fe1rx Ohlins Installation

[arcticcat522]

your like a wizard....I agree, thanks

[Bullitt]

One of the best threads evvarr!

Many thanks fe1rx.

^ I second the wizard comment lol

[yllwwgn]

OP, I love this thread, but just be aware that by the time you are done with it, for the vast majority of us readers you may have to translate some of the knowledge shared in the form of:

Thing #1 (be it springs, sway bars, a particular damper setup, etc) = GOOD
Thing #2 = BAD

Ha ha ha.... just saying... you are bound to be approached with a question at some point in the form of " "But duz it rub yo?" Ha ha ha...

[E82MSport]

First I wanted to say thank you for the extremely detailed thread, it made my installation go very smooth.

Question - I decided to go with the supplied springs form Ohlins and found the ride and handling quite good although the front wheel/tire fitment seems to be limited by the lower spring perch as noted in your write-up. I'm currently running the stock wheel/tire fitment (M Sport) with 215F and 245R. I purchased a set of Apex ARC 8's in the 1-series fitment and have held off installing tires due to this issue. I engaged Apex on this and due to the various combinations of suspension setups they couldn't give me a definitive answer if a 235/40/18 would fit without spacers. Based on your write-up it looks as though I might be limited to a 225/40/18 in front using the supplied springs, would you agree? I'm in no real rush so maybe the Swift spring conversion from HPA might be the answer.

Your advice would be greatly appreciated.

[fe1rx]

Originally Posted by E82MSport

First I wanted to say thank you for the extremely detailed thread, it made my installation go very smooth.

Question - I decided to go with the supplied springs form Ohlins and found the ride and handling quite good although the front wheel/tire fitment seems to be limited by the lower spring perch as noted in your write-up. I'm currently running the stock wheel/tire fitment (M Sport) with 215F and 245R. I purchased a set of Apex ARC 8's in the 1-series fitment and have held off installing tires due to this issue. I engaged Apex on this and due to the various combinations of suspension setups they couldn't give me a definitive answer if a 235/40/18 would fit without spacers. Based on your write-up it looks as though I might be limited to a 225/40/18 in front using the supplied springs, would you agree? I'm in no real rush so maybe the Swift spring conversion from HPA might be the answer.

Your advice would be greatly appreciated.

I am glad this post was helpful. Coincidentally, I am working on a detailed look at the various wheel/tire fitment issues but have been too busy getting ready for my first track day of the season to finish it. As it happens, I have a set of the ARC 8 (8.5" x 18" ET45) wheels that just got 235/40R18 Nitto NT01 tires mounted on them yesterday. I haven't had a chance to put them on the car, but they will fit with my spring perch location. If your spring perch is not very close to the top you are likely to have a problem. If you let me know exactly where your perch is located I can adjust my suspension drawing to show you how much clearance you will have (give or take the variation between tires of the same nominal size).

The "perfect" tire size for these wheels would be 245/35 but selection in this size is very slim. Given that I think that 235/40 is the best size. But I am getting ahead of my next post ...

[E82MSport]

Thanks fe1rx for the reply. When I initially asked Apex about tire fitment I wanted to know if a 245/35 would fit in front, which seemed doable based on what I've read in the forums. I abandoned that for the 235/40 fitment after realizing the spring perch was soo low with the Ohlins kit but it seems that might be out of reach with this setup. I've taken some pictures, let me know what you think.

I didn't mean to high-jack your thread, so if you want to talk about this off-line let me know.
This is the sock tire size 215F



11mm of threads, maybe 11.5


BTW I set ride height/spring perch to recommend based on Ohlins and it came out within ~3mm of my pre-install ride height. I could raise it up 25mm but that would be far above the stock M-Sport ride height.

[fe1rx]

Originally Posted by E82MSport

Thanks fe1rx for the reply. When I initially asked Apex about tire fitment I wanted to know if a 245/35 would fit in front, which seemed doable based on what I've read in the forums. I abandoned that for the 235/40 fitment after realizing the spring perch was soo low with the Ohlins kit but it seems that might be out of reach with this setup. I've taken some pictures, let me know what you think.

