1967 MGB - Father/Son build

[boodlefoof]

Hi all! Long time no see. It has been a few years since I've been on PT.com. Projects have started and stalled and remain yet to be completed. I had a bunch of kids. Things have been crazy. Now, my oldest is about to start driving.

And since he's about to start driving... little boodlefoof and I are doing a father and son build. We're going to hold one another accountable to get the job done! We're also going to document the project here. That project will be a very upgraded 1967 MGB.

In 2016 I found and purchased a rolling 1967 MGB body (I had no place to put it, but knew I'd get to it someday). The car had no parts other than enough suspension to allow it to be rolled around. That was OK. Knowing myself, I would want to upgrade or replace everything anyway! Fortunately, the prior owner had replaced the floor pans, rockers, and inner rear fenderwells. It appears he did a good job too. So, what we have is a solid blank canvas to work with.

198769
198770

To begin the project, we began by taking inventory of what parts we had, what we did not have, and what we would need to fabricate or buy, while keeping in mind that we wanted to keep costs to a minimum. We discussed the project and our goals: to make a fun, reliable, inexpensive, and competent fair-weather sportscar that we could have fun with at track day events. Oh yes, and he is going to learn to drive a stick-shift with this car. The general concept came together as follows:

Transplant in a modern, more powerful, fuel-injected four-cylinder engine and a better manual transmission. Map the existing suspension and reconfigure to optimize handling; in particular replacing the rear leaf spring suspension with a link-type suspension, and minimizing front scrub radius and bump steer. Looks-wise, we plan to simply do without a convertible top, instead opting for a speedster look with a shortened and raked windshield. Some lightweight aluminum wheels would look nice, potentially reduce weight, and help reduce tire scrub with additional offset. With the extra get-up-and-go of the aftermarket engine, we contemplate brake improvements. The interior will be minimalist, but functional. We will likely also make some minor body modifications.

This is also going to be a learning experience and we plan to present the project write-up accordingly by discussing our modifications and improvements in as much detail as we can.

Brief preview: We've already begun planning and working on the project. Out back we're doing a 3-link suspension with a Watt's-link, and using the stock 3.91-geared rear end. Up front, we're fabricating a new crossmember and tubular control arms with coil overs. The engine will be a Toyota 3SGE Beams Blacktop that was only available in Japan (2 liter, inline 4-cylinder with 11.5:1 compression that makes 210hp, 163 ft. lbs. of torque, and redlines at about 8,000rpm). Transmission is a Toyota J160 (6-speed). When done, the car will weigh under 2,000 pounds.

Looking forward to being back on here.

[Rick Dorion]

Great to see you back. Following this. I have a soft spot for the MGB and MGC.

[andrewb70]

This should be fun!

Andrew

[fordtc87]

Oh, I love something different... Best of luck, this will be a build worth following.

[boodlefoof]

After setting our goals for the project, we decided to start with the design and build of the rear suspension.

As I mentioned, this is a learning experience for the both of us, and we hope to share our experience with you. So, we will try to get technical in our discussion of what we're doing. That does mean lots of text... Also, we don't want to propagate bad information and I'm a bit rusty on the knowledge I once had on the subject of suspension design (it has been a few years). If I've gotten off-track on any of the theory below, please chime in! If you want to skip the theory, you can go straight to the pictures.

Factory Suspension:

We began with a review of the baseline provided by the factory front and rear suspension. Removing the wheels and blocking the body to our desired ride height, we used a plumb-bob to mark the wheel centerline, front and rear, and draw lines on the garage floor representing those centerlines. From each line, we marked a grid on the floor that would be used for our measurements. Going back to the plumb-bob, we marked on the floor and measured with the grid (as best we could - probably within 1/8'' on all dimensions) the side-to-side and fore-and-aft contact points of each suspension pickup point: each control arm pivot, spindle attachment, tie rod mount point, leaf spring eye, etc. We also measured the height of each point from the floor. We then plugged these X, Y and Z coordinates into Suspension Analyzer (made by Performance Trends) to model the suspension. With these inputs, the program can measure camber and caster gain, bump steer, anti-squat, swing arm lengths and the like.

The rear leaf spring setup shows a roll center height ("RCH") of nearly 13’’. Most here are probably already familiar with RCH. For those who aren't, RCH can be thought of as the imaginary point about which the car’s body wants to lean (“roll”) in a turn. The higher this point (the closer to the car’s center of gravity height), the less prone to roll the body is - that is, the suspension's design resists body roll. While this might seem like a good thing (as we don’t want the body to roll too much in a turn), a high RCH results in more lateral force of the turn being transferred into the outside tire (potentially causing loss of traction) instead of that force being transferred into body roll. A lower roll center height, paired with other roll-control devices (such as a sway bar) can reduce body roll and avoid unpredictable loss of traction. People wiser than me suggest that for a solid axle rear (non-IRS), an RCH of between 5'' and 10'' is a good place to be.

