MR2 Owners Club Forum banner

1 - 20 of 46 Posts

·
Premium Member
Joined
·
13,212 Posts
Discussion Starter #1 (Edited)
In parallel with, and as an add-on to, my rear knuckle project I have been working on something a little extra special for the rear of my MR2. I have been hesitant to say anything about it until I felt it was ready, but I think it's time to throw it out there to get some feedback from the community.

Here's the rear view in SusProg 3D:
73591


And here's the basic layout in Solidworks:
73593

Overall layout is quite similar to the Porsche 996/997 rear suspension. Obviously it requires some substantial cutting and welding on the chassis to install those two boxed mounting points for the upper links. It also requires a custom lower control arm (quite a bit shorter than stock), a custom cross member with the control arm and tie rod points moved outward, AND super high offset wheels. And even then it will likely require a widebody or substantial fender flares to get any decent amount of wheel width. (I'm thinking 17x10 +66 on my car, and that would fill out my flared fenders).
See later posts for details, but I think I can avoid needing some of these items. At least I think I can re-use the stock control arm mounting points on the cross member. For my car I will likely use somewhat longer control arms than stock just to maximize link length and fill out my fenders with my current wheels.

So, not without it's drawbacks and additional expenses. But look at these camber curves. Blue is the multi-link, gray is my current (optimized) strut.
73594


73595


I think I would be able to run significantly less static camber and still have better camber on the outside wheel in a turn, not to mention having much less excessive camber on the inside wheel. Roll center of course is higher than the strut, and much better behaved as well (doesn't move around much due to body roll). Toe curves can be basically whatever I want, although I think I will shoot for something similar to the stock 93+ curve, but with less total toe change.

In addition to the improved kinematic camber curves, there is also substantial flex in the strut that doesn't get talked about. My analysis from designing the knuckle predicts about .06-.08 inch flex outward at the strut to knuckle interface (at 1.4g), mostly based on the strut itself flexing. That's about 0.4° camber loss due to flex, so the actual strut camber curve is even worse than what is shown above.
 

·
Premium Member
Joined
·
13,212 Posts
Discussion Starter #3
Alex are you just aiming to bolt that on to the stock shock?
Yeah, the upper adapter for the multi-link (which replaces the strut tab on my knuckle) has a single spherical bearing instead of two bolt holes for the strut bracket. So it should work to use the existing coilover. About the ONLY part that won't have to change that way!
 

·
Registered
Joined
·
776 Posts
Cool stuff!

What’s the reason for the new rear crossmember? The stock pivot point just not in a good spot?

What’s the reason for moving the hub out so far? Just getting much better kinematics in the available space?

You planning to make this a sort of kit? I imagine very small market, but would be cool to see it more widespread.

Front in future?
 

·
Registered
Joined
·
212 Posts
Cool stuff!

What’s the reason for the new rear crossmember? The stock pivot point just not in a good spot?

What’s the reason for moving the hub out so far? Just getting much better kinematics in the available space?

You planning to make this a sort of kit? I imagine very small market, but would be cool to see it more widespread.

Front in future?
lower arm would be WAY too long if you used the stock pivot. It may be possible to just modify the existing crossmember, but knowing Alex he already has a tubular subframe in the works. Because of the location that the top arm has to be, the lower needs to be close to that same spot. That's also why you would have to run super high offset wheels to get the wheel back inset in the car.

In the past Alex told me that the front isn't as important, note that many Porsches have MacPherson strut on the front end and their rear ends are fancy German multilink suspension
 

·
Premium Member
Joined
·
13,212 Posts
Discussion Starter #6
Cool stuff!

What’s the reason for the new rear crossmember? The stock pivot point just not in a good spot?

What’s the reason for moving the hub out so far? Just getting much better kinematics in the available space?

You planning to make this a sort of kit? I imagine very small market, but would be cool to see it more widespread.

Front in future?
My original goal was to be able to use at least the stock control arm position on the cross member, but that ended up with too large of a length difference between the upper and lower arms. The result is a less linear camber curve, primarily effecting the rebound side of the curve. So you get camber curves like this: (yellow is the shorter arm version for reference).

73596


73597


It doesn't effect the really important outside wheel much (and actually, makes the camber curve there ever so slightly better), but it increases the unwanted camber gain on the inside wheel. And it just seems like a bit too much compromise. Granted the shorter lower arm does increase the wheel scrub (sideways movement due to suspension movement) some. I'm not sure how important that is, but there is only so much you can do about it given the space available.

Moving the hub out is just a play to get the longest possible arms in there. Even longer would be better, but again, only so much you can do with the space available. Those boxed attachment points set into the frame rail are for the same purpose. Doing that vs just welding tabs onto the side of the frame gains me about 2 inches of length on those upper links.

I won't rule out the possibility of a kit, but like you said, probably a small market. Not that many people are willing to go to these lengths. Front is also a maybe someday project, but is of lower importance in my opinion.
 