I presume from your pictures that you don't have camber plates and that you set your ride height pretty much as Ohlins recommends, maybe just a bit higher. This just works for the OE wheels and 215 tires. With my GC camber plates the perch is a bit higher with the Ohlins springs, and it just works with a 225 tire. I don't see this working at all with 8.5 x 18 ET45 ARC 8 wheels in any tire size.

The post that follows documents my thought process in choosing a track wheel and tire combination. As I had this in mind early, I dropped the Ohlins springs right away.

Incidentally, I just got from BMW the ride height specifications for our cars. Using their method (which you have done) OE ride height is:

Front: Series 607 mm, Low-slung sport 592 mm, Performance Sport (SZP5A) 592 mm
Rear: Series 599 mm, Low-slung sport 584 mm, Performance Sport (SZP5A) 574 mm

[fe1rx]

No rubbing …

I must say that I didn’t quite appreciate what that goal really meant when I set it. I assumed that it would be possible to have an installation that would not rub under ANY condition between full droop and full bump travel. That is true at the rear, but is not possible in the front. That said, I have yet to experience any rubbing up front, but I certainly know that the right combination of events will cause a rub. That brings me to the concept of the no-rub envelope – that boundary of suspension travels and steer angles below which the wheel will not rub the fender, the bumper cover or the fender liner. Obviously the tire must also not rub on the strut tube or spring perch. The question is then, what envelope is satisfactory?

The no-rub envelope is a function of wheel width, tire size (width, profile and diameter) and static camber. It can be altered by modifying the fenders, fender liner and bumper cover, but mine are all stock. It would be nice to come up with a general solution to the question “what is the no rub envelope” for the hypothetical configuration we are considering, but there really are too many variables (which explains why so many people on this forum have rubbing questions, if not issues). And of course, any wheel and tire fitment experiment is an expensive one. We want to get it right. Hence this attempt to try to figure out what will and won’t fit.

My own starting point is the OE style 261 wheels with 225/40R18 Direzza ZII tires all around. They fit fine and are not the issue. I want to add a square track configuration with wider rubber on wider wheels. Ideally the wheels will be identical front and back. I am only interested in 18” wheels.

So here so here is how I started my own experiment.

Step 1 is to understanding the problem. The basic problem of wheel fitment is achieving the following conditions:

- clearance between wheel and suspension components under all possible conditions
- clearance between tire and suspension components under all probable conditions
- clearance between tire and bodywork under all probable conditions

My approach is to document my starting point so that I can see what effect any given change might produce. I do this by measuring, sketching and calculating, but I find a scaled drawing to be the most useful tool to visualize with. Because I have target static ride height and camber settings, these are built into the drawings. The drawings include the critical suspension arm lengths and motion arcs and are checked against measured camber gains to validate the model. Because tire size will be limited by the front geometry, I am focusing on the front.



This drawing shows how well the 8.5” x 18” ET 45 wheel fits the available space.

The question is now one of tires. Here are the potential candidates I considered with typical dimensions (recognizing there is significant variation in section width and corner profile among tire brands):

Size Diameter Typical Section Width
215/40R18.....24.7”..........8.8”..(OE Front Size)
225/40R18.....25.1”..........9.1”..(Current Street Size)
235/40R18.....25.4”..........9.6”
245/35R18.....24.7"..........9.8”..(OE Rear Size)
245/40R18.....25.7”..........9.8”
255/35R18.....24.9”..........10.2”

Fitting these tire sizes on the 8.5” x 18” ET 45 wheel I get the following drawings:











These simple drawings are 2-D slices through the vehicle at the wheel centers. They ignore potential rubbing in many places but they are a useful starting point. I have drawn the tires conservatively square in corner profile. My conclusion from this analysis is that while I would like to stuff something bigger in, the widest tire I am sure will fit is 235/40R18. I look a little closer at the tire to fender clearance over the full range of suspension movement (zero steering angle) and at my selected static camber setting it looks like I will have fender clearance.



At this point I want more information than can be gathered using my simple 2-D approach. With neither the Apex wheel or the 235/40R18 tire on hand, I instead mocked up the configuration by using spacers and the front OE wheel and the 225/40R18 tire I have mounted on it. This tire fits fairly squarely on the OE 7.5” wheel, as does the 235/40 tire on the 8.5” wheel. Accordingly, by spacing it out appropriately I can mock up the tire to fender clearance.