In addition to the high rear roll center, this MGB was also fitted with a 5/8’’ rear sway bar. The roll stiffness provided by the high roll center height, paired with the sway bar results in a rear suspension very resistant to roll. Having mapped both the front and rear suspension (we will discuss front suspension observations when we get to that part of the build), the front suspension of this car, even with a front sway bar, is not nearly as resistant to roll. This results in a tendency for the car to oversteer: that is, for the back end to slide out in a turn. The leaf spring suspension also features a lot of anti-squat (AS), which is the suspension's resistance to squat upon acceleration. Very high AS can even cause the rear of the car to rise upon acceleration (even though weight is transferring to the rear).

Theory of the 3-link alternative:

To better pair with an improved front suspension, we decided to eliminate the leaf springs and replace them with a stable and predictable 3-link-type suspension. The 3-link uses (as the name implies) three trailing arms that run fore and aft (generally parallel to the car) to locate the axle. Two links are typically located outboard (near the wheels) and below the axle housing and the third link is attached on top of the center of the axle housing, sometimes offset to the right (passenger's) side to counter-act rotational forces in the driveline that can cause unequal loading on the rear wheels. The frame mounts for all of the links are forward of the rear axle.

To provide neutral steer (no under- or over-steer), the lower links are situated to be parallel to the ground. The vertical distance from the axle centerline of these mounts (and the upper axle mount) is important - less distance puts more stress on the links. Some recommend mounting the lower links as low as practicable while maintaining ground clearance and mounting the axle link as high as possible as well (without causing too much upper-link angle). To provide AS, the upper link angles downward towards the front of the car (higher at the axle, lower at the frame). The higher the angle, the more AS. However, the higher this angle, the shorter side view swing arm length (SVSAL). SVSAL is the distance at which the upper and lower links would intersect in side-view if they continued forward to the point of intersecting. Too short a SVSAL can result in brake hop. Herb Adams recommends no less than 42'' of SVSAL. For both the upper and lower links, longer is generally better, as longer links experience less change in angularity as the suspension moves. This means SVSAL, AS, and the pinion angle, do not change as much as the suspension moves.

Our 3-link design:

Given the available space under the car, we determined our upper link could be about 17'' long at most. Towards the front, we run into driveshaft/trans tunnel clearance issues. At the axle housing, we gain a little length by offsetting the upper mount behind the axle centerline. We set the length of the lower links at 27''. Our available space woudl allos an upper link adjustable from 17.5 - 19'' in height at the axle side, so we will be able to tune anti-squat a bit. At maximum third-link height, SVSAL is 61'' and anti-squat is 47%. At 3'' of bump, anti-squat is still positive (18%) and SVSAL is 44''. On the lower third-link height setting, AS comes down to 30% at ride height and SVSAL is longer. We also will make the upper link mount (on both the axle and frame) adjustable from side to side, so that the mount point can be between 1.5'' and 3.5'' to the right of center. This should help us tune and balance tire loading.

Lateral axle location:

Now, the 3-link only locates the axle front to rear. A 3-link requires a separate lateral locating device, and this device sets the RCH of the rear suspension. The most common lateral locating device is the panhard bar. A panhard bar is a long single link running from side to side, attaching to the car’s frame on one side and to the axle on the other side to keep the rear axle from moving from side-to-side. However, as the suspension moves up and down, the arc of this link will cause the rear axle to shift back and forth from side to side ever so slightly. This is typically not an issue, but if tire clearance is tight it can be a problem. An effective panhard bar should be built low, long and level (to the ground at ride height). Making the bar long and level minimizes side to side movement. Roll center height is equal to the height of the mid-point of the panhard bar. So, as the car hits a bump, RCH moves down ½ the distance the car’s body moves down.

Another option is the Watts link. A Watts link is slightly different. Two links attach to a centrally-located bellcrank that pivots as the car moves up and down, effectively making the two links act as though they had a variable length. As a result, there is no side-to-side movement of the axle: the arc of the links that would shift the position of the rear end is offset by the motion of the bellcrank. RCH is the height of the bellcrank pivot point. If the bellcrank is mounted to the frame, RCH will move up and down equally with the body in bump, making for a very predictable ride. If the bellcrank is mounted to the axle, RCH will not change at all in bump.

