This is *G o o g l e*'s cache of http://www4.ncsu.edu/~ndhuffma/Miata%20Suspension%203125.htm as retrieved on Jan 4, 2006 07:07:55 GMT. *G o o g l e*'s cache is the snapshot that we took of the page as we crawled the web. The page may have changed since that time. Click here for the current page without highlighting. This cached page may reference images which are no longer available. Click here for the cached text only. To link to or bookmark this page, use the following url: |http://www.google.com/search?q=cache:u9dmJsvhMn0J:www4.ncsu.edu/~ndhuffma/Miata%2520Suspension%25203125.htm+&hl=en&client=firefox-a| /Google is neither affiliated with the authors of this page nor responsible for its content./ ------------------------------------------------------------------------ The Miata Roll Center By Nic Huffman *Disclaimer:* In the absence of decent legal advise, I recommend that you do not modify your car in any way. Always wear your seatbelt and obey traffic laws. Hell, wear a helmet. The author is not connected in any way to Mazda North America or any company whose products are mentioned. Said author is not a professional vehicle dynamicist; he is a chemist. It is a testament to just how out of hand the American judicial system has gotten that I feel the need to include a disclaimer. The following information is intended to be educational and to help those who disregard the above warning avoid some problems and expenses I have incurred. Information taken incorrectly or incorrectly implemented could result in injury or death, so it is best to double-check any proposed modification with the references in my review of automotive literature or in your local library. Your safety is ultimately your responsibility, not mine. It is also upon you not to injure people or property with any information taken from this site. Finally, the author is too broke to be worth suing or he would be writing this about the Elise. *Abstract: *The M1 Miata suspension is good in stock form and has qualities that make it work very well when modified. The geometry is a good compromise for racing, and can be improved via springs, dampers, bushings, etc., without any welding or fabrication. This leads to a car with phenomenal handling characteristics in classes that do not allow changing control arms and attachment points such as the CSP class. This page is a compilation of first generation Miata information and an overview of performance modifications mainly for street-legal cars. I hope that the information on this page will serve as a starting point for those wanting to modify their Miatas AFTER learning a little suspension theory. I have included a lot of my experiences and the general experience of the Miata.net forums. I also include sections on testing to make sure our theory is working. I make an effort to use indefinite wording when dealing with hypothetical situations (X /should/ cause Y) to keep things scientific. While I have 80k miles, 5 sets of tires, 8 autocrosses, and several dozen books worth of experience, my time and funding have been limited. Hopefully, I will get a job with a tuning company and get paid to work with vehicle dynamics. *Stock Suspension: *The M1 Miata in stock form has roll centers located 61mm above the ground plane at the front and 120mm rear. These figures are just a bit higher than suggested as a basic guideline for purpose-built racers in A. Staniforth's 'Competition Car Suspension.' This is probably a good compromise since the center of gravity of a street-based racer is higher than that of a true racer. These roll center heights are biased to decrease jacking at the penalty of rolling more in turns. The equivalent swing arm lengths are in the middle of the range of what is commonly used according to Staniforth. This is a good balance between camber control in roll and in dive, i.e. the suspension is compromised between providing optimal camber in steady state turning and optimal camber for braking and accelerating. The equivalent swing arm length at the front is a bit shorter than at the rear. This produces less camber loss at the front when the chassis rolls. The rear suspension is given a bit more static negative camber to balance this out. The OEM springs are soft (2.79kg/mm F, 1.68kg/mm R or 156lb/" F, 94lb/" R) and with motion ratios of 0.692F, 0.766R, provide wheel rates of* *1.33kg/mm F, 0.99kg/mm R (74.7lb/"F, 55.2lb/"R). The suspension bushings add to the spring rate. Mine measured 0.6kg/mm F/ 0.9kg/mm R at the wheel, bringing total wheel rates up to 1.93kg/mm F and 1.79kg/mm R. The springs are also supplemented by rubber bump stops. My 95 came with the older 48mm bump stops in the rear and the newer design 60mm bump stops in the front. The R package used the longer bump stops at both ends. At the limit, a stock Miata will be riding on the bump stops. The combination of spring and bump stop yields a smooth ride when traveling at moderate speeds which firms up when driven more aggressively. The spring/bump stop combination acts as a progressive spring. This set up works well with the narrow, tall sidewall tires that come on the car. My 95 came with thin spoke wheels labeled 14X6. The manual shows a 14X5.5 wheel size (thick spokes) and an optional 15X6. The stock 185mm wide tires provide more tread width per ton of vehicle weight than many sporting cars. *Baseline: *Improving any suspension should begin with making sure everything is working as intended by the factory. Your tires must be of fairly high quality and properly inflated to produce much cornering force. Stock shock absorbers (the black Showas, not the yellow ‘R package’ Bilsteins) are known to go bad within 30k miles and are really horrible at 60k. A Miata on blown shocks will feel darty in turns and you can feel the car smash into the bump stops when slalomed. This is not confidence inspiring and I suspect it has really hurt resale of these cars. The stock suspension bushings stiffen over time, which improves handling until they become so hard that they tear apart. Bushing wear occurs with age and especially with hard use. Bad bushings are not overly easy to diagnose, as they usually don’t cause noises or other red flags to tip you off. The alignment should also be checked to make sure it is not causing bad behavior. Some tire shops will check your alignment for free in hopes of selling you an alignment. Comparing your alignment numbers to those in your manual, you should hope for maximum positive caster and maximum negative camber listed. The numbers need to be fairly even side-to-side. From a wear perspective, you do not really need an alignment unless the car pulls to one side under neutral throttle or under braking or if you have uneven tire