I installed the mockup configuration on my car with the spring removed from the front suspension. This allowed me to investigate all combinations of wheel center height and steering angle and to identify the extent of the no-rub envelope. As a benchmark, I repeated the same examination for the front wheel installed without a spacer, because I know from experience that no-rub envelope is sufficiently large that I have not had fender rubbing issues at the track or on the street (with M-sport springs and struts).



From this I expect that the wider 235 tire on an 8.5” wheel will result in a smaller no-rub envelope. The tires are much closer to rubbing when turned outward, but then the inboard turning wheel is generally more highly loaded. When turning inward the issue is generally the outer diameter contacting the fender lip at the top of the fender. When turning inward the issue is generally the corner of the tire contacting the fender lip or the bumper cover or the clip that connects them. Larger diameter tires will also come closer to the oil cooler outlet grille.





I concluded from many trials of various widths and offsets that Apex has nailed the 18” square configuration with their 18” x 8.5” ET 45 wheel. The early 1-Series pioneers went through many trials to get us to this point, so I certainly am not going to reinvent the wheel. Accordingly, I tentatively committed to a set of Apex ARC 8 18” x 8.5” ET45 wheels. With that variable pinned down, I redrew the suspension with this size wheel. That confirmed that suspension clearance is minimal, but positive, and that fender clearance is promising but tire-dependent. Of course I had previously determined that a shorter front spring (shorter than that provided in the Ohlins 3-series kit) was required to maintain tire clearance from the lower spring perch.

I am happy with the 225/40R18 street tires on the OE wheels. For track tires, I wanted something reminiscent of the old Toyo RA1, of which I have fond memories, for both its reasonable grip and its long life. With the demise of the RA1, the R888 would seem like a logical choice, but I was convinced to try the Nitto NT01 instead. As much as I would like 245s all around, they are available only in 245/40 and I saw the increase in overall diameter as raising additional rubbing concerns. 235/40 nicely fits an 8.5” rim with only a slight increase in OD from my street tires.

At this writing, the tires are now installed and scuffed in. As expected from all my drawing, strut clearance is no problem (with no spacers).



Neither is oil cooler grille clearance, although it is noticeably less than with the 224/40 tires.



Fender clearance appears to be quite reasonable both front and back. This is definitely not an “aggressive” fitment, but is a functional one without any fender rolling.





Speaking just from scuffing the tires in on the street and a few on-ramps, the NT01 tires are somewhat noisy (different but not really louder than the ZIIs), have decidedly more grip and less squirm/slip angle, but feel slightly numb on center. Just off center they load up quickly but the sidewalls feel a bit softer than the ZIIs.

I will be testing them at the track this weekend with GPS data logging so will have more to report after that.

[fe1rx]

As a comparison, here is the oil cooler grille clearance with the 225/40 ZII tires on OE wheels:



And the strut tube clearance:

[fe1rx]

Until this year I have tracked my car with OE brake pads, which, when nursed, worked. Tired of being out braked, I decided to install Hawk DTC pads for track use (DTC 60 rear, DTC 70 front). This choice was motivated by my being happy with the DTC 60s I installed on my FC RX-7 track car. The choice of Castrol SRF fluid and titanium shims was similarly motivated.

My track experience prior to the 135i has been primarily in a track-prepared FC RX-7 Turbo II. From that I learned that titanium shims work at insulating the calipers and fluid from the pad heat (raising the temperature of the pads in the process), that Castrol SRF can realistically last a whole season (and nothing else I tried could in that application), and that I liked Hawk DTC 60 pads.

This experience has guided my approach to a track brake setup for the 135i which is now riding on Nitto NT01 race rubber and an Ohlins Road and Track suspension (60 N/mm front, 120 N/mm rear). Clearly with those modifications the OE brakes would be completely overwhelmed at the track. Because of the sticky rubber, I opted for the DTC 70 front / DTC 60 rear option, which in theory should suit the higher deceleration rates and greater weight transfer that race rubber should provide.

