The Integra
and Civic chassis that Honda produced from the late eighties to the early 2000s
were set apart from average FWD cars because they could handle well out of the
box. With modification, they could handle even better.
One common
problem with FWD cars is understeer. Because the front wheels are doing
the accelerating, turning, and the majority of stopping, the tires tend to
"fall off" sooner than their RWD counterparts. To combat this,
FWD race car builders increase the chassis frequency, or "stiffness,"
of the rear of the car using sway bars or stiffer springs.
In this
post, readers will see how EJ2 Track Rat was able to create a low-budget
solution for drivers looking to increase rear roll stiffness via use of a stiffer
sway bar in the rear of the car.
Background
Alex's
TrashTeg was driving fairly well. He was able to pass competition school
with NASA Northeast in it and he learned a fair amount about racecraft during
the course of the season. He was also having a ton of fun. However,
Alex was beginning to realize the limits of his budget-built car-- Partially
because his car had too much dive, squat, and roll under braking, acceleration,
and cornering, respectively.
When
TrashTeg was first built, it had the following setup:
205 Toyo RRs
At a ride height of ~5.5-6 inches
With a front spring rate of 650 pounds/inch,
Rear spring rate of 750 pounds/inch,
No front sway bar,
A 25 mm GodSpeed (eBay) rear sway bar,
And some Koni Sports he got for free from our friend
Anthony
Our goal was
to increase the roll stiffness of the car while still keeping the car
controllable at the limits with a healthy amount of rotation in corners.
We had two options: increase front and rear spring rate, with a stiffer rear
spring rate than front, or increase front spring rate and throw a big rear bar
on.
Since I only had a pair of 800 pound/inch springs in my garage, we figured we'd
throw the 800 pound/inch springs up front, and increase rear roll
stiffness without springs. The easiest way to do this is with a bigger sway bar, which
A-Spec Racing (ASR) makes. Their 32 mm hollow bars come in different
thicknesses and since they're hollow, they have the advantage of less weight
vs. their solid counterparts for the same amount of torsional resistance.
ASR Hollow Rear Sway Bar Kit
However, ASR
bars are not cheap at ~$500 for a full kit and Alex had already purchased a
$300 eBay sway bar kit. He wasn't about to spend $500 more for the ASR
product as a result. But, he and I were both interested in doing
something creative.
Much debate
was had on this thread-- specifically whether one can weld to a sway bar and
expect it to work because sway bars go through a forging process that allows
them to become "spring steel." Spring steel is highly
malleable, but returns to the shape it was bent from fairly easily.
Welding to spring steel changes that property and increases brittleness.
Example of a Cracked Sway Bar (Not Ours)
Doing the Deed
Since we don't care about armchair engineers and we buy parts for our
destined-for-doom race cars on eBay, we decided it'd be worth a go to attempt
something similar. The new setup would have the following changes:
Ride height lowered to 5" at pinch welds
Front spring increased to 800 pounds per inch
A "custom" xxx mm solid rear sway bar, courtesy of
the plumbing aisle at Home Depot
We started by purchasing some steel black tubing with an inner diameter of 1" and a wall thickness of 0.133" for about $20. A 1" inner diameter equates to 25.4 mm which is only slightly larger than the OD of our eBay sway bar. Our plan was to cut this bar to length so that it would fit between the pillowblocks that mount the sway bar to the subframe. Then, we would cut the bar in half, length-wise, and weld half of the Home Depot pipe to the sway bar. This would result in a solid sway bar with a diameter of ~28 mm.
Sway Bar Diagram from ASR
Our Plan from EJ2 Track Rat
We laid out the eBay sway bar after purchasing the pipe from Home Depot and we started by measuring where our addition would be and how short we'd need it cut.
Laying Out the Sway Bar
Next, we cut our Home Depot pipe to length using a cut-off wheel. Ideally this would be done with a band saw, but with a pipe-vise it's easy to do with a $30 Harbor Freight cut-off wheel. The cut just isn't as clean.
Alex Cutting Home Depot Pipe to Length
Next we cut the pipe in half, lengthwise using our cut-off wheel and our old school bench vise from eBay.