Our design:

We decided we would make a frame-mounted bellcrank, with a pivot point 7’’ off the ground. This will give us a 7’’ RCH at ride height. As the bellcrank and lower of the two Watts links will have to ride below this point, we can’t go too low or we lose ground clearance. For example, with a 7’’ high pivot and a 2’’ distance from pivot to lower link, we retain a little over 4’’ of ground clearance. Because the bellcrank pivots as the car bumps, the lower link actually swings away from the ground in bump, giving us some breathing room as regards ground clearance.

But what dimensions do we need for the Watts? That is, how long does the bellcrank have to be? What must the distance be between the center pivot and the attachment points for the Watts links? How long do the Watts links have to be? If the links, or the bellcrank radius, is too short, our suspension travel will be limited because the Watts will run out of room to pivot and will bind. To test our theoretical Watts in reality, as opposed to on a computer program, we decided to build a model out of cardboard and wood to see how much suspension travel we could get for given Watts dimensions. We determined that a 2’’ bellcrank radius (distance from pivot to either Watts link connection), paired with 16.5’’ long links gave us full suspension travel with room to spare, but the links would not be so long as to interfere with other suspension components.

And finally for pictures...

After taking measurements and mapping some things out in Suspension Analyzer, we began by removing the trunk pan for clearance. Since we plan to reuse the factory pan (sleeper until you actually look under the car or under the hood? Maybe...), so we laboriously drilled all of the spot welds to get it out...

198881
198882

Then, I set little Boodlefoof to sandblasting the rear axle so we had a cleaner part to work with...

198884

His little sister wanted to help too... She is so cute.

198883

[boodlefoof]

Testing Watts-link dimensions with a template:

The cardboard is serves as the frame mount. Another piece of cardboard is the bellcrank. Two sticks of wood serve as the links, while a larger piece of wood substitutes as the axle. We moved the axle through a large range to ensure there would be no binding.

198885
198886

Fabrication:

We put the rear axle back under the car and carefully measured to ensure it was properly positioned, we measured our mount point for a new crossmember to run from frame rail to frame rail to support our bellcrank bracketry. This bracketry will need to be far enough behind the rear axle to avoid interference, but remain as close as possible to minimize the length behind the axle that we will have to fabricate our axle mounts for our two links. Longer mounts (sticking further out from the axle) require heavier material and more bracing.

With measurements confirmed, we cut our crossmember to fit out of a piece of 2’’ x 2’’ x .083’’ wall tubing. To this crossmember we welded two triangular shaped pieces of 12 gauge sheet metal to locate our watts bellcrank, which will bolt between these two extra-large mounting tabs. These triangular shaped pieces are reinforced by being bolted together with spacers and 5/16’’ bolts at various points for strength. If this were a heavier car with wider, stickier, tires we would likely have used heavier material. Here, we're trying to keep weight and cost down. Overall the bellcrank bracket and crossmember weighed 11 pounds.

Little Boodlefoof fabricating the frame bracket...

198887

Frame bracket mocked in...

198888

[boodlefoof]

Fabricating the Watts bellcrank:

To make the bellcrank, we began with a piece of 1'' ID DOM tubing into which we could insert a pair of shouldered brass bushings that have a 1'' OD and .75'' ID. We cut the tube to 1.75'' (shoulders on the busings are 1/8'') so that the bellcrank would fit in the 2'' space in our mount bracket. To this tubing we welded a pair of 12-gauge plates, spaced just far enough apart for our 5/8'' heim joints to fit between them, along with a pair of heim joint dust boots, while still leaving enough space on the outside of the two tabs so that bolt heads (we had to shave them down a little) and thin nyloc nuts can fit between the mount bracket tabs.

To get things lined up and set the depth for welding on the first mounting plate to the tubing, we made a very simple jig from a block of wood.

198893
198894
198895

And in the car with the links attached.

198896
198897

[boodlefoof]

Now that the Watts link frame mount and bellcrank are in place, we turned to figuring out how to attach the links to the axle. Since we will be fabricating a 3-link, the lower link axle mounts seemed like a logical place to potentially hook into. We're already going to have a sturdy mount hanging down from the axle, might as well make it multi-purpose. The lower axle mounts for the 3-link also seemed like a good potential mounting point for the coil-over shocks too.

So, we cut a piece of 1.5'' x 2'' x .120'' tubing for the lower link mount and tacked it to the factory leaf spring perch. Then we mocked-in an upper coil-over mount on the frame and a piece of .120'' wall square tubing extending backwards from our lower axle 3-link mount to serve as a lower attachment point for the coil-over. With budget in mind, we're using a steel-bodied, non-adjustable coilover from AFCO that has 5'' of travel and a 12.5'' mounted length. We're using a 125# spring.