wear. *Target:* Now that your suspension is up to stock spec, you should define exactly what you want the car to do. A car cannot be a luxurious status symbol, a serious street machine, and a dedicated track car at the same time. Most cars are somewhere in between. The idea is to change parts out to improve certain elements of the car as opposed to buying every item in the catalog and assuming they will work together. There is no magic upgrade that will instantly transform a car. A number of well thought out enhancements will transform a car. Many of these enhancements are expensive, but I will try to include as many cheap/free upgrades as possible. Until we learn how to change the laws of physics, every machine is a collection of compromises, and the best we can do is to change those compromises to suit us. For example, a car optimized for performance in dry conditions can be sketchy in the rain and completely undrivable in the snow. *My Miata:* My goal was to make my Miata as fast as possible given my funding level. My car is not set up for a specific class, but I could make it CSP legal fairly painlessly. I wanted a car that would help me learn driving skills, so my modifications have mostly been aimed at improved cornering performance. As of mid-2004, my set-up is as follows: 95 ‘popular equipment package’ with most of the popular dead weight removed. There is no power steering, cruise control, windshield washer, spare tire, jack, or audio equipment. I have fully gutted the trunk and removed all heat shields. The sound deadening is gone, but the stock carpet remains in place. I installed a Hard Dog Hard Core roll bar, Simpson 5 point harness, and Ultra Shield rally seat all in the name of safety. Kumho MX’s are mounted to Team Dynamics 15x7x40mm wheels that weigh 13.2lbs each. Braking is accomplished by Hawk HPS+ brake pads acting on stock rotors. The calipers are filled with Castrol LMA fluid supplied by Goodridge woven stainless brake lines. Suspension consists of Koni Sport shocks wrapped in Ground Control adjustable perches with Eibach springs rated 8.04kg/mm F, 5.36kg/mm R (450lbs/in F, 300lbs/in R). The rear suspension rides on Flying Miata upper shock mounts, while the front retains the stock mounts. Koni bumpstops are used at all 4 corners. Roll is restrained by a 15.9mm Racing Beat bar out back and a hollow 27mm Jackson Racing bar in the front mounted on a Racing Beat reinforcement kit. The bars connect to the control arms via Moss Heim jointed end links. The suspension is set to minimum stock ride height to retain shock travel and keep bump steer reasonable. Alignment is set at 5.0 degrees caster, -1.25 degrees front camber, and –1.75 degrees rear camber. The engine remains stock, but breathes through a Jackson Racing cold air intake. The CAI, throttle body, and intake manifold have been mildly ported to remove casting flaws. The exhaust exits via a rusty Brospeed header, Flying Miata cat, and Thermal cat-back. The flywheel is a Fidanza unit meant for the 1.6l engines that weighs 3.2kg (7.1lbs). The clutch and pressure plate are from Racing Beat. The rear tires are driven through a stock Torsen LSD. I have yet to cornerweight the chassis, but I estimate a weight of 2250lbs without driver. *UPDATE 12-8-04: *Nearing the 100k-mile mark, I pressed new SuperPro brand polyurethane bushings into my control arms. I decided to cornerweight the car, so I removed the anti-roll bars, their end links, and the RB reinforcement kit. I replaced the stock front bar, stock bushings, and stock end links. The car weighs 2399lbs with me in it. I account for 200 of those. The cross weights are within 15lbs of each other. The LF corner is a bit heavy, resulting in 52% front weight bias and 52% left weight bias. The car is now aligned to 3.5^0 caster, -1.5^0 front camber, 1/16” front toe out, -1.8^0 rear camber, and 1/16” rear toe in. There is a bit more understeer now, but traction coming out of tight corners is really amazing. The transitional response is about the same as the mid-2004 setup. The new spec is very confidence inspiring and breaks traction more gradually than the previous setup. I intend to use the new spec for a while for evaluation and testing before moving back to stiffer bars. *UPDATE 1-7-5:* Reasoning that the reduced load transfer of my newest setup should require less tire pressure, I road tested my car at 28psi. I found that the tires gained about 10^o C on a ~15^o C day with fairly even temperature across the tread. There was no significant roll over, despite repeated instances of understeer and oversteer. The car understeers enough that I think I will reinstall the stock rear bar. *Rubber: *Tires are the focus of any road vehicle's handling ability. Every part on a vehicle should be designed to put the tires in an optimal operating condition. Tires produce the most grip when they are perpendicular to the ground at optimal pressure and temperature and have the least load on them that you can arrange. Tires deform laterally when cornering loads are applied. This deformation is a detriment to producing cornering power. This can be minimized by internal design, sidewall height, and rim width, inflation pressure, etc. High performance tires have stiffer sidewalls, which resist distortion of the carcass. Shorter sidewalls tend to be stiffer and resist distortion to a greater degree. Using wider rims (no wider than recommended by the tire manufacturer) will also reduce distortion. Optimizing these wheel specifications results in a more uniform loading of the contact patch and more grip. Every car’s handling can be adjusted by changing tire inflation pressure. For performance applications, the first order of business is to eliminate roll over. Most local level racers will use stripes of white shoe polish on the sidewalls of their tires to show how far the tires roll over. I prefer to use a product called Dial-in from Greddex, which comes off a bit more easily. With either product, you are hoping to wear the mark off on the tread, but not off the sidewall. Inflation pressure is increased in increments of a few psi until rollover is eliminated or the maximum pressure listed on the sidewall is reached. You will probably want to use the lowest pressure that eliminates rollover. Higher than necessary inflation pressures will increase the ride rate, but will decrease cornering performance due to a smaller contact patch and lower heat generation. This stiffening effect makes the car feel like it is cornering better, but traction is actually reduced. The correct way to stiffen up a suspension is to install stiffer springs, anti roll bars (ARB’s), shock absorbers, etc., not running 40+psi. Mazda recommends 26psi at all four corners for the NA