I was unhappy with the idea of running the pads without shims for two reasons:

1) shims protect the nose of the pistons from abrasion
2) shims reduce heat transfer into the caliper and protect, to some degree, the boots from cooking

Hard Brake makes titanium shims for the front calipers, so I got a set of them in 0.02" (0.5 mm) thickness. I wanted also to wrap around the trailing edge of the pads as done with the OE brake pads to provide a smooth bearing surface to react the pad torque reaction force. I designed the stainless shims to provide full coverage of the pads and had them waterjet cut from 0.036” (0.9 mm) 304 stainless.

Rear Shims:



Rear Shim with Pad:



OE pads have shims that wrap around both ends and which provide a close fit in the caliper. Race pads lack shims and so are a rattling fit in the caliper (about 2 mm clearance when cold). Some additional clearance is required for the race pads vs. street pads because they expand more under their potentially higher operating temperatures. Rather than using thinner shims that wrapped both ends, I used a thicker (actually same as OE) material but only wrap around the thrust end of the pad, providing essentially 1 mm additional expansion clearance.



I took a similar approach with the front pad, first installing the titanium shim:



Then adding the stainless shim:



The bent end of the shims is always located at the top of the caliper so that it carries the torque reaction thrust from the pad:



Testing to date has been one track day at Shannonville Fabi. This track is fairly hard on brakes as revealed by the following velocity and longitudinal deceleration traces. (Speeds are km/h)



I drove approximately 60 laps – enough to burn a tank of gas. Brake cool-off consisted one cool down lap at the end of each session. Had I been driving with street pads I would have been more religious about cooling off the brakes, but I wanted to give the shims a decent test.

On track the DTC pads have a noticeably lower friction coefficient when cold, but quickly reach a temperature where they are noticeably “stronger” than OE pads. Cold friction levels can easily be lived with on the street, but the pads are abrasive and noisy when cold and so not a good choice for the street. Their abrasiveness when cold is a feature not a flaw in that the DTC pads quickly clean off the rotors for a change back to street pads. I experienced no hint of either pad or fluid fade on the track.

Changing back to OE pads, I noticed several things. At the rear to brake heat was sufficient to melt and partially burn the paint used on the pads. The melted paint is messy and tends to glue the piston boots to the shims.



I took this opportunity to glass bead blast the pad backing plates and shims to remove the paint residue so that next track day this should be less of an issue.



At the front, brake temperatures were obviously higher and the paint was largely incinerated rather than melted. I used no compound between the brake pads and the shims. All the residue is from the paint on the backing plate, or from the singed piston boots.



The new piston boots I installed are now singed and damaged, despite the shims.



Lack of heat colours in the shims indicates that they did not get excessively hot, but the piston nose design puts the boots directly in contact with the shims at all times. This is not really ideal as it cooks the boots. The residue of cooked boots can be seen on the stainless steel shims. A piston nose design that put a small air gap between the pad/shim and the boot would be preferable.



The front pads were worn to an overall thickness of 0.580” with a taper of 0.010” end to end. A new pad is nominally 0.616”, implying a total wear of 0.036”. At that rate these pads should last 10 similar lapping days.

[Ginger_Extract]

Very curious that the smaller coil-over body limits your tire selection so greatly. I was always under the impression that your options would open up w/ the extra inner clearance to work with. I'm using OEM M-Sport springs and dampers (now blown, will soon be Konis) w/ Vorshlag camber plates. On that setup, my front wheel/tire setup: 17x8.5+40 w/ 255/40 Hankook is a rub free fitment with more clearance than what you are reporting, despite the positioning of the tire with the 17" wheel. Just looking at the data, and the pictures, I'd wager that a 9" wide front wheel w/ the appropriate 255 section width tire could be accommodated with relative ease. You certainly have the camber (outer clearance), and suspension (stiff enough, greater inner wheel well real estate) to do so.

As always, excellent reporting. You are a true asset to this community.

[MarkkyyMan]

If this thread doesn't deserve to be a sticky, I don't know what does. Well done.