199044

Looks like this will work and there will be enough clearance from the coil-over for the Watts link attachment.

199045

To mount the link, we will use a piece of heavy wall 7/8'' tubing, tapped for the 5/8-18 bolt that will attach the rod end to the axle mount point. We cut this piece long so that it can be attached to the axle and lower 3-link mount along its entire length for strength. We then began welding tabs to this piece of tubing and to the lower 3-link mount, actually making our attachment double-shear.

199046

And more progress on the passenger's side... We pulled the axle out from under the car to work on it once we got the position right.

199047

[boodlefoof]

And here is progress on the driver's side. This Watts link is the lower height of the two Watts links, so the mounts are different, but we followed the same process.

Starting with the axle under the car to get the position right...

199048

Then we pulled the axle out from under the car for space as we further braced the piece of threaded tubing that the Watts link bolts into. Multiple ribs attach the tubing to the lower shock mount and lower 3-link mount. Then, a plate goes over the ribs to tie them together.

199049

And then putting the whole thing back under the car to take another gander at how this is going to look under the car (still more to be done - and yes, the body is not in its ride height position in the picture... nor is it even level!).

199050

[boodlefoof]

With the Watts link in place, we turned to fabrication of the 3-link, beginning with the upper link and building it to the specifications noted above that we mapped out in Suspension Analyzer.

First, we measured the position of our frame mount at the back of the transmission tunnel and fabricated a mounting bracket that would help spread the load across the car's existing structure in the transmission tunnel area.

199355

And tacked in place at its 15.5'' mounting height.

199357

We then made the axle mount.

199358

With the mount attached, we checked for interference. Maximum suspension bump will be limited to 3''. At full suspension bump the upper link just barely contacted the body structure. We cut a recess to provide plenty of clearance. We will box that in.

199359

With the upper link in place, we turned to the lower links. We already made the axle-side lower link mounts while fabricating the Watts link (you can just see those in pics above), which put the link mount at 6.25'' high. Mocking our lower links into place, this put our frame mount position in dead space under the car's floorpans with no substantial structure to which they could be attached. To solve the problem, we made this structure with some 1.5'' x 2'' x .120'' tubing, which will attach to the existing transmission crossmember (at the front) and the factory leaf spring mount points (towards the rear).

199369

I didn't end up getting a picture of the lower links mocked in. Will do so at some point. Then, I'm going to have to figure out the best way to tip the car to get access to weld these mounts in. I don't want to weld while laying on my back under the car - been there, done that... no fun.

[EVO]

This is gonna be great! I built a C10 with my son and loved every minute of it. Watching with interest!
E.

[Twentyover]

I was surprised to hear the car had a rear bar- This appears to be a 65-67 narrow tunnel push button door variant, but then I saw the rear axle was a post '76 with integrated swaybar mounts. It appears that in some previous life an owner replaced the banjo axle with a tube axle (for those unfamiliar w/ MG lexicon, a banjo has an drop out center section like ford 8 & 9" rears, the later tube tube has the axle tube stuck into the diff housing casting like a Ford 8.8.)

Cars from launch until I think it was 76 had no rear swaybar, roll center was as you indicate. Later cars (post 75) raised the ride height roughly 2", for Federal bumper and headlight height compliance, without additional suspension changes (They raised the body on the suspension with spacers welded to the crossmember in the front and plates welded to the rear frame stubs in the rear.) In 76 I think they added a rear swaybar as the raised CG height was causing handling issues.

May want to consider the bajo axle- same gearing (along with a few other ratios) and the dropout is aluminum- about 30 lb weight savings over the tube. There are three other tube ratios- 3.7 (Late MGC W/ OD), 3.31 (Early MGC W/ OD and MGC automatic) and 3.07 (MGC 4 syncro/no O/D and MGB GTV8)

Raced an MGB Tourer about 20 years ago- even w/ stock type suspension they can be a hoot

Need to study the thread when I have a bit more time.....

[boodlefoof]

Greg,

Yes, I believe someone did swap in a later rear end. I wish I had the banjo and looked at finding one to re-fit back under there, but they are highly sought-after and pricey it seems. I even contemplated trying to get a very lightweight Alfa Romeo rear to stick under there, but they are also kind of hard to come by at a good price. So, we stuck with what we had.

[boodlefoof]

OK. It has been a few weeks. Time for a few updates that will likely take a few posts...

First, another member contacted me to discuss the upper link axle mount attachment to the cast iron differential housing. I should have used a different welding wire to attach the mild steel to the cast iron: ENiFe-Cl wire using a 98/2 gas mix. I'll end up cutting off the mount and re-attaching with the proper materials.