Miatas. The Miata club of America recommends 28psi all around. I ran around 36psiF/34psiR when I had 195’s on 6in rims and now run 34psiF/32psiR with 205’s on 7in rims. Tire pressure increases by 1psi for every 10 ^0 F gain in ambient temperature, and also increases with use. I try to inflate my tires when they are still cold in the morning. Wider rims support tires better, allowing lower inflation pressures, while maintaining low rollover. This produces higher grip. The general rule is to use a wheel 25mm (1in) narrower than the tire’s listed tread width. You can use wheels up to ~50mm (2in) narrower than your tire at the expense of support, but only do so if competition rules or monetary limitations apply. Since we wish to use tires of approximately 23in overall diameter, sidewall height will be a function of rim diameter. Larger diameter rims will of course decrease sidewall height, firming up the ride and increasing grip up to some point. Sidewall heights are measured in a really backward way. The ‘series’ number (the ‘50’ in the tire size 205/50R15) is the height of the sidewall as a percentage of tread width. Accordingly this size tire should have 102.5mm sidewalls (205mm x 50%). A section cut from my old ES100’s in the same size actually measured more like 85mm. To get back on track, this measurement standard complicates comparison of tires of different widths. As an example, the 205/40R17 tires on the Mazdaspeed MX-5 have ~82mm sidewalls, where the 265/35R18 tires on the rear of a Boxster would have ~93mm sidewalls. Many of the best handling road cars today have a calculated sidewall height in the 90-105mm region. Many dedicated racecars use shorter sidewalls, but they run on much smoother surfaces than streetcars. The benefits of short sidewalls only apply up to some point, after which weight, lack of compliance, lack of sufficient load rating, etc. conspire to reduce performance. In my untested opinion, wheels over 16in in diameter will not make a Miata faster. Big wheels are for bling, not performance. Another problem with short sidewall tires is a lack of torsional compliance. A sudden stab on the accelerator is more likely to cause wheel spin with short sidewalls. Taller sidewalls can twist a bit more, and in general, should resist spinning better. Drag racers have known this for a long time, and even the wrong-wheel-drive ricer drag crowd uses tall sidewalls. A common misconception is that wider tires give you more grip due to a larger contact patch. Contact patch is actually a function of inflation pressure, and not width. Lower inflation pressures yield larger contact patches. Of course, if you try to lower your tire pressures too much you will incur rollover and are also more likely to suffer a blowout. Wider tires generally produce less heat, since the contact patch does not deform as much. This allows the use of softer rubber compounds, hence the increase in grip. It should be noted that lower tire pressures, softer rubber compounds, and stiff suspensions generate more heat in the tires. Tire width can be compared between different cars (somewhat unscientifically) by dividing the weight placed on the tire by the width of that tire. Performing this calculation shows that the Miata comes stock with a very wide tire for it’s weight. Upgrading to a 205mm wide tire gives you more tire width per ton than many Porsches and Ferraris. Differences in tire loading between front and rear can help you understand the chassis tuning a bit better. Some high powered cars, for example, will have more tire width per ton in the rear to decrease throttle induced oversteer. Differences in tire loading between axles may suggest stronger roll resistance at the more lightly loaded end and/or an effort to avoid the need for an LSD. Tire loading also explains why front-heavy FWD cars with equal sized tires at each corner are so prone to understeer. See my ‘Tire Data with Pithy Comments’ page for data on several high performance models. The pithy comments are accessed by moving the cursor over the red tabs in the corners of the cells. Paul Haney’s “The Racing and High Performance Tire” is the best book I have read on tires and is where you should start if you need more information. *Rubber I’ve Abused: *My car came with OEM 14X6 wheels, so I used the popular *Dunlop SP8000* tire in 195/55R14 size. These tires were a nice upgrade from worn out stock rubber. The grip was improved and tire chirp pulling away from stoplights was eliminated. These tires have large tread blocks, which resist distortion and tearing. Small tread blocks as on Michelin MXV4’s tend to tear off when pushed hard, drastically reducing tire life. This is not accounted for when a wear rating is assigned for a tire. I found the Dunlops to last as long as the French tires, but with better traction. Long life tires only live long if you drive like grand ma (no, not into a tree). Small tread blocks are used for wet weather and grip in snow. I went through 3 sets of SP8k’s, obtaining mileages of 24k to 14k miles (my driving style changed just a bit in this time). Eventually, I changed to 15x7” wheels, along with *Yokohama ES100’s* in the 205/50R15 size. This size has a larger diameter and the reduction in acceleration (~5%) was apparent. These tires improved a lot after the first couple hundred miles. Grip was good until they wore to their wear bars (at 10k miles) at which point they turned to stone. Pyrometer testing of the petrified tires in constant turning showed that the tires would work best at ~32psi and more camber than –1.25F/-1.75R. Of course, these values were only valid for that particular test on my car and may not be applicable to your car or conditions. I try to buy tires in the fall so that the extra heat generated by the full treads is countered by cold ambient conditions. I was surprised to find that I couldn’t get these tires above luke warm in the winter without really abusing them. My current (2004) tiers are *Kumho MX’s* in 205/50R15. These grip better than the ES100’s. I have personally heard a lot of complaint about Kumho’s, but in every case in was in relation to the 711/712 series. According to Tirerack.com, the MX’s produced only 0.02g less lateral acceleration than the reigning Goodyear KD’s on a fat and soft Lexus. The MX’s seem to heat up somewhat better than the ES100’s in the cold. Pyrometer testing with scrubbed-in MX’s again pointed to not enough negative camber, along with tire pressures of ~34psi all around. *Suspension Springs:* One of the first things to address when planning a suspension is spring rate. Overly stiff springs reduce mechanical grip and make ride uncomfortable. Soft springs yield a sloppy ride and can be fully compressed by the car’s