  • Like
Reactions: SixAppeal

·
Premium Member
Joined
·
13,212 Posts
Discussion Starter #7
lower arm would be WAY too long if you used the stock pivot. It may be possible to just modify the existing crossmember, but knowing Alex he already has a tubular subframe in the works. Because of the location that the top arm has to be, the lower needs to be close to that same spot. That's also why you would have to run super high offset wheels to get the wheel back inset in the car.

In the past Alex told me that the front isn't as important, note that many Porsches have MacPherson strut on the front end and their rear ends are fancy German multilink suspension
What he said!

Initially I plan to modify the stock cross member. Tubular, maybe some day.

And, I'm not saying I wouldn't rather have a double a-arm front to go with it! But it's certainly secondary. For one thing, the camber curves on the front are already quite a bit better than the rear. For another, the Porsche example.
 

·
Moderator
Joined
·
5,343 Posts
Nice work, I've always wanted a rear multi-link on my deuce, similar to an S2000 (which also has crazy high offsets, 8.5" +55 rears if I recall).

While I'll leave the technical feedback to others, I can comment the obvious that a beautiful design like this would only appeal to a handful of owners ($$$) should you start marketing it.
 

·
Registered
Joined
·
141 Posts
In addition to the improved kinematic camber curves, there is also substantial flex in the strut that doesn't get talked about. My analysis from designing the knuckle predicts about .06-.08 inch flex outward at the strut to knuckle interface (at 1.4g), mostly based on the strut itself flexing. That's about 0.4° camber loss due to flex, so the actual strut camber curve is even worse than what is shown above.
This is something I had not considered before, cool insight!
 

·
Registered
Joined
·
168 Posts
Nice to know that the ruler and eyeball measurements I did months ago weren't far off. I think I found that you would need such as high offset too. I am in to see the development of this.
 

·
Registered
Joined
·
83 Posts
Man, that's an awesome concept. I'm already 100% set on getting the geometry correction kit when I lower my car a bit, but this is just nuts. I'm not really at the level to contribute to the discussion much technically, but this would be amazing to see fully realized on an SW20. A kit would certainly interest me, but installation difficulty would certainly be a significant thing to consider.
 

·
Registered
Joined
·
776 Posts
Question - how bad would the geometry get if you were to make it a more bolt-on solution to the SW20 with maybe some weld in captive nuts to bolt on brackets?

I know it's difficult to design around one set input being a bit non-optimal, but it might open it up to being a kit that's one step farther forward than the extreme top mounts.
 

·
Premium Member
Joined
·
13,212 Posts
Discussion Starter #14
Question - how bad would the geometry get if you were to make it a more bolt-on solution to the SW20 with maybe some weld in captive nuts to bolt on brackets?
Depends on how far you take that. As it is now my upper links are about 8.25" long (in the rear view, 9.4" actual length considering the diagonal). Lower control arm is about 10".

If you skip cutting into the frame and just surface mount the inner end, you lose about 2", so you are down to 6.25" length on the upper arms. If you keep the lower arms the same, and keep the high offset wheels, your camber curves look like this:
73705


Actually looks quite good in roll, lots of camber gain on the important side of things.

73706


But the camber gain in bump is pretty excessive. HOWEVER, maybe you can now get away with running only -1° static camber? Honestly it seems to work surprisingly well, maybe this is worth considering.

It's interesting how this mostly effected the curve in bump, and as I showed before lengthening the lower arms mostly effects it in rebound. I think this is mostly down to how the relative angles of the arms change as you change the length, but it can be a little tricky to keep track of what effects what.

I think personally I would still prefer to cut and box the frame and instead skip the high offset wheels. Considering I have 3 sets of wheels at the moment, it would be nice not to have to start over there!
 

·
Premium Member
Joined
·
13,212 Posts
Discussion Starter #15
If you just skip the high offset wheels and keep everything else the same, you lose about .75" of length on the upper links.

Camber vs roll is slightly better:
73707


Camber vs bump is slightly worse:
73708


BUT, keep in mind I can also achieve the roughly same camber curve changes by lowering the lower ball joint (ie, more roll center adjuster).
 

·
Premium Member
Joined
·
13,212 Posts
Discussion Starter #16
Basically, longer links improve the relationship between the bump camber curve and the roll camber curve. That's why formula cars tend to have super long a-arms:
73710


For example, if I keep the chassis side the same and lengthen the lower arms by 4", and then drop the lower ball joint until the roll camber curve is roughly the same as before, you get camber curves like this:
73711


73712


Roll camber is roughly similar, but we have less excessive camber gain in bump. But of course this widens the track width by 8", and at this point the suspension isn't really fitting in the car any more. Have to work at least somewhat within the constraints of the chassis.

All that said, I think it's worth considering that the high offset wheels aren't gaining me all that much, but I think the frame boxes are worth it.

