]

It looks to me like this cutaway makes the whole ball joint safety question go away. In a WRX, the ball joint is essentially unloaded. In the 818, it's loaded in tension. We were assuming that Subaru didn't use a tension ball joint, but that looks like a tension unit to me.

That's true, and it's clear that the ball joint won't immediately fail. But joints that are designed to carry loads have bearing surfaces made of teflon or other low friction material, as well as being greased and often having lub nipples attached. They also typically have wear indicators to let you know when the joint has to be replaced, before catastrophic failure. They are also much beefier, even for light vehicles like classic Beetles and Miatas.

This one has metal to metal surfaces that are designed for precision, not load, so the bearing surfaces are likely to wear much faster under load than they do in their design application. And because it's in tension, I imagine the joint won't even feel loose until the surfaces wear out and it pulls through. You could (and probably should) unload it frequently to check for looseness, but otherwise there would be no sign of an impending failure until it popped.

So, yes, it could have been a worse design than it is. But that's not a terribly comforting way to look at it to me.

Here's a video showing what the inside of a load-bearing ball joint looks like.
https://www.youtube.com/watch?v=04DJ1R5cSy4

[Santiago]

That's true, and it's clear that the ball joint won't immediately fail... You could (and probably should) unload it frequently to check for looseness, but otherwise there would be no sign of an impending failure until it popped.


I had the same thoughts once seeing the cutaway. Impending doom it is not; something of a frequent & regular maintenance item it is.

So, even if this does not make the whole ball-joint issue go away, perhaps this means focus should be on sourcing a ball-joint replacement. Howe is pretty good to work with - I use their joints in the SLA conversion on my Mustang. They may even be open to specific new design if (a) there's adequate market interest and (b) what they have now won't do.

I still think a redesigned spindle is a worthy project, but at least for me the motivation would shift away from the safety concerns of the ball joint towards performance concerns (principally weight savings and possibly ease of brake upgrades). I can live with frequent ball joint maintenance on my R since I'm already used to seasonal inspections, tear-downs, and upgrades.

Best,
-j

Best,
-j