roll or pitch, leaving no suspension travel to soak up bumps. Overly soft springs also allow the suspension to move to less favorable (poor camber) portions of the range of travel. Many modern cars use soft springs working in series with long bump stops. This provides a progressive suspension that retains some travel at increased rate during hard cornering/braking. This is great for the non-enthusiast, but less than ideal for the go-fast crowd. This regime is representative of the stock suspension as well as suspensions using soft lowering springs. I do not know if substantially stiffer lowering springs also follow this regime, but I suspect they do. As a point of reference, I consider soft lowering springs to be around 30% stiffer than stock and stiff lowering springs to be around twice as stiff as stock. Spring rates can only be compared directly for the same vehicle. The spring rates used on different vehicles are completely irrelevant to each other, as the stiffness of a suspension is a function of leverage ratios, sprung weight, etc. To compare the stiffness of suspensions on different vehicles we have to talk in terms of natural frequency, the units of which are cycles per second or Hz. A car with a ~2Hz suspension will be fairly stiff and a car with a ~1Hz suspension will be overly soft. Fred Puhn’s book gives some first-order equations for natural frequency should you feel like estimating your set up. Note that these equations yield a higher frequency figure than the actual figure since they do not account for the spring rate of the tires, etc. Fortunately, if we are talking Miatas, we can directly compare spring rates with only one hiccup. That hiccup is weight. If you strip your Miata down to 1800 lbs, it will ride much stiffer than a fully optioned Miata with a 75lb turbo and 25lbs of subwoofers, given the same springs. If this is not working for you, picture fat Jared bottoming out and then some on his pogo stick vs. thin Jared barely compressing the spring on his now much happier pogo stick. You may also notice that a sedan is smoother with 4 occupants, but crisper with just the driver. The word “progressive” has been used to sell a lot of springs. I am not at all sold on the idea that off-the-shelf progressive springs really improve handling. Progressive springs make the car’s response non-linear, which may hinder the driver’s ability to feel what the car is doing. This will negatively impact speed, especially given the skill level most of us are at. Additionally, the dampers have no way of altering the damping force with the change in effective spring rate, further confusing things. Note that a Miata whose bump stops come into effect already has a progressive suspension. If you add progressive springs you now have a suspension with 2 sources of progression. The graph of spring rate versus load would be far from linear, and I cannot imagine that helps drivers feel what they are doing. Most progressive springs can be identified by a change in the spacing between coils or a change in the outside diameter of the spring as in the ‘barrel’ spring. Note that the FM springs are not progressive in the normal range of travel since the closely spaced coils are fully compressed as installed. The only time I could see a valid case for using progressive springs on the Miata is on the rear suspension of a very powerful car. Done correctly, this would take a lot of calculating and testing. It would also likely take the form of 2 separate springs mounted in series with a special spacer holding them in alignment. *The Case Against Lowering:* You will never see this in the magazines since the manufacturers of lowering springs pay for advertisements, but lowering a car is not a good idea unless you are willing to do a lot of work. While the lowered center of gravity (CG) and the better aesthetics cannot be denied, many factors conspire to make a hastily lowered car handle worse than before it was lowered. The factors that immediately come to mind are: 1. Lowering changes your suspension geometry. This can affect jacking properties, amount of camber gain, and amount of roll. It is also possible for a roll center to pass below the ground plane (especially in front), which reverses one of the 3 means by which load transfers. This negatively influences feedback to the driver. 2. Lowering changes bumpsteer, which makes the steering wheel pull when subjected to bump or droop. This makes driving at the limit of traction more difficult. 3. Lowering can increase the angle of the CV joints, sapping power and increasing wear. 4. Lowering reduces bump travel. Unless you use specially shortened shocks or special shock mounts, many bumps (and corners) will compress the spring far enough that the bump stop comes into effect, which reduces mechanical grip. More on this later. 5. Lowering tends to make Miatas scrape on speed bumps. With all of this said, my car felt amazingly decent on soft lowering springs. This is a testament to how much bump stops affect handling. I tested to see how often my car hit the bump stops by placing a dab of white grease on the bottom of the bump stops, driving around, and checking to see if the grease was now on the top of the shock body too. Even in moderate turns my car was on the bump stops. At the time, about 25mm of shock shaft was visible between the bottom of the bump stop and the top of the shock body. You can get away with mild lowering by using stiff springs, but dropping a Miata more than ~25mm requires more work. The proper way to lower a car requires detailed analysis of the suspension geometry. From there, the suspension pick up points usually need to be relocated. Next, bumpsteer must be eliminated by altering the position of the steering rack relative to the suspension sub frame and/or substituting different tie rods. Finally, the alignment of the half shafts must be made parallel to the ground plane. In addition, provision for adequate shock travel must be made. This can add up to several kilobucks. *Proper Coil-Overs: *I later switched to a Ground Control adjustable coil-over setup and was shocked to find that my car rolled MORE in corners despite the ~3x increase in spring rate. This can be attributed to the new springs being stiff enough to keep the car off the bump stops, but less stiff than when the car was riding on the bump stops. The general ride became stiffer, but the big bumps are now soaked up a lot better. I appreciate the ride, but I suspect most would not, though I have never had a complaint from a passenger. My car is significantly stiffer than a Sport package Boxster or a C5 Z06. I am running the 8.0F/5.4Rkg/mm springs, which are one step more aggressive than the 6.7F/4.5Rkg/mm springs usually