Cutting the Home Depot Pipe in Half Length-Wise
To weld our cut-off half of the Home Depot pipe to our eBay sway bar, we would need to first clamp the OD of the Home Depot bar to the sway bar. After grinding the surface of the Home Depot pipe and the sway bar clean, we clamped the Home Depot bar to the sway bar using a set of Milwaukee vise grips.
Half of a Home Depot Pipe Clamped to an eBay Sway Bar
I ran a bead across a section to the left and right of the vise grip, then moved the vise grip down the length of the sway bar and repeated the process...the entire thing was done with flux wire.
Close-up of Beads Ran Next to Vise Grip
The entire bar welded up looked amusing, to say the least. It really was also great practice for welding. I needed some, so this was of benefit to me.
Picture 1 of Welded Sway
And Picture 2...
Finally, we sprayed the bar gloss black and mounted it onto the car. Once on the car, we checked to ensure clearance between our modified bar and the subframe brace was satisfactory.
The Bar Mounted to the Subframe
Checking Fit with Bar at Full Droop
Checking Fit with Bar at Full Compression
And after installing the rear bar, we installed my front springs and manually set ride height as we always do-- the LoBuk way.
Alex Setting Ride Height to Clock Bushings
Another View of Alex Setting Ride Height to Clock Bushings
Impressions
Alex tested TrashTeg at Watkins Glen in October of 2019 with NASA NE to see how the changes improved the car.
Alex Suiting up for a Race at The Glen
On Track, he noticed sharper turn-in due to the stiffer front springs, but the larger rear sway also allowed the car to have less mid-corner push. The new sway bar we created held up for a full weekend of abuse at one of the best and most demanding tracks in the Northeast.
After our racing shenanigans on Saturday we attended a NASA Northeast BBQ and had a good time. The car held up, the mods made it faster, and we had a great time.
In my last post, I discussed the first upgrade for my budget transmission-- moving from a helical limited slip differential to a clutch-type differential. If you didn't read that yet, I suggest you go back and take a look.
A pic from my Instagram of Damien and I at New Jersey Motorsports Park (Photo by Windshadow)
The second of two transmission modifications allowed in the National Auto Sport Association's (NASA) Honda Challenge H4 class is a final drive gear with a different gear ratio from stock.
So, what does a final drive do? In this post, we'll delve a little into that and we'll discuss how you should choose one for the road course.
What is a Final Drive?
The final drive is basically the last gear between the transmission and the drive axles. By altering the final drive gear ratio, you alter the amount of wheel rotations per rotation of the selected gear.
So, a 4.2:1 final drive will rotate the driven wheels 4.2 times for each rotation from the transmission and a 4.7:1 final drive will rotate the driven wheels 4.7 times for each rotation from the transmission.
If we select a "shorter" final drive, we are saying we want a final drive with more rotations at the wheels per rotation at the gear. This gives us more torque at the wheels and more acceleration, which we gain in return for a loss in top speed. We call this a "shorter" final drive because the amount of time between gear shifts is now not as long as it once was.
2 Fast...
On the flip side, if we choose a "longer" final drive, we are trading straight line torque and acceleration for top speed and a potential gain in fuel economy. The longer final drive gives less rotations at the wheel per rotation in the selected gear. It takes more time between shifts, so we call it "long."
Why Should I Change My Final Drive?
In a drag racing car, we can go faster by turbocharging our cars or modifying the engines to generate more torque and horsepower.
However, if the stock final drive is too short, one disadvantage is a loss of top speed.
If the stock final drive is too long, one disadvantage is a loss of torque at the wheels.
So, drag racers must find the proper balance between torque to the wheels and top speed by selecting the correct final drive. Many other factors play a role in this phenomenon, but they're beyond the scope of this article.
A Camaro at the strip
In road racing cars, the sessions are typically longer, the tracks have corners and elevation changes, and horsepower isn't as important as cornering speed in most amateur classes.
So, to go faster we can't rely on adding power at the cost of reliability, choose a final drive that will give us the best straight line speed, or choose the final drive with the best acceleration. We must consider other factors.
Ken, Kallie, and Brian at New Jersey (Photo by Windshadow)
How to Choose a Final Drive
There are five steps outlined in this article for choosing a final drive.
