Best-Engineered Cars of Lemons – The FrankenMiata
Part III: Making it handle with a hacksaw.
by Dave Coleman
The FrankenMiata is a handling machine. Sure, it slices through fields of $500 Craigslist rejects, but even when you measure it against real cars, this pile of Mazda junk stands tall. Early in its development, it beat the slalom speed of a brand-new 911 Cabrio on a set of bald, flat-spotted $60 Falkens, and months of development later, it works even better. So what's the secret? Cheaty coil-overs aged with a spray can of undercoating? Poly bushings hidden under a layer of old rubber? Hardly. Cheating is easy, we race with artihmetic.
Springs
Most Lemons amateurs seem to think lowering their car is the key to cheap handling success. Though a lower center of gravity certainly does some good, constantly pounding into your cracked old bump stops does vastly more harm. Cars behave very unpredictably when bouncing around on their bump stops, and predictability is key to not crashing into the car next to you.
We managed to increase our spring rate 50% without lowering the car an inch, and without spending a penny. Here's how:
This miracle of suspension tuning takes two simple steps. First, you stiffen by removing active coils from the spring. Second, you restore your ride height by raising the spring perch. The details, of course, are far less simple.
 |
| Our rear spring started out 86 lb/in. After cutting 1.5 coils, it was 107 lb/in. After setting it on a Bilstein spring perch, it was 128 lb/in. |
When you cut a spring, the spring rate increases directly in proportion to the number of active coils you remove. If you have a 100 lb/in spring with 10 active coils and you cut 1 coil off, your new spring rate will be 100 x 10/9, or 111 lb/in. Simple. (Watch out for the “active” part of that explanation. Coils that are stacked solid at ride height, as you'll find on any progressive rate spring, or coils that are sitting still on a spring perch aren't active.)
Obviously, when you cut a spring, you also lower the car. How much? That's also reasonably easy to at least guesstimate. A spring will compress by the weight put on the spring divided by the spring rate. Put 100 pounds on a 100lb/in spring, and it will compress 1 inch.
So all you have to do is figure out how much each corner of the car weighs and what each spring rate is, and you can figure out the ride height, right? Not so fast.
Sadly, the spring almost never moves in a 1:1 ratio with the wheel, and this vastly complicates the math. Before you calculate anything, you're going to have to figure out the motion ratio between the wheel and the spring. You can do this by measuring pickup points on the suspension, figuring out the angle at which the spring is mounted, and doing some math. Or, since you have the suspension apart already, you can just measure it directly.
 | ||
| Thanks to the shock's mounting point a few inches inboard on the lower control arm, and its angle from vertical, the front spring only moves 0.69 inches for every inch of wheel travel. |
Once you remove the spring you're about to cut anyway, you can reassemble the suspension without it and you'll suddenly have the ability to move the suspension through its entire stroke with relative ease. Figure out some way to measure the wheel's position relative to something directly above or below it (any angle will throw off your measurements), and the distance between the spring perches. Now, move the wheel an inch and see how much the spring perch distance changes. This, in theory, will tell you your motion ratio, but with such a small measurement, any small measurement error will have a drastic affect on the final number. The motion ratio also changes slightly through the suspension travel, since the angle of the spring will change slightly. To minimize both these problems, try calculating the motion ratio based on the the full travel instead.
On a Miata, the front motion ratio is 0.69:1 and the rear is 0.75:1. So, if you put 100 pounds on the front suspension, the spring is going to see 145 pounds (100/0.69).
Next, we took our springs to Muellerized and borrowed their spring tester to figure out our starting spring rates. Our front springs started out 160 lb/in and our rears a paltry 86 lb/in.
Now, let's focus on the front spring to see what we did. Knowing the spring rate (160 lb/in), the number of active coils (5.5) and the unloaded height of the spring (11.25 inches), the first thing we did was take all this info and stick it in a spreadsheet. Then we calculated the spring's loaded height, with the car's weight sitting on it.