supplied with the GC kit. We will have to get used to using big boy units at some point, so why not start now? Anyway, one nice thing about the coil-overs is that you can buy a different set of 4 springs for ~$200 if you feel the need to play around with spring rates or lengths. Eibach and other brands have a huge selection of springs from close to stock rates to over ten times as stiff as stock. I have ~150mm tall springs up front and ~175mm springs out back, but some will choose a 200mm rear spring, which should allow slightly more wheel travel before coil bind at the expense of not being able to lower the car very much. I would have gone for this option if I had known about it when I bought the kit. As a point of reference, the Spec Miata racecars use 12.5F/5.8Rkg/mm springs while the FM Track Dog uses 11.6F/6.2Rkg/mm springs as far as I know. Both of the racecars mentioned above have full roll cages, so they can get away with a bit more spring stiffness. One of the limiting factors for spring stiffness is chassis torsional stiffness. The Miata has a stiffness on the order of 3000ftlbs/degree. After a certain point, using stiffer springs will flex the chassis too much and generally accepted chassis tuning rules may no longer apply. The suspension geometry of the Miata allows the use of softer springs than cars with McPherson struts. Cars with poor suspension geometry need relatively stiff springs to keep the wheels in favorable alignment. The Miata has a larger favorable range of travel, so it can be sprung more softly and further benefit from increased mechanical grip. Finally, springs are one of the most highly stressed bits on a car, so I am wary of the ricer kits that cost half what the real deal costs. *Dampers: *Dampers are commonly called shock absorbers, which is a misnomer. They do not absorb shocks. Energy is put into the dampers by suspension movements. Following the law of conservation of energy, this mechanical energy is converted into thermal energy. This heat radiates from the damper’s body. Note that a damper isn’t the same as a dampener. A garden hose is a dampener, and I have a real aversion to sounding dumb. Anyway, dampers come in two basic flavors, monotube (or gas) and double tube. There are good books on the subject, so I will stop at the most basic differences. Monotube dampers have pistons nearly the diameter of the damper body. Moving the large piston in a monotube damper displaces a good deal of oil, allowing the damper to damp very small bumps. Dual tube dampers have smaller pistons, which displace less oil when moved, and consequently are less effective on small bumps. When the travel of your damper is only ~20mm in each direction, having the first 5mm of travel basically without damping can be a problem. The monotube dampers have better heat dissipation characteristics than double tube units, but advances in anti-foaming oils have made this a moot point. Of the more affordable dampers (<$150 each), Bilsteins are monotube, while Konis, KYBs, and Tokicos feature double tube construction. All of these dampers are valved for near-stock spring rates, though the last 3 can be adjusted for more damping. Konis are thought to work well with springs up to around 8kg/mm. Konis can be rebuilt by Truechoice Motorsports, Pro Parts, and others. They can be revalved for stiffer damping, as well as upgraded to double adjustable. A lot of autocrossers use Konis since they can be rebuilt to suit racing conditions. Some high-end dampers, like Teins and JICs, come packaged with springs and valving to suit them. *Suspension Travel:* Accurate measurement of suspension travel has to be measured with expensive data acquisition systems. The low budget way of measuring suspension travel is to put zip-ties around the lower ends of your damper rods and measure how much they move up the rod as the damper compresses under certain conditions. A fair first order approximation of droop travel is the negative of the bump travel you measured. This approximation is only good for lateral forces, as different wheel rates front to rear complicate the matter in a longitudinal sense. Zip-tie testing is best done in a large, closed off parking lot. You have to be fairly scientific about this to get good data. If, for example, you hit a bump during testing, your measurement is of that bump and is not representative of suspension travel (due to roll, etc.) on a smooth surface. To measure body roll, you must slowly bring the car up to speed as you turn. Sudden acceleration, braking, or hitting a bump means you have to reset your zip-ties and try again. When measuring brake dive, you have to stop at the maximum rate, but you also have to build up to it slowly. Quickly slamming on the brakes will cause inaccurate results. These tests are mostly useful in calculating how much travel you are using and how much roll you have. The biggest problem with this method is resetting and measuring the movement of the zip-ties. I used an old bicycle spoke to reset the zip-ties. I measured the movement of the zip-ties by placing short sections of drinking straw on the end of the spoke, fishing it through the spring, and comparing it’s height to the displacement of the zip-tie. I would trim the straw and re-measure until it matched up perfectly, then measure displacement directly from the length of the straw. I would estimate this to be accurate to ~1mm or so. Another way of measuring roll is to set a camera up at the height of the horizontal styling line that goes around the car and measure roll from the resulting picture, as in the N. Garrett book. The picture should be taken when the car is at the limit and settled into the turn. If you throw the car into the turn your data will reflect it. The most telling data from the zip-tie method is that obtained by driving the car in the conditions you are tuning it for. If you take the measurements at an autocross, you can see if you are hitting the bump stops in the worst case or not. I found that my car as of mid-2004 used 19mmF/22mmR of damper travel in steady state turning on a smooth surface. This gives me a calculated roll angle of ~3.1 degrees, which is in the ballpark for a stiff Miata. Under braking, I recorded 24mm of damper travel up front. From drawings I made of the suspension, my front roll center should be very close to the ground plane with this amount of dive. In these smooth conditions my bump stops never come into effect. Driving around town for a week, I found damper travel to be about twice the amount that I recorded for smooth conditions. The bumpstops do not come into effect around town either. I could probably get away with the softer standard GC spring rates from a damper travel perspective, but