Step 1: Find the Powerband
Step 2: Find your Corner Speeds
Step 3: Compare Final Drives for Each Corner for Each Track
Step 4: Repeat Step 3
Step 5: Test
Step 1: Find the Powerband
Let's take a look at the stock, 94-01 Acura Integra LS/RS/GS-- a popular Honda for road racing due to its suspension geometry and operating costs.
The stock Integra that loaned its motor to my race car
The powerband for a car is a colloquial term which refers to the engine's RPM range at which it operates most efficiently. If you're low on money and have a relatively stock motor and no dyno sheet, a great way to find your car's powerband is through the internet.
Wikipedia states that the stock 1998 Acura Integra generates peak torque at 5,200 RPM and peak power at 6,300 RPM. So, we will conservatively assume that the powerband is from ~5,000 to ~6,500 RPM, allowing for shifting inaccuracy and any changes in engine efficiency due to age and use.
So, we know where the car's RPMs need to be to remain happy, but how do we apply this to the track? You may be racing Summit Motorsports Park Main with a 0.55-mile-long straight or Lime Rock Park which is a Miata track. You may be racing somewhere with anywhere from ten to twenty-five corners, all with different average speeds and elevations.
Some corners are extremely important and can lose races if the car and driver are not fast through them and out of them-- like Turn 1 at Watkins Glen. Some corners are not important at all and are referred to as "throw-away corners,"
Disclaimer: This is where we must analyze our cornering speeds-- it is an art and our conclusions may not always be the same. However, i'm going to do my best to break it down so you can make your own decisions in the end. The most important thing is that you trust your gut and supplement that with real-world data.
Using an AiM Solo or an equivalent lap timer, take a look at all of the tracks you race in a typical season in your area. For each track there will be several key corners where exit speed is critical because momentum is lost. For example, below you'll find a trace of one of my fast laps at NJMP Thunderbolt last year with a 4.2 final drive:
Data from an AiM Solo
There are three graphs drawn versus track distance in the above figure. In order from top to bottom they are: lateral G's, longitudinal G's, and speed. We use the longitudinal G and lateral G graph as a reference point and the speed trace to see what our corner speeds are.
In this figure, I have highlighted several corners where a lot of momentum is lost due to braking which is required to make the tight-radius corners. Those corners are 1, 2, 5, 7, and the exit of the octopus. Below you'll find a track map and video for reference:
NJMP Track Map
A lap of Thunderbolt in my old single cam HPDE setup
Looking at the AiM data, we can see the following corner exit speeds for a 4.2:1 stock final drive in an H4 Acura Integra:
Approximate minimum speed per corner at NJMP Thunderbolt
Watching video and using common sense, I can look at these speeds and remember what gear I'm in while on track, which is third gear for every. single. corner! Why is this useful again?
Step 3: Compare Final Drives
The internet doesn't just have the powerband for a stock B18B1, it also has calculators that allow you to look at the RPM vs Speed trace for a stock Integra transmission. The calculator I like to use for Hondas is from Zeal Autoworks. It's been around since forums were hot and it's still in use-- so you know it's good.
You can choose the Honda transmission you're interested in and alter gear ratios as well, comparing two at a time. Tools for other cars are likely to exist on the internet as well.
The screenshot above shows the max MPH for each gear using a stock 4.266 final drive in an Acura Integra. Using this calculator, we can also get the values for a transmission with a different final drive. For our example, we'll look at a comparison between a stock 4.2:1 and an aftermarket 4.7:1 final drive. Trans 1 uses a 4.2:1 final drive and Trans 2 uses a 4.7:1 final drive:
We can see that the RPM increases much quicker for a given change in speed with the 4.7:1 when compared to the 4.2:1. However, you'll also note that the top speed in 5th gear is much lower with a 4.7:1 final drive than it is with a 4.2:1. This reinforces what we stated above, that a longer final drive has less acceleration ability, but a higher top speed, and vise-versa.
Now, we know the corners we want to optimize our car's final drive for, we know the minimum speed through the corners, and we know the power band for our car. All that's left is to compare graphs of RPM vs. speed for different final drive ratios to find the most optimal for your car. For the sake of simplicity, we'll analyze the comparison between a 4.2:1 and 4.7:1 transmission further.