[fe1rx]

Originally Posted by Ginger_Extract

Very curious that the smaller coil-over body limits your tire selection so greatly. I was always under the impression that your options would open up w/ the extra inner clearance to work with. I'm using OEM M-Sport springs and dampers (now blown, will soon be Konis) w/ Vorshlag camber plates. On that setup, my front wheel/tire setup: 17x8.5+40 w/ 255/40 Hankook is a rub free fitment with more clearance than what you are reporting, despite the positioning of the tire with the 17" wheel. Just looking at the data, and the pictures, I'd wager that a 9" wide front wheel w/ the appropriate 255 section width tire could be accommodated with relative ease. You certainly have the camber (outer clearance), and suspension (stiff enough, greater inner wheel well real estate) to do so.

As always, excellent reporting. You are a true asset to this community.

There is actually no difference between the strut diameters. OE and Ohlins are both 52 mm.
You are probably right on the tires. My approach was quite conservative because I wanted to "guarantee" no rub without any rolling and I was trying to figure it out without actually having the wheels or tires on hand to try different options. dcaron9999 has managed 255s with a similar alignment on OE springs and shocks. The large variation between nominal and actual section widths could bite me with some tire brands though. Next set of tires I may try something wider, but I am pretty happy with the Nitto NT01s in 235. I am not sure a wider tire would give me any real world benefit. What would I gain except the bragging rights of wider tires? I would need to design a test to figure that out ...

[Ginger_Extract]

Have you actually tracked the car with the car as it is? There is a reason that 1M's, and other vehicles with similar size, weight, and power figures are trying to run 285+ tires on 10"+ wheels.
In regards to increasing section width without wheel width, 255/40/17 w/ 17x8.5 is a VERY common sight in the competitive AutoX community. Is it optimal? Likely not, as I mentioned, a 9" wide wheel would certainly make better use of the tire, especially in the front. That little bit of extra stretch would aid potential poor tire wear (if applicable), and sharpen up the responsiveness of the tire.
In due time, I will be figuring out how to get a 17x9, approx ET35-38 w/ 255/40/17 up front, and then roll with the de-facto 18x9.5 ET62 APEX w/ 275/35/18 in the rear. Although, as I previously mentioned, considering the tire:mass ratio for these cars, ideal would really be nearer to 10" front and 11" rear. However, that's only feasible w/ full 1M conversion, including quarters which is a messy business.
Sorry to go so off topic, but when considering the car setup as a whole, a salient issue to consider.

[02rsxpilot]

I experienced the same pad paint melt with my DTC 70s up front. Glad to know it's not an issue that would affect performance.

I'm going to have to take a closer look at my piston boots when I swap them back in this week.

[fe1rx]

Originally Posted by Ginger_Extract

Have you actually tracked the car with the car as it is?

Yes I have. Before I post some data though, this is how I collect the data:

The AIM Solo DL is a GPS-based lap timer and data logging system. It provides actual lap times and predicted lap times based on 10 Hz GPS data and internal accelerometers. In addition it can be connected to the vehicle OBDII port or directly to the CAN bus to log additional vehicle data.

http://www.aim-sportline.com/eng/pro...solo/index.htm

Real-time lap times, and in particular predictive lap times provide immediate feedback as to how well you are driving a track. The logged data provides even more information for analysis after a run or after an event. Once you are capable of driving consistent, repeatable laps at a track, the AIM Solo DL or other similar devices will make you a better driver. That is my sales pitch.

I had a Racepak G2X in my previous track car so I have some experience with GPS data loggers, but the Solo DL is substantially simpler, more powerful and has better analysis software. Most compelling was its ability to tap into the large number of chassis sensors built into our cars.

INSTALLATION

A minimal installation needs only a place to mount the device as it has an internal battery with a reasonable life. I got my DL with a Ram windshield mound, which is very sturdy. With no connections to the car at all, the DL will log the following data:

GPS Speed (km/h or mph)
Lateral Acceleration (g - from accelerometer in DL)
Longitudinal Acceleration (g - from accelerometer in DL)
Vertical Acceleration (g - from accelerometer in DL)
GPS Lateral Acceleration (g - calculated from GPS positions)
GPS Longitudinal Acceleration (g - calculated from GPS positions)
GPS Yaw Rate Gyro (deg/sec - calculated from GPS positions)
GPS Number of Satellites
GPS Positional Accuracy
GPS Path Slope (deg - calculated slope of path)
GPS Path Heading (deg - calculated track heading)
GPS Altitude (ft above sea level calculated)
Internal Battery Voltage (AIM Internal Battery)

AIM’s analysis software can download the data to produce track maps, segment times, lap analyses, etc. It is not my intent to sell you on the usefulness of GPS data logging and lap analysis. Others do that quite well.