A big thank you for the note! I'd hate to have this come apart on me while thrashing on it!

Second, below is the picture of the 3-link in place such as it is. I will post some better pictures from underneath the car when I can get it up in the air to complete the weld-up of the frame brackets.

200291

Third, I have the engine (pic below) and began fitting it into place to figure out suspension clearance limitations. At first, it looked like it would almost drop right in. Then, I realized the crank pulley was right where the steering rack would need to go, so the engine would have to move back several inches. As such, fitting the engine will require a substantial bit of firewall modification. It also means I'll need to re-locate the shifter forward a bit. More on engine stuff in later posts.

200292

For now, on to the front suspension...

[boodlefoof]

Front Factory Suspension:

Just as we did with the rear suspension, we began our front suspension portion of the project by reviewing what was already there. We already had in mind that the front suspension (at least the crossmember) would likely need to be modified for engine clearance. We also suspected that we could design and build something better than factory.

So, we blocked the car at (approximately factory) ride height, removed the wheels, marked the front axle centerline and made a grid on the floor underneath the car.
We then got to work with our plumb-bob and measuring devices to plot all of our front suspension points into Suspension Analyzer. Here were our observations on the factory setup:

1. Scrub Radius. One obvious issue identified is the large factory scrub radius. The scrub radius is the distance between the wheel centerline on the ground and the point on the ground where an imaginary line running down the length of the king pin (the steering axis) intersects the ground. The tire wants to rotate about the imaginary line running down the king pin. However, if the center of the tire is not on this line, the tire “scrubs” (is dragged across the pavement) by the force of turning the steering wheel. This picture explains it better than I can...

200293

A large distance between these two points on the ground (i.e., a large scrub radius) has several negative side effects. First, on a manual steering car, a large scrub radius increases the steering effort needed to turn the steering wheel, particularly at low speeds. In addition, the length of the scrub radius acts like a lever arm, amplifying through the steering and suspension the force of each bump of the wheel. This puts more stress on components and transfers more bump to the steering wheel.

After getting results from Suspension Analyzer, we double-checked scrub mathematically since we know what our tire height will be (24.8''), know the king pin inclination (7 degrees), and could measure the distance between the face of the hub and the steering axis line at the wheel centerline using a dial caliper. With our planned 24.8'' diameter tire, a zero offset wheel (like the Panasports that people frequently put on MGBs), we confirmed scrub is a whopping 2.9''. The steel wheels the car came with had 1'' of positive offset, still giving us a scrub radius of 1.9''. Still not very good. By way of comparison, Porsches and Corvettes have been keeping scrub radius to 1’’ or less for years.

One can compensate for the excessive steering effort caused by excessive scrub radius by employing negative caster (or power steering). Negative caster leads to a propensity for the front tires to want to wander, requiring more frequent steering wheel correction while driving in a straight line. We didn't want to do that. We also didn't want to go with power steering. We wanted to decrease scrub radius. But how?

One could change the spindle entirely, using one with a higher king pin inclination, making that imaginary steering axis line intersect the ground at a point closer to the hub centerline. One could also keep the factory spindle and try to replace the factory wheel hub with one that moves the wheel mounting surface further towards the center of the car. The factory hub is about 2.75'' from the edge of the inner bearing to the wheel mounting surface. We contemplated briefly whether we could find a different hub that would work since the MGB spindle uses a common Set 6 inner and Set 2 outer bearing. It actually looks like Wilwood makes an aluminum hub for the 70-78 Camaro that would fit and move the wheel mount surface about 1'' inboard. However, at $300 a pair they are a bit pricey. Also, they would change our wheel bolt pattern. While that might be a bonus since few wheels are available in our 4 x 4.5'' (or 4 x 114.3mm) lug pattern, it would mean we would need to change the rear bolt pattern too... more money.

Ultimately, we could spend an extra $400+ on hubs and wheel adapters to try and minimize scrub, but we decided instead to keep the factory hub and do what we could by changing the wheels (and altering the existing spindle – more on that later). Increasing backspacing (positive offset) of the wheels tucks the wheels further inboard and reduces scrub radius. However, wheel offset is limited – there simply aren’t wheels made in the sizes that would fit under the MGB with enough offset to get us to zero scrub radius. We decided on the Enkei RPF1 wheel, which is a lightweight aluminum wheel that actually is made in our oddball 4 x 4.5'' (114.3mm) bolt pattern. With a 43mm offset our scrub radius is reduced to 1.2’’ (and we will further reduce scrub by modifying the spindle – more on that later). In addition, with its 16’’ diameter we will have more room to fit larger brakes than we would with the factory 14’’ wheels.