Dame and I going at it at Lime Rock Park
Below you'll see the graph we looked at previously, but this time I've overlaid two y-axis lines and several x-axis lines. The y-axis lines are in red and they represent the powerband and the x-axis lines are the minimum speeds for the corners we deemed most important at NJMP Thunderbolt.
Overlaying pertinent powerband and track corner speed data on the RPM vs speed trace for different transmission gearing
Corner by corner, let's look at what the data tells us...
Step 3a: Compare Final Drives-- Corner by Corner Analysis
In this next sub-step, we will go through each of the corners we deemed most important and select the final drive ratio that benefits us most.
Turns 1 and 2:
Below is a graph of RPM vs speed for a transmission with a 4.2 final drive and a 4.7 final drive. We can see that when proceeding through Turns 1 and 2 in a 4.7:1 final drive transmission, we have two options for gear choice:
We can stay in fourth gear through turn 1 and keep our feet to the floor as we progress through turn 2, minimizing brake input. Or, we can keep the car in 3rd, turn into corner #1, and shift into 4th out of turn 2. Either way, we need to row through the gears to optimize mid-corner to corner exit speed with the 4.7:1 final drive.
Graph of RPM vs. speed for a transmission with a 4.2:1 final and a 4.7:1 final
Yet if you look at the green line in the above graph, you'll see that through turn 1, you can maintain 3rd gear, through turn 2 you can maintain 3rd gear, and once through turn 2, onto the back straight, you can shift into 4th all while staying in the powerband.
This implies that a 4.2:1 final drive is most advantageous for turns 1 and 2.
Turn 5:
The next corner we'll analyze is the slow left-hander out of the high-speed, sweeping right-hander at NJMP Thundebrolt-- also known as Turn 5.
The graph shown below is similar to those above, but the x-axis line in this graph is for Turn 5.
Turn 5 min speed, stock Integra powerband RPM range, and RPM vs speed plots on one graph
Driving a 4.2:1 transmission, we have two options for gear choice-- 2nd or 3rd. In either gear, we're significantly far from the powerband while close to the minimum speed for this corner. 4.2:1 is far from optimal for this corner.
However, a 4.7:1 final drive puts us at the intersection of the powerband for our motor and the minimum speed for Turn 5 when in 3rd gear.
From this, we can conclude a 4.7:1 final drive is better for Turn 5.
Turn 7:
The second-to-last corner we'll analyze is Turn 7 at NJMP Thunderbolt. This corner is a lot of fun because it requires a lot of braking, but if you choose the proper line and throttle/brake inputs through here, passing can be completed against a fierce competitor.
Looking at the graph below, it appears that a 4.2:1 final drive would be the best from the middle to the end of turn 7 as it requires less shifting and starts at the bottom of the powerband.
Keep in mind, however, that exiting turn 7 is not a straight, but a decreasing radius right-hander that requires grip. Not all speed through this section is due to final drive selection. It could be due to other factors not covered in the scope of this article, such as driving style or differential setup.
With situations like this, it's best to rely on your experience driving the track. Feel the corner and use your intuition to gauge what would be most advantageous in the big picture. I may think this is an important corner for final drive selection, but maybe it's not! Remember what I said in the beginning disclaimer-- this is an art.
Take a look at the graph and map below:
Graph of turn 7's min speed on an RPM vs speed graph for a stock Integra with different final drives
Diagram of T7 at Thunderbolt
The Exit of the Octopus
If we zoom out on Turn 7 and The Octopus together we can better put things into perspective.
The "back half" of NJMP Thunderbolt
From the exit of The Octopus, it's easy to see that the 4.7 final drive is better positioned to exit the corner than the 4.2 final drive is. The 4.7 AND the 4.2 final drive are both in 3rd gear which means they still have a 4th gear to shift through before they reach the abysmal, long fifth gear in these cars.
Engine speed vs car speed chart for the exit of The Octopus
However, a car must be tuned within the context of the track and therefore it is an art as mentioned above. Keep in mind that after the exit of The Octopus, racers in a low-powered car will never lift until they get all the way back to Turn 1 at NJMP Thunderbolt.