On the subject of making it more "bolt on" one thing I have considered is adding holes to the flanges on those upper link boxes and pop riveting them on. A 1/4" diameter steel rivet has a tensile strength rating of 1500lb, and you could easily use 10 of them per box. Even 3/16" rivets will get you 800lb each. Upside to welding is it will tend to more uniformly distribute the load into the side of the frame, and the box will do a better job of reinforcing the frame where you cut a chunk out. The downside to welding is it's hard if not impossible to rust proof the back side of those welds. And you have to trust your welds.
 

·
Premium Member
Joined
·
13,212 Posts
Discussion Starter #17
Last consideration, if we go back to the stock lower control arm length and space the lower ball joint down we can lessen the impact on the camber curve. It depends of course on how much you space the ball joint down, but with an extra 1/2" of RCA you can end up with this.

73713


73714


Not too bad I guess... If I don't touch the cross member at all, I end up with a toe curve roughly similar to the 93+ curve, using the 91 cross member, so that also works out relatively OK. And it wouldn't be too hard to modify the cross member to shorten the toe arm a couple of inches, just add a weld nut and extend the flange on the outside. Considering the reduction in complexity, maybe this is worth it.
 

·
Registered
Joined
·
776 Posts
I would caution using rivets in tension. They have some ultimate capability there, but small prying effects across the rivet head make it very fatigue limited as a joint. Also, a steel 3/16” rivet is an absolute bear to pull with a hand rivet gun (and would need a very big rivet gun at that). 1/4” steel rivets would be extremely hard to install. Keep in mind rivets are almost “no preload”, which makes them very susceptible to fatigue in tension unlike a preloaded fastener.

While welding does need to a certain level of skill, weld metal is 70 ksi ultimate tensile strength min, so you don’t need much weld metal to get a huge amount of strength.

There’s also the problem of a rivet joint is still going to dump the load into your thin stock frame rail. Without doing calcs, I would be nervous of that side everything taking the load reliably.

I think a weld on bracket as 1 piece is likely the best thing to help dump both the tensile punch load, but also deal with the differential punch load as you accelerate or brake (twisting frame rail with one being tension, one being in compression). It also makes positioning easier. I could envision it being a simple .080-.100” thick piece with a bent flange to give it more bending stiffness, then weld on the Clevis ears to mount the rod ends.

If someone did want to go back to struts for some reason, you can cut off the clevis ears and just have a fancy welded on frame rail reinforcement.


I personally like the idea of trying to keep as many things consistent between a “good strut setup” and “good a arm setup.”
 

·
Premium Member
Joined
·
13,212 Posts
Discussion Starter #19
I would caution using rivets in tension. They have some ultimate capability there, but small prying effects across the rivet head make it very fatigue limited as a joint. Also, a steel 3/16” rivet is an absolute bear to pull with a hand rivet gun (and would need a very big rivet gun at that). 1/4” steel rivets would be extremely hard to install. Keep in mind rivets are almost “no preload”, which makes them very susceptible to fatigue in tension unlike a preloaded fastener.

While welding does need to a certain level of skill, weld metal is 70 ksi ultimate tensile strength min, so you don’t need much weld metal to get a huge amount of strength.

There’s also the problem of a rivet joint is still going to dump the load into your thin stock frame rail. Without doing calcs, I would be nervous of that side everything taking the load reliably.

I think a weld on bracket as 1 piece is likely the best thing to help dump both the tensile punch load, but also deal with the differential punch load as you accelerate or brake (twisting frame rail with one being tension, one being in compression). It also makes positioning easier. I could envision it being a simple .080-.100” thick piece with a bent flange to give it more bending stiffness, then weld on the Clevis ears to mount the rod ends.

If someone did want to go back to struts for some reason, you can cut off the clevis ears and just have a fancy welded on frame rail reinforcement.


I personally like the idea of trying to keep as many things consistent between a “good strut setup” and “good a arm setup.”
Yeah, I agree on the rivets. Maybe technically OK per the strength spec, but I just don't like it. Just something I considered, but had more or less discounted.

I do feel like keeping the stock cross member, stock lower arm lengths, and stock wheel offset might be more feasible than I had originally thought. Since the camber curve in roll is actually better you can run less static camber to offset the excessive camber curve in bump, and in the end I think it works out OK.

You cannot imagine the circles I have been in on this design trying to figure out a way to make it all work in the space available. Any suggestion that causes me to circle back to a simpler solution is helpful!
 

·
Registered
Joined
·
776 Posts
The ultimate solution in my mind is at most very slight notching of stock frame rail At most a little clearance for rod end protrusion into frame rail at most with say a 3/4” hole. This would also allow you to fog the frame rail with primer + paint after welding. Maybe not perfect corrosion prevention, but probably fairly close.

So that’s a flat plane with the frame rail.

Stock cross member with only maybe a bolt on bracket for toe arm at most.

Stock offset wheel. I’d probably baseline a 9” +35/9.5” +38 as being the most typical size.
 
1 - 20 of 46 Posts
Top