I would risk sending my front roll center below the ground plane when entering turns under braking. The FM rear shock mounts on my car are a bit unnecessary since I am running the minimum stock height, but they do give me a bit more bump travel for really bumpy surfaces. The maximum amount of suspension travel at the wheel is equal to the maximum travel at the damper divided by the motion ratio. The motion ratios on the NA Miata are 0.692F/0.766R. Motion ratios have no unit, but you can think of it as (mm damper travel / mm wheel travel). The stock Showa dampers have ~118mm travel and my Konis have 125mm. Suspension travel for racecars is often split so that 1/3 is available for droop and 2/3 is available for bump. On the NA body style Miata this is not the case. The rear suspension really doesn’t have a lot of bump travel from the factory. This isn’t a real problem with the stock suspension, but can be a big problem on a lowered car, especially when the springs are soft. A lot of people try to solve this problem by cutting the rear bumpstops. I have never tried this approach, as it doesn’t make sense to me. A cut bump stop would allow the suspension to soak up small bumps better when the bump stop does not come into effect. On larger bumps, however, the suspension would quickly come into contact with the shorter, stiffer bump stop. This would dramatically increase the loading on the corner of the suspension in question and reduce traction at that wheel. Problems with lack of bump travel in the rear suspension usually show up as oversteer. There are 2 good solutions I can think of to fix the lack of rear suspension bump travel other than raising the car to near stock height. The first is to install the FM upper mounts. Alternately, you could possibly order some dampers with shortened bodies. Whatever you do, do not drive your car without bumpstops. There are 2 products on the market that are often mistaken as adding additional shock travel. The first is a set of upper shock mounts from Japan that apparently lower the car a bit, yet add no bump travel. The second is a set of spacers that fit between the stock upper shock mount and the chassis. These are intended to help people with tire rub and will not increase bump travel. These products have their place though, and that is to change ride height without affecting the distribution of travel between bump and droop. If your lower shock bushings are removable (Konis, etc?) you may be able to make some custom offset bushings and gain ~5mm of bump travel, but this would be breaking new ground. To get your gain in bump travel you would have to raise the coilovers by whatever amount of offset you used. Another method of gaining rear bump travel is to update to the upper shock mounts of the second generation Miata. You have to make some provision for change in the angle of the damper shaft or your dampers will fail. I have not played with this, so it is best to search Miata.net if you intend to perform this mod. Note that stiff coilovers setups may not be able to use the full range of travel since the springs offer a limited amount of compression before they go into coil bind. My current springs have ~90mm of travel at the front and ~105mm at the rear. This should not be a problem if their range of motion is within the range of damper travel. The reduction in total travel warrants making sure your bump stops actually come into effect. You may find that your coils will bind solid before the damper body contacts the bump stop. Split nylon spacers that fit over the damper shaft are available to adjust the point at which the bump stops come into play. *Antiroll Bars:* Antiroll Bars (ARB’s) are used to limit body roll and modify the balance of traction between understeer and oversteer. ARB’s function by transferring weight from the inside wheel to the outside wheel. This transfer of weight acts to suppress the natural tendency of the vehicle to roll towards the outside of the turn. Reducing roll helps keep the tires more perpendicular to the pavement, which increases traction. Unfortunately, the grip tires produce is reduced as more weight is applied. The load sensitivity of tires usually predominates over the gains from being more perpendicular to the road. As a point of interest, Racing Beat claims that the RX-8 is an exception to this rule. They claim to actually increase grip as the result of adding their thicker ARB’s, though this may be due to a better balance of traction front to rear. Most cars actually lose maximum grip potential when larger ARB’s are installed. In most cases, installing a pair of stiffer ARB’s will improve your slalom speed and decrease your maximum grip. The benefit comes in quicker transitional response and usually in driver confidence as well. The roll angle can also be controlled by increasing the spring rate, which does not transfer any weight. The downside to increased spring rate as a means of roll control is that overly stiff springs can reduce mechanical grip. In addition, there is a practical limit to how stiff your springs are based on how much of a beating you are willing to endure as you drive the car. Overall, using ARB’s to control roll is more effective. ARB’s can also be used to change the response of a car between oversteer and understeer. The generally accepted rule is that stiffening the ARB at one end will cause that end to slide more easily. This is a little counter-intuitive, since we usually try to fix the end that is not working as opposed to decreasing the performance of the end that is working. Note that the above rule is not true in every situation, especially if something strange is going on in your suspension. ARB rates can be compared numerically to determine the ratio of stiffness front to rear. The stiffness of an ARB is a function of several dimensions of the bar itself. The only dimension that changes much between different bars for the Miata is the outer diameter of the bar. For our purposes, we can assume that stiffness is equal to the diameter of the bar raised to the 4^th power. We can then compare different combinations of ARB’s by dividing the stiffness of the front bar by that of the rear (front O.D.^4 / rear O.D. ^4 ). A relative stiffness of ~8:1 is close to neutral. Higher relative stiffness should yield understeer, while a lower figure leads to oversteer. Some of the sets of bars out there are not very well matched. One set I tried was 3.5:1 and it made the car oversteer horribly. My current set is ~7:1. Tubular ARB’s have a weight advantage over solid bars. Some tubular bars are over 3 times as stiff as stock, yet they weigh less than stock. Try to twist an empty beer can and you will get an idea of the torsional stiffness tubes can achieve at really low weights. To calculate the stiffness (as we use it above) of a tubular bar, subtract the inner diameter raised to the 4^th power from the outer diameter raised to the 4^th power (O.D.