We got out the Wheelrite measuring tool to see how wide a wheel we could fit. The rear is particularly tight, but we determined that a 16x7'' wheel could fit back there. The RPF1 comes in a 16x7'' size, and each wheel weighs only 14.2 lbs. Not bad. We found a screaming deal sale on these wheels and got the whole set of 4 delivered for under $800!

200294

[boodlefoof]

Front Factory Suspension (Continued):

2. Bump Steer. Another obvious issue present in the factory front suspension is substantial bump steer. Bump steer is the tendency of the wheels to change toe angle (i.e., to turn without steering wheel input) as the suspension cycles up and down. It is caused by the out tie rod's movement scribing a different arc than is ideal based on the inner and outer pivot points of the control arms. Again, a picture probably explains it better than I can.

200295

Ideally, the inner tie rod end on the steering rack should fall upon the imaginary straight line segment you could draw between the upper and lower control arm mount points on the frame in front view. The outer tie rod end on the spindle should fall upon the imaginary straight line segment you could draw between the upper and lower control arm mount points on the spindle in front view. Further, the ratio of the distance: (A) between the upper control arm mount and tie-rod; to (B) between the tie-rod and lower control arm mount on the spindle side should equal the same ratio between these mount points on the inboard (frame) side. This ensures that the arc of movement of the outer tie-rod end is equal to the arc drawn by movement of the upper and lower control arms, resulting in zero bump steer.

However, the factory MGB steering rack places the inner tie rod pivot point too far outboard and makes the overall tie rod length shorter than the effective length of the control arm pivots, resulting in bump steer that measures approximately 2/10 of an inch of toe change (unwanted steering) per inch of bump. That's a lot of bump steer. See the graph.

200298

3. Caster Adjustment and Spindle Design. The design of the spindle is worth noting, particularly as it relates to designing caster into a new suspension. When looking at the spindle in side view and when mounted to the control arms, caster is the lean of the spindle to one side or the other. If the top of the spindle leans back towards the driver, you have positive caster. If it leans forward, you have negative caster. As mentioned in an earlier post, negative caster can reduce steering effort, but increases the tendency of the car to wander. Positive caster increases straight-line stability, encourages the wheels to return to center (pointing straight) when coming out of a turn, and helps increase effective tire camber in turning. Again, a picture...

200297

The MGB came from the factory with a spindle that does not use ball joints like most cars of today (or even most cars of its day). Rather, it used a long kingpin running through two mounting points on the spindle, and about which the spindle rotates. The upper and lower control arms mount to this kingpin not with ball joints, but with thru-bolts running through some tubular bushings. So, the upper and lower control arms have to be aligned directly on top of one another to avoid bushing bind. Those bushings handle only up and down movement, while rotational movement is handled by the spindle rotating on the king pin. This makes leaning the spindle backwards to gain positive caster difficult. Instead of simply locating an upper control arm and ball joint further rearward than the bottom, the entire suspension crossmember has to be shimmed so that both the upper and lower control arms rock forward or backward (in side view) in unison to set caster (and, simultaneously, anti-dive). Further, one cannot build in caster gain (causing the spindle to lean further back during bump) by having the upper control arm at a different angle (in side view) from the bottom control arm. Again, you would bind the bushings.

We wanted to build in caster and adjustability. The bushings would have to go. More on that later.

4. Other General Observations. Apart from the issues above, the front suspension at this ride height showed a roll center height of about 0.5’’, camber gain of -0.45 degrees per inch of bump (or roll), approximately 50% anti-dive, and a front-view swing arm length of 183’’. More on these items when we discuss the re-designed suspension...

[boodlefoof]

Designing and Building the Front Suspension:

With the engine more-or-less in place and having selected our wheels*, it was time to begin designing the front suspension. We would work our way from the outside in, while concurrently running numbers on Suspension Analyzer.

* Quick note on tires. We decided on the 205/55R16 Hankook Ventus V12 evo2. This tire has a section width of about 8.4’’ on our 7’’ wide rim and is 24.8’’ tall. It weighs 19 pounds and features a somewhat soft rubber compound (UTQG rating 320 AA A), but not so soft that this tire can’t be driven in colder weather - think Maine in any time other than June - September. We contemplated the Dunlop Direzza ZIII in the same size as an alternative, but with its ultra-soft compound and the propensity for cold nights in Maine even during driving months, we opted against it.

All together, this 16’’ wheel (at 14.2 lbs each) and tire (at 19 lbs each) combo comes in at 33.2 pounds per corner. That is a full 5 pounds lighter at each corner than the 14’’ steel wheels and 185/65 tires that came with the car.

Beginning the Design.