This reinforces the point that tuning a car cannot focus on only one aspect-- road racing or for the drag strip...
Using different ratios, now it's time to experiment with which one works best for your driving style for a given track. Additionally, remember to consider corner speeds for other tracks you regularly drive. Lime Rock Park and Watkins Glen may deserve two totally different final drives for example.
Step 5: Test
Lastly, get your car on the track and see if it works. Compare data using your on-track lap timer. This is an essential tool.
Conclusion
As stated above, a final drive can have a large influence in corner exit speed for a track car. While it also factors into the top speed of a track car, there are other factors to consider such as gear ratios for each individual gear and overall horsepower/torque for the car. Additionally, there are factors that will make the car corner faster, thus potentially affecting final drive choice.
It was some time ago, in December of 2018, where I was in the middle of "off-season improvements" that I decided to build my own transmission.
Welding exhaust flanges in the winter
However, well-built, competitive Honda Challenge transmission builds can get expensive. For example:
*Aftermarket final drive = ~$600
*MFactory metal plate differential = ~$900
*94-01 Integra "Brass Master" transmission rebuild kit from Syncrotech = ~$650
*New transmission gearset and case = ~$150-300 used (depends on seller/condition)
The above combination of parts doesn't consider the use of a better differential either, which can run upwards of $1500 (OS Giken).
OS Giken assembly
So, my goal was to assemble a competitive transmission setup for Honda Challenge H4 using as little resources as possible. Similar to a lot of other parts on my car, I planned to hunt Facebook Marketplace, forums, eBay, and Craigslist for sales or deals on used parts for the build.
Custom eBay cold air intake, MSD coil bracket made from an oil pan, and an entire MSD setup I got for ~half-of-retail price
The first part of this series will cover selecting the differential.
What Does a Differential Do?
Whether in a front-wheel, rear-wheel, or all-wheel drive car, the differential takes power from the transmission's final drive and distributes it to the left and right axles.
The simplest and most common differential is termed an "open" differential by car enthusiasts.
I'm not going to delve into specifics about this differential type. The most important thing to remember about an open differential is that by design, the wheel with the least grip spins the fastest.
This means, when going around a corner, you'll notice the car has a tendency to spin the inside wheel while the outside wheel isn't getting power. Not only is this slow, but it's also lame. In drag racing, this tends to result in one-wheel burnouts.
For specifics on an open differential, check out the video below:
LearnEngineering's video on open differential function
The next type of differential is called a "limited slip" differential (because it limits the slip of the wheel with the least grip).
There are two common types of limited slip differentials-- a torsen or gear-type differential and a metal plate or clutch-type differential.
A torsen differential uses gears to limit slip and transmit power more evenly between the two axles. Since I don't want to delve into specifics, I'll explain the pros and cons of a torsen:
Torsen pros:
*low-maintenance-- change fluid once a season for a road racing car
*easy to drive at low speed-- good for street driving
*typically cheaper than clutch-types
Torsen cons:
*Corner exit speed is not as high relative to a clutch type
Cannot adjust lockup percentage like in a clutch type
Cannot adjust lockup on decel or accel like in a clutch type
A torsen differential's function is described in the video below:
LearnEngineering's video on torsen differential function
A clutch differential uses clutch packs to transmit power from the transmission's final drive gear to both axles in an even manner which "limits slip." A description of their function is shown below:
Torsen diffs are durable, relatively cheap, relatively fast, and great for daily driving since they don't provide a lot of low speed steering resistance. So why would we want to run a clutch type diff? Let's review the pros and cons below:
Clutch-type cons:
*Expensive fluid with friction modifier generally needed
*Requires fluid change after every weekend to avoid excess wear
*Can fail sooner than torsens if maintenance not kept up
*Expensive to purchase
Clutch-type pros:
*Generally faster than torsens in a FWD car mid-corner to exit if set up correctly
*Adjustable for lockup on decel and accel
*Amount of lockup is adjustable
Seems like there are a lot of cons to running a clutch-type. So why would we run it? Well-- real-world experience and data suggest there's a benefit in lap time from the lowest Honda Challenge classes to the highest.