M-World has a simple tutorial that is worth a look:

http://www.m-world.us/aimsupport/Dat...20-%202013.pdf

Racelogic has a more detailed look at data logging and analysis using a VBOX, but the principles are the same:

http://www.racelogic.co.uk/_download...it-driving.pdf

My intent is actually to review the details of installing a Solo DL in the 135i. I have seen some posts regarding the device in the 1M, but at a detailed level there are vehicle differences, and my own experience was not straight-forward.

The AIM manuals and analysis software are available for free on line. I strongly suggest their aim-sportline.com website. For some reason they (don’t) maintain another one that is hopeless by comparison. The manuals are not great. Their email and phone technical support is good, although they were completely wrong on a BMW-specific question.

http://www.aim-sportline.com/eng/download/index.htm

Richard at M-World was a great resource in getting my unit fully functional as he has experience with various BMW installations.

The device can be connected to the vehicle CAN bus via the OBDII port, or it can be connected directly to the CAN bus at the ECU. I bought my device locally, figuring that the OBDII connection was simplest, only to find that all taps to the CAN bus are not created equal and the data that I thought I would get, I couldn’t get.

Connected to the OBDII port, the following additional data parameters can be logged:

Calculated Gear (current gear based on speed and rpm)
OBDII Engine Coolant Temperature (°C or °F)
OBDII Fuel Level (%)
OBDII Intake Air Temperature (°C or °F)
OBDII MAF (I get no data on this channel)
OBDII Manifold Absolute Pressure (mbar, bar, or psi)
OBDII PPS (% - appears to track TPS)
OBDII RPM
OBDII Speed (actual not derated speed)
OBDII Throttle Position Sensor (%)
External Battery Voltage (Vehicle Battery)

Prior to logging anything, the DL needs to be configured with a logging protocol. To log data from the OBDII port the protocol has to be set as follows:

ECU Manufacturer = OBDII
ECU Model = CAN

This list of data is hardly compelling, and not why I bought the device. Logging engine data with an AccessPort yields far more.

AIM tech support advised, sorry that is as good as it gets. Richard at M-World advised otherwise so I followed Richard’s advice and bought a plug-and-play CAN bus harness from him. With that, I could use the following protocol to get 28 lines of vehicle data:

ECU Manufacturer = BMW
ECU Model = BMW_PT6

The M-World harness is a standard AIM CAN/RS232 harness with some very convenient crimped pin terminations that allow the ECU connection to be made without any cutting of wires. Here is the process:

1 Disconnect the battery



2 Remove the cover below the glove box to access the connectors



3 Remove the blue connector



4 Make the connection to the CAN+ and CAN- wires per the M-World instructions



5 Secure the harness



6 Route the harness to the device location



Connect the battery, ensure the DL configuration is set properly, connect the device. Then turn it and the vehicle on to confirm that it is receiving data.

Just to be clear, the BMW protocol does not work with an OBDII connection. Clearly not all CAN data is present on all branches of the CAN bus.

Incidentally, removing the side vents deserves a DIYof its own. Some of the DIYs I have seen are simply instructions on how to break things. The vent has 4 clips, the two inboard of which are only accessible through the grill. Working each of these clips in turn and pulling up on the bottom of the grill the whole thing pops out fairly easily. Brute force is not necessary. Disassembling the vent means breaking it. Putting the vent back in is extremely easy. Once you figure out how to retract the 4 clips getting the vent out isn’t to bad either.