We began by mocking our front wheels in place under the car. Even with the stock steel wheels, the tires looked tucked too far in under the fenders. We decided we could widen the track width to 50.5'' (up from the factory 49'').

We then turned our attention back to the spindles. As we mentioned before, the king pin and bushing design was not ideal. By only accommodating up and down movement, the bushings really limited our ability to build in caster and caster gain, and adjustability generally, to the suspension design. While one could leave the spindle and king pin relatively stock and just replace the rubber bushings with “uniball” type high-misalignment spherical bearings, we decided to eliminate the kingpin and trunnion entirely and replace it with a ball joint at the bottom of the spindle and a heim joint where an upper ball joint would otherwise be. In addition to providing the freedom of movement we would need for caster, the use of the ball joint and heim joint would allow us to obtain other benefits by increasing the spindle’s height, giving us better roll center characteristics and camber gain.

We fit the existing spindle into the wheel to see what kind of space we had to play with at the top and bottom of the spindle if we were to modify it to make it taller. Spoiler Alert - new front disk brakes in picture (more on that later).

200304

We noted that there was actually quite a lot of clearance to make the spindle taller.

Spindle Modifications

We began by disassembling and measuring the existing spindle.

200306

200305

In its factory form, the spindle has a 7 degree king pin inclination and the distance between the upper and lower mount centerlines of the rubber bushings (the spindle height) is 8.75’’.

Due to the amount of axial load the lower spindle attachment point will experience when hitting a pothole or something similar, we opted to use a ball joint on the lower mount. The upper mount does not see the same kind of axial loading because the coilover shock will not be attached to it. So, for the top mount we decided to use a heim joint on a stud that we would mount in the old upper king pin hole. By using a stud, we can fine tune the height of the spindle for roll center migration and camber gain.

But first, to fit a ball joint in the lower hole we had to find a way to put a 7 degree tapered hole of the right diameter into our approximately 1’’ hole. Fortunately, weld-in tapered inserts are available and we found some at TMR Performance. These come in a variety of sizes. We opted for a 1’’ outside diameter sleeve with a 1.5’’ over 12’’ (7 degree) tapered inside diameter because it would drop right into our hole. This insert is designed to fit a Moog K719 style lower ball joint with a .780’’ large end on the tapered stud. It drops right in to be welded in place.

To extend the lower mount position downward (increasing the distance between upper and lower pivot points on the spindle, making it taller), we chose a screw-in ball joint from QA1 with a ½’’ longer stud. With parts in hand, we tested our fit and clearance.

200307

On the upper end, we needed a way to attach a spherical rod end. We decided to use a piece of 0.875’’ outside diameter DOM tubing with a .120’’ wall and slide it into the upper mount hole. It fits with just the slightest bit of play. We would use this play to our advantage, angling the piece of tubing inward (towards the center of the car) at the top. This allowed us to increase KPI to 8.5 degrees (putting our scrub radius under 1’’). Welding that piece of tubing into the spindle’s upper hole would give us a through hole that is the proper inside diameter to fit a 5/8’’ bolt on which our rod end can mount. By cutting the piece of tubing longer than the length of the upper mount hole in the spindle, we can tailor the height of the spindle after modeling our suspension on the computer. To further reinforce the strength of the .120’’ wall tubing insert we would then slip a second piece of tubing (internal diameter of 0.875’’, outside diameter of 1.25’’) over the first piece and weld it to the spindle as well.

Having already begun modeling the suspension on the computer (more on that later) and determining the preferred height for our spindle, we increased our spindle height from the stock 8.75’’ to a modified height of 11’’.

With the design set it was time to modify the spindles!

[boodlefoof]

Designing and Building the Front Suspension (Continued):

To modify our spindles, we first sandblasted them.

200308

Then, we welded in the first upper stud tubes and ball joint inserts.

200309

Then, we welded the second (reinforcing) upper stud tube over the first upper stud tube and mocked things together.

200310

Then, we painted it up to make it pretty and fit it back in our wheel to verify clearance. We have plenty of clearance. Just need to trim some threads off the long upper bolt and these are done!

200311

200312

[jaybee]

This is a very cool project. All the more so because it's a platform with virtually no pro touring support. You'll have plenty of horsepower for such a tiny car, should be incredibly fun.

[boodlefoof]

Long time no posting... things have been busy, but progress continues slowly. We left off on posts at front suspension design.

With the wheels and spindles mocked in place, and our front suspension design considerations in mind, we set about the iterative process of modeling out the front suspension. We began by inputting into Suspension Analyzer our fixed variables: tire diameter and track width, lower ball joint location, and upper control arm mount location. We then worked inward, using the ranges of space available to us for frame-side upper and lower control arm mounts.