For example, exiting corners, I sometimes have to wind my wheel out sooner and take a wider line to avoid cooking tires while competitors with clutch-type diffs can stay tighter, taking up less distance and thus less time. Data reinforces this conclusion from the "butt dyno."
Watch corner exit speed in the in-car vs the silver hatch
Another example was when I drove the white DA Integra on the left and laid down a 0.7-second-faster lap time than in my car on my fourth lap ever in the DA. Mid-corner and corner-exit grip were notably better due to the diff.
NOTE: There were other variables in that case. The car's final drive helped with corner exit speed at the track we were at. The DA also had better, double-adjustable Konis in it than I had which made curbing a breeze to drive over. And lastly the Integra was more stiffly sprung in the rear, which helped it rotate.
Brian's DA Integra and Damien's EK Hatch with an E30 in tow
What's the Diff?
I eventually found a Spoon Sports 1.5-way B16/B18B1 diff for ~$350 used. Not knowing much about diffs and willing to take a risk, I drove down to Delaware, very close to where I lived at one time in my life, and I picked up this diff without asking many questions.
Diff with B18B1/B16 ring gear diameter
These diffs typically cost more than $1100 new on the internet. I considered myself lucky at the time.
Making it all Work
The first thing I noticed when I brought it home was that the ring gear where the vehicle speed sensor (VSS) goes was chewed up. This was concerning:
This was likely caused by the previous owner not using the correct VSS gear. Typically, when you buy a new diff, the diameter of the case is larger, so the manufacturer will include a different VSS gear to match up with the diff. This is most common with:
*OS Giken
*Cusco
*Kaaz/Spoon
So, I began looking for VSS gears for a Spoon differential. This part is hard to find as people don't typically sell it used and you usually have to buy it straight from the manufacturer. However, since Kaaz makes the Spoon differential for the B Series tarnsmission, I was able order a new one for $150 shipped through their US sales department.
Next, I opened up the differential and looked at the internals to see how bad the clutch packs were and if any of the gears were damaged. Kaaz has a handy video that shows you how to disassemble their differentials shown below:
Differential assembly and disassembly
The plates had some of their clutch material worn off but none of them were missing teeth or extremely burned up. Pictures of the plates are below:
All twelve clutch plates from the Spoon/Kaaz differential
The belleville washers were in rough shape. Lots of scoring around the inner and outer diameter of the washer with a slight burr formed on the inner diameter of one of the belleville washers. Additionally, the thrust washers were of different IDs and ODs and thicknesses. In the pictures below, Thrust Washer B was thicker than A which suggests the previous owner tried altering initial lockup of the diff.
When I sent pictures of just these four pieces (NOT even the plates), Kaaz recommended a full rebuild kit-- about $250 dollars (wasn't gonna happen). I requested the individual parts to no avail. Kaaz would only sell all or nothing:
Thrust washers, belleville springs, side-by-side
To see how messed up my diff actually was, since Kaaz stated they "don't recommend patches," I began doing some research.
I spent a long time trolling the internet looking at the condition of others' diff plates, experiences they had with their Kaaz diffs, and reassembly and maintenance recommendations. I read probably somewhere between 20 and 30 Honda-Tech.com threads, watched a bunch of YouTube videos, and stumbled across some less-known, but extremely helpful resources such as a blog called Pinderwagen.com.
Pinderwagen's tricked out Volkswagen
Pinderwagen was extremely useful in helping me determine that my clutch plates were probably in good condition. I read their article on obtaining a used Kaaz unit similar to mine and saw the condition of their plates.
After reading about their experience tracking the car with the plates in the as-found condition for a full season, I found out that their car performed very well. I also saw that their data confirmed my understanding of speed from a plate-type vs. a torsen diff. As a result, I decided to rebuild my diff as it sat-- also per directions on Pinderwagen and from other sources.
Pinderwagen's article on rebuilding a Kaaz diff
Rebuilding the diff is as follows:
1.) Clean the internals of the diff with solvent
2.) Lube all components with fresh differential fluid (I use Torco RTF per recommendation from other racers)
Torco RTF
3.) Reassemble the diff in order from either Pinderwagen's article or the Kaaz video on YouTube.