With the BMW_PT6 protocol up and running the following additional lines of data are now available in addition to the basic GPS data:

Calculated Gear (current gear based on speed and rpm)
RPM
Accelerator Pedal Position (%)
Speed (BMW derated) (km/h or mph)
Speed (BMW actual) (km/h or mph)
Wheel Speed Front LH (km/h or mph)
Wheel Speed Front RH (km/h or mph)
Wheel Speed Rear LH (km/h or mph)
Wheel Speed Rear RH (km/h or mph)
Steering Wheel Angle (deg)
Clutch Switch (Inactive on 6MT)
Brake Switch (Inactive on 6MT)
Brake System Pressure (bar or psi)
Brake Pressure Front LH (bar or psi)
Brake Pressure Front RH (bar or psi)
Brake Pressure Rear LH (bar or psi)
Brake Pressure Rear RH (bar or psi)
Engine Coolant Temperature (°C or °F)
Engine Oil Temperature (°C or °F)
Outside Air Temperature (°C or °F
Outside Air Pressure (mbar, bar or in.Hg. - this is not MAP)
Gear (Inactive on 6MT)
Longitudinal Acceleration from ECU (m/s^2)
Lateral Acceleration from ECU (m/s^2)
Yaw Rate Gyro from ECU (deg/s - not deg as stated in software)
Distance (useless data)
Fuel State (%)
External Battery Voltage
Fuel Injector? (Inactive on 6MT)

Most data can be sampled at 10 Hz, some can be sampled at higher frequencies.

Not found in the AIM manuals is the requirement to run the Device Calibration routine from the RaceStudio software, while the device is installed in the vehicle in its normal orientation. Without doing that the DL accelerometer data will be incorrect.

Making up for the poor manuals, AIM has some useful tutorial videos that cover many subjects. They can be found at:

vimeo.com/aimsports/videos

I am interrupting the Ohlins post to cover this because the data from this device is necessary for some of the posts that will follow. For me, the single most compelling piece of chassis data is steering wheel angle, and the ability to correlate it with lateral g and speed. With this data, understeer and oversteer can be quantified. Also, individual wheel speeds and brake pressures permit the function of the e-diff and the actual benefits of a true LSD to be quantified.

Stay tuned …

[Ginger_Extract]

Interesting. Way to take advantage of the AIM product. Personally, I am a RaceLogic fan boy, with my VBOX getting a hearty work out over the years.
I will be curious to see your findings on e-diff v. true LSD. Going from my personal expectations and experiences, the e-diff can work quite well on smooth, and level pavement with smooth directional transitions. Think about your favorite 4th gear corner at your favorite road course. However, the e-diff truly falters with tight and technical changes of direction, and can be easily overwhelmed with autocross, especially on rougher pavement.

[fe1rx]

Time to report on some results.

Prior to heading to the track I wanted to get an objective measure of my handling balance. One method of measuring handing balance is by plotting steering wheel angle vs. cornering g at a constant radius (i.e. on a skid pad). The method is subject of an SAE Recommended Practice which I have approximately followed (although my 65 ft skid pad radius is less than the specified 100 ft minimum). I used the AIM DL data logger to record both cornering g and steering wheel angle at 10 Hz on a level skid pad.

According to theory the graph of steer angle vs. lateral acceleration should be a straight line at lower g levels, then increase non-linearly. In the linear range, the slope of the graph is the understeer gradient in degrees (of front wheel steer) per cornering g. If the line is horizontal, the car is neutral. (Sounds good, but neutral represents the knife edge between understeer and oversteer. It may actually better to be “close” than “there”.) The steeper the graph the more the car is understeer. The test is very sensitive to a lack of smooth driver inputs. Any lack of smoothness will corrupt the data.

I have done this test precisely once, on a cold (10°C) damp skid pad with sticker Nitto NT01 track tires. The results are interesting and deserving of a lot more experimentation to look at the effect of tire changes, tire pressure changes, damp vs. dry conditions, higher tire temperatures, etc. Below I have plotted the steering wheel angle against lateral acceleration for a left turn.



The yellow triangle represents the steering wheel angle associated with zero tire slip angle (i.e. the steering wheel angle corresponding to the front Ackerman angle). The Ackerman angle in radians is equal to the wheel base divided by the turn radius, which are both known. The yellow triangle corresponds to the Ackerman angle (8°) divided by the steering ratio (reportedly 1:16). Given that the logged data, if projected, should intersect the y-axis at this point, something is clearly wrong. Fiddling the numbers it looks like the steering ratio is actually 1:17.5 on center. The average steering ratio when calculated from lock to lock is presumably 1:16.