We ultimately came up with a design that uses a 13.5’’ lower control arm and 9’’ upper control arm. The LCA is angled down very slightly towards the ball joint from the frame, while the upper control arm is angled upwards from the frame to the upper control arm mount. The result is a RCH of 2.88’’, and 0.78 degrees of negative camber gain per inch of straight bump, and caster gain of 0.33 degrees per inch of bump, while maintaining an 83’’ FVSAL at ride height; a little short, but not too bad. Because the ratio of the control arm lengths to one another is optimized for our given mount points, FVSAL change as the suspension movement is slowed, resulting in it remaining greater than 48’’ throughout maximum suspension travel. By way of comparison, the C5 Corvette has an FVSAL of over 300’’ at ride height, but that number drops to 65’’ at maximum suspension travel.

Also because of our optimized control arm lengths relative to one another, lateral RC migration is virtually eliminated. The C5 Corvette sees a maximum of 5’’ of lateral RC migration in a brake and turn scenario. Our design sees a maximum of approximately 2’’ of later RC migration in the same scenario. The picture below shows RC migration in a this scenario. With the brakes applied (giving 1.5'' of front suspension bump), the highlighted numbers show likely RC lateral migration outputs as steering input increases and body roll increases as a result.

211969

With about 5 degrees of positive caster at ride height (resulting in 1.2’’ of caster trail – time will tell if that is too much), and minor caster gain in bump, our camber curve on both the inside and outside wheel should keep the tires in a good position for maximum contact with the pavement in a roll/turn scenario. With increased caster comes increased steering effort, so we plan to retrofit a Saturn Vue electric steering assist to the car in the future.

On steering, we also paid close attention to bump steer and rack position. We were able to get bump steer down to 0.010'' of steering change across 5'' of suspension travel. The image below shows the bump steer curve - it looks wavy, but keep an eye on the Y-axis units of measure. That is change in tire direction in inches.

211970


Physically modeling the suspension, we also determined that we could mount the coilover spring to the lower control arm very close to the lower ball joint, reducing the wheel’s leverage over the spring. This let us run a relatively light 225# rate front spring rate to achieve a front suspension frequency of 1.5hz, right where we want it to be for our performance street car that will see some track time. Paired with our 125# rate rear spring, giving our rear suspension a 1.7hz frequency, our front and rear frequencies should be well-suited to one another. Using a 1’’ hollow front sway bar and ¾’’ solid rear sway bar with adjustable arm length, we can dial-in front/rear roll characteristics for neutral steer. As modeled, with the sway bars, the car will experience 2.15 degrees of roll per G of later force. This thing is going to handle like a go-kart!

[boodlefoof]

With the suspension designed, we moved on to physical mock-up to visualize what we had to build and how it would fit. We mounted the spindles on a jig and slid them into place under the car. Then, we made dummy control arms out of wood and mounted them to the spindles. We also mounted the steering rack to the jig to check engine clearance.

211972

211973

211974

211975

[boodlefoof]

Then, we began actually building the pieces. The control arms were made out of 7/8'' heavy wall tubing that we could bend in the tubing bender and tap for 5/8'' heims. We made a jig to measure the proper bend up and sizing of the arms. Here is a lower arm set in the jig for fitting.

211976

We made a removable frame crossmember out of some 3/16'' angle steel that would bolt to each frame rail, and which would be joined together by some 1.5'' x .095'' DOM tubing that would serve as the structure to which everything would attach. We routed the crossmember tubing just in front of and just behind the oil pan.

211977

211978

With the crossmember roughly made, we leveled a work table to be our fabrication space and marked a grid of dimensions on it. We then transferred the crossmember to the table and mounted it to the table to keep it in place while we fabricated all of the mounting tabs for suspension points. We also mounted jigs to the table to hold the spindles and lower control arm pivot points in place to make it easier to locate the necessary brackets onto the crossmember.

211979

211980

[boodlefoof]

Progress continued...

211981

211982

Until we ended up with a crossmember looking about like this (still more gusseting to be done).

211983

Then, we mounted it under the car!

211984

211985

Then, we put the wheels on.

211986

211987

If I had it to do over, I would have angled the ball joint threaded sleeves to better match the KPI. However, we moved the suspension through its full travel and the ball joints have plenty of space to move without any potential binding.

[jaybee]

Looks great. Love the way you've worked hard to analyze what's already there and work within it rather than gutting it to the shell and more or less cutting and pasting an aftermarket solution. Nothing wrong with that, but it's more or less a cubic money approach. There are far more people who don't have six figures to drop on a toy than those who do. Besides, I like watching the thought process.