^4 – I.D.^4 ). Only front bars are available in tubular construction. *End-links: * The stock end-links are non-adjustable units that use rubber bushings to prevent binding. The bushings are liable to flex somewhat, causing yet another source of non-linearity to confuse the driver. My stock end-links would not allow the use of all of the adjustment holes in the rear bar. Adjustable links with stiffer urethane bushings are available from Racing Beat. Better yet, Heim jointed adjustable links can be purchased from Moss Motors and others. If you are comfortable with your homegrown engineering skills, you could make a set of adjustable Heim jointed end-links for about half the cost of commercial units. Heim joints can be purchased for $5-10 from Truechoice, Pegasus, and others. You will need a threaded section to fit between the joints. You may be able to use threaded studs, but do not be tempted to cut your own with a die. Reading Carroll Smith’s book on fasteners is a really good idea before attempting anything like this. You would also be smart to use ‘safety washers’ at the single-shear end just in case the Hein joint separates from it’s housing. In the event of separation, these washers prevent the link from coming completely off the end of the ARB. 28g of prevention are worth 454g of cure. *Installation notes for ARB’s: * The front ARB mounts are known to tear out when subjected to stiff bars and spirited driving. The load on these mounts is related to how much the bars are twisting and how much the bars resist twisting. A heavy bar used in conjunction with the stock springs will put a lot of load onto the mounts. A car with soft lowering springs and whose roll is determined largely by the bump stops will then place less load on the mounts. A reinforcement kit is available from Racing Beat. The kit can be ordered with or without a cross member. I cannot visualize what the cross member actually does. Several other vendors offer stronger mounts, including Good-Win Racing and Mazda Motorsports. The rear bars have relatively short torque arms. This may introduce some geometric problems, so I set the rear bar at full soft to keep the torque arms as long as possible. I then balance the car with the front bar. The rear bar also has a habit of sliding sideways in its mounts while turning. Some bars have stops to prevent this from happening, but I have found they end up sliding too. I made some stops from split heater hose and worm gear clamps that really work. I once found evidence of contact between the front bar and an arm in the front suspension. It is worth making sure this doesn’t happen in your suspension. The actual rate of the ARB’s will depend on how much its mountings flex. The rubber bushings in which the stock bars are mounted can be replaced with polyurethane for increased stiffness. I have a chunk of Delrin on my desk that may become a set of ARB bushings at some point. I am really tempted to try using stock bars with Heim jointed end-links and stiffer bushings just to see how much stiffer than stock it is. Several Miata.net posters have found that the polyurethane bushings provided with aftermarket ARB’s are too long, causing them to bind up when the mounting bolts are tightened up. This can be fixed by trimming the end of the bushing flush with the face of the mounting bracket. *Bushings:* The stock bushings consist of a metal tube with an outer diameter of ~20mm bonded inside a rubber tube with an outer diameter of ~40mm. The metal tube contacts the suspension subframe (or rear upright) on its ends, clamped by a bolt running through the center of the tube. The rubber tube makes contact with the control arm along its circumference. Relative movement is accomplished by twisting the rubber tube. There are no surfaces sliding on one another, hence no need for lubrication. The stock rubber bushings always attempt to return to an untwisted state. The untwisted state is determined by the position of the control arms when the bolts running through the bushings are tightened. To prevent excessive wear, these bolts are always tightened when the car is on the ground. An interesting consequence of using rubber bushings is that their contribution to the spring rate is dependant on displacement of the control arm. The farther the control arm is twisted from its equilibrium position, the greater force is required to maintain the new state. When the suspension is compressed, the rubber bushing adds to the spring rate. When the suspension is unloaded, the bushings subtract from the spring rate. For example, when the car is in a turn, the spring rate of the outside wheels increases due to the effect of the bushings, while the spring rate of the inside wheel is reduced. I also noted that when the suspension is manually displaced and allowed to spring back to the equilibrium position, about 2 oscillations occurred before movement stopped. This indicates that the stock rubber bushings act as an underdamped system. The spring rate contribution of the stock bushings was found by first placing the car on jack stands, and removing the wheels, spring/damper units, and ARB’s. A weight of 13kg was then hung from a wheel stud and the displacement of the suspension at the hub was recorded. The front suspension was displaced by 23mm, while the rear was displaced by 15mm. Assuming that the spring rate of a twisting bushing is constant, I calculated spring rates of 0.57kg/mm F and 0.87kg/mm R. From the experience of changing the dampers and the reproducibility of the displacement figures, I feel that these numbers are at least in the ballpark. Aftermarket polyurethane bushings operate in a different regime. The inner steel tube is virtually the same as in the stock rubber bushings. The difference is that the polyurethane is harder than the rubber and that the polyurethane is not bonded to the steel tube. There is a sliding motion between the 2 parts, and this necessitates the use of grease. The poly bushings have no tendency to return to an equilibrium position since the polyurethane slides instead of twisting. At least on initial instillation, I noticed that these bushings have a firm resistance to movement and this resistance is not dependant on displacement of the suspension. This situation is very similar to the ‘stiction’ inherent to low quality telescopic suspension forks found on mountain bikes. I am hoping these bushings wear in a bit, reducing this effect. Since poly bushings rotate instead of twisting, they do not add to the spring