Our car, stripped of luxuries like a hood, adjustable seats and frivolous fenders, and weighed down with a roll cage, a Dodge turbo and a driver, tipped the scales at almost exactly 2,200 pounds. With nearly perfect 50/50 weight distribution, each tire has 550 pounds sitting on it. From this, we had to subtract the weight that wasn't sitting on the spring. The 14″ wheels, skinny Falken tires, teeny brakes and dainty double-wishbone suspension. We just guessed all of it would add up to 50 pounds, leaving 500 pounds of sprung weight.
At the spring, that looks like 725 pounds (500 pounds / 0.69 motion ratio), which means the spring is compressed 4.5 inches (652 lbs / 160 lb/in.), leaving it 6.7 inches long (11.25 – 4.5 in.) at ride height. Since we didn't want to lower the car at all, 6.7 inches was our bogey.
Next we tried punching in various options, like chopping off half a coil, one coil, two coils, etc., noting how much each increased the spring rate, and how much each shortened the loaded height. For this calculation, you again have to figure out the free height, by subtracting the height of the active coils you cut off, and the height of the dead coil that's wound flat at the bottom of the spring. That dead coil is simply as tall as the wire diameter (0.4 inches), then calculating the new spring rate and how much the spring will compress with 725 pounds of load on it.
 | ||
| Bilstein spring perches don't fit on the base model shocks. We had to open up the ID of the perch a bit to clear the flange on the top of our shocks. |
With that calculation automated in the spreadsheet, we next started thinking about how to move the spring perch. Fellow MotoIQ nerd Chuck Johnson, who spends his days as an engineer at JE Pistons, stiffened his B13 SE-R LeMon springs by cutting them and then raising the car back up to stock ride height with an elaborate aluminum spacer he machined from a discarded WWII airplane piston. We neither had, nor needed such an awesome solution. Mazda already provided something easier.
The Miata's optional Bilstein shocks use a cast aluminum spring perch with a spiral spring seat designed for an open-coil spring. Robert Davis, a fellow Miata racer who also happens to be Executive Vice President of Mazda R&D (and thus my boss in a big way), had offered us a set of stock Bilsteins from his vast bucket of spares. Knowing the LeMons BS judges would have a seizure at the sight of any yellow suspension parts, we declined on the shocks, but took the aluminum perches.
 | ||
| Our shocks also had a lip above the spring perch designed to hold the bottom of the dust boots. We had to machine this off as well. |
An open-coil spring sitting on one of these perches loses one active coil, since the bottom coil is supported by the perch and no longer has to bend when load is applied. That means we could take our 5.5-coil springs, cut one coil off, and set them on these aluminum perches for a total of only 3.5 active coils. That bumped our spring rate from 160 to 251 lb/in (160 * 5.5/3.5), an impressive 56% increase in rate.
And, since the Bilstein spring perch doesn't sit properly on the base model shocks, you gain an extra 3/4-inch of spring height. Add it all up, and the whole thing stacks up to 6.8 inches at load. That's 0.1″ off from our stock bogey, or damn well close enough. Given the number of estimations made during this calculation, we'd be lucky if this calculation is within half an inch of reality, but even half an inch is pretty good for LeMons.
The same process was followed at the rear, but 1.5 coils had to be cut from that spring to achieve the same 50% increase in spring rate and stock ride height.
 | ||
| The final result is hideously elegant. The gap between the cut end of the last coil and the spring perch is the 0.75-inches the Bilstein perch adds to our spring height. The gap between the spring perch and the bottom coil, on the left, shows one of the many small errors in our calculation. We calculated the spring rate as if the entire coil were inactive. Clearly, only about 2/3 of the bottom coil are inactive at full droop. As the spring compresses, more and more of that last coil will hit the perch and go dead, giving us a slightly progressive spring rate, and a lower ride height than we calculated. The bolts wedged under the perch were the only thing within reach that could be used to make the spring perch sit level. Turns out they work fine and have lasted through 94 hours of endurance racing and roughly 50 hours of track testing. |