rate. They should have a damping effect due to the shearing action of the grease and the stiction. In theory, switching from stock rubber bushings to poly bushings should decrease spring rate (in bump at least) and increase the relative damping level. In practice, the switch makes the car feel a little bit harsher. The real reason we change bushings is to limit compliance in the suspension. Back in the dark ages and way before I was squeezed out, cars had metallic bushings. These were troublesome devices that required occasional lubing to prevent creaking. A ‘lube job’ was an automotive service, not a mixed drink or anything else. The metallic bushings also transmitted a lot of vibration to the chassis and offended delicate American rumps. On the up side, they kept the suspension aligned, even when subjected to cornering forces. To combat the atrocity of a little NVH, rubber bushings were invented. These softened up the ride, but introduced compliance whenever the suspension was stressed. This compliance tends to go in the wrong direction. For example, the tires of a rubber bushed car in a turn will both change camber in the wrong direction. Deceleration is accompanied by the tires toeing out at all 4 corners. Acceleration causes toeing in, especially on the driven wheels. We tend to try to accommodate all this flexing by adding a little more negative camber all around and some toe in at least in the rear. These are band-aid solutions. Ideally, we would do something to stop the compliance instead of trying to negate its downsides. Stiffer bushings are a step in the right direction, since they permit less compliance than the stock pieces. A step farther is to move to hard plastic Delrin bushings, which almost eliminate compliance. These also transmit a lot of NVH, and the general consensus is that they are too stiff for street driving. There can also be problems with the steel sleeves in these bushings rusting, which causes wear and noise. This could probably be cured by using stainless steel sleeves. Most purpose-built racecars use heim joints or spherical bearings, though these wear out and can also have a lot of stiction. The most interesting method of allowing movement without causing stiction is using flexures. These are blade shaped sections of carbon fiber which can flex in one direction, but are relatively stiff in the other. To visualize this, try bending an old hack saw blade. It bends easily (and without stiction) in the plane of the blade, but is very hard to bend in a direction 90^0 to that plane. This is F1 technology though, and I have never seen it used elsewhere. It is worth noting that the ball joints used in the front suspension are nearly perfect. They have no compliance, move freely, and last a long time as long if you keep good boots on them. *Caster:* Ground Control has a really good tech article about caster that got me thinking about caster and its relation to the Miata. The zero order theory of caster says that, since caster cambers BOTH front wheels in a favorable direction in a turn, we always want to use the maximum amount of caster available. A better understanding of caster can be had by looking at its’ 2 principle effects and how they relate to the entire vehicle dynamics picture. The first effect-- that of cambering both front wheels in a favorable direction while turning— increases grip at the front of the car. This increase in grip is dependent on the steering angle, so the more steering you add, the better the front grip. Unfortunately, there is no easy way to compensate for this at the rear of the car. If you balance the car at one steering angle, that balance will be lost as soon as you have to use another steering angle. In effect, this first property of caster causes oversteer with increased steering angle. The second effect is that weight is increased across the inside front-outside rear diagonal pair of tires. This effect also increases directly with steering angle. Spring rate determines how much weight is transferred due to the caster angle, so stiffly sprung cars notice this effect more than soft ones. This transfer of weight across a diagonal set of tires is called ‘wedge’ by the left-only crowd. I think pushrods belong in suspensions and not in engines, so I just remember that this translates to oversteer. This oversteer also increases as more steering angle is applied. In tight turns, the effect of positive caster leads to oversteer. At the same time, we are liable to be in a low gear and have a lot of thrust available. This may result in an aggravating inability to put power down. So why do we want any caster? Mazda designed our beloved cars to be sold to a wide customer base. The vast majority of that base likes centerpoint steering, which reduces kickback at the steering wheel on rough roads. To arrange centerpoint steering, Mazda had to use a high kingpin inclination angle of 11^o 20’. Kingpin inclination, also known as steering axis inclination, is the angle between the steering axis and the vertical axis when viewed from the front (caster is viewed from the side, so kingpin inclination is measured on an axis 90^o from that of caster). The kingpin inclination works opposite of the caster angle in that it cambers both wheels in the wrong direction. Consequently, we can use some caster to counter the negative effect of the high kingpin inclination angle. I am still playing around with caster angle on my car. I can’t yet offer a direct comparison since I changed a lot of factors between my last 2 alignments. I moved from ~5.5^o down to 3.5^o . Other modifications have changed the balance of oversteer and understeer. The lower setting has reduced steering effort on my de-powered rack, but not too much. The car still centers itself well coming out of corners. If I had the time and money, I would do back-to-back testing of caster angles. My conclusions so far are: * Too much caster should result in oversteer, especially on stiffly sprung cars. * Too little caster should result in understeer due to the effect of KPI. * Too much caster makes steering effort heavy on de-powered racks. * Too little caster may feel to vague and may not self-center well. * We use more steering angle than road racers, and should therefore pay caster more attention. Another possible conclusion is that a lot of us use too much caster and try to balance the car with light rear bars. The following bits are still under construction: *Bumpstops: *buy nice ones and stay off them! *Ackerman:* we have it. It sucks. I see no clear solution. Stiff front bars could reduce the inherent drag, but not increase front grip. Toe out adds to Ackerman, but maybe not to the extent that running a lot of toe in will increase grip that much.