Plotting the data against front wheel angle is actually more interesting. Since we are interested in relatively small steering angles from center, I have assumed 1:17.5 in making the conversion from steering wheel angle to front steer angle. The data for both left and right turns together is easier to see when the y axis the absolute value of the lateral acceleration.



The data is very easily corrupted by the slightest lack of smoothness in either the throttle or steering. The right turn data above shows some of the irregularities that are introduced by any lack of smoothness. Next time I repeat the test I will try using the cruise control to maintain and nudge speed upward to see if this improves the data.
What is remarkable is how neutral the car is up to about 0.8 g. Above 1 g there is a lot of noise in the data. Perhaps some of this can be attributed to the tires being cold. They never got warmer than slightly warm to the touch. I have to say I was surprised at how neutral the car is “out of the box”, with just my best guess suspension, damper, alignment and tire pressure settings. I am glad that I did not make any anti-roll bar changes as part of the modification, because clearly, with my current spring rates, ride height and alignment, no change is necessary. A stiffer front bar would make the car understeer more.

We can say that my car is Neutral Steer (NS) in steady state cornering. This is not to say that it is NS in all conditions. At the track, this likely corresponds to mid corner NS behavior. At an autocross (at least the ones I attend, Kgolf31’s look a lot like race tracks) the car never really sees steady state conditions.

Confident that the handling balance was good and the grip levels were great, I headed off to Shannonville for a lapping day on the Fabi circuit. Fabi is a flat, 9-turn circuit with 6 right-hand turns and 3 left-hand. Two of the turns are 2nd gear hairpins. The long (180 km/h) straight is followed by a 100 km/h sweeper. One of the turns is flat from turn-in to exit. It is a nice technical track that I am quite familiar with.

Complimenting the track tires were Hawk DTC 60/70 brake pads, which handled the braking requirements of this track without the least complaint. The day was cool (12°C) and overcast but the track remained dry all day. Handling was remarkably composed (all nannies off, of course) and the car very easy to drive. I made no adjustments to tire pressures or damper settings from my street settings at this track. Even driving at this level, it was difficult to get heat in the tires under the conditions (I probably should have started with lower tire pressures, I was at 36 cold all around).

A track map:



Speed vs. distance:



G-G plot for one lap:



For this same lap, here are the brake system pressures:



It is interesting to note the higher front brake system pressures, which will be due to some combination of electronic brake force distribution and the brake bias valve. Also of interest is the rear brake activity which is not symmetric left to right. This represents e-diff brake activity. To highlight this activity I have defined a math channel in the analysis software that is the difference between the left rear and right rear brake pressures. As symmetric pressures cancel, we are left only with the e-diff activity:



In the tight right-hander leading onto the long straight you can see that 33 bar of brake pressure was applied to the inside rear brake in order to prevent excessive wheel spin. This compares to 44 bar being applied symmetrically to the rear at he heaviest braking point on the track.

Math channels present an interesting possibility to calculate the understeer balance at all points in the track. I have done this in the following manner:

1) the instantaneous radius of a turn can be calculated from the lateral acceleration and the speed:

R = V^2/A ,where R in meters, V in m/s, a in m/s^2)

2) the NS front steer angle in degrees can be calculated from the wheelbase and the instantaneous radius by geometry and the small angle approximation:

gamma = (180 / Pi) x (WB / R) , where WB and R are in the same units

3) this can be converted to the corresponding NS steering wheel angle in degrees using the on-center steering ratio:

BMW_NS = gamma1 x 17.5

4) the instantaneous front steering wheel angle, STEER_ANGLE, is logged by the DL in degrees.

Plotting these two parameters on top of each other it is readily apparent when the car is US (actual steer angle greater than NS angle) or OS (actual steer angle less than NS angle).



In summary, the car is generally US but slightly OS between turns 1 and 2, 2 and 3, the exit of 6, the exit of 8 and the exit of 9. These angles are at the steering wheel. It is more useful to know the actual US or OS at the front wheel. That is the difference between the two steering angles divided by the on-centre steering ratio. If multiplied -1 for right turns and +1 for left turns this new function is positive for US and negative for OS.



The prevalence of US and OS in this lap is now even clearer. Both are limited to about 2 degrees at the front wheel (35 degrees of steering wheel input). In the absence of comparative data it is hard to know exactly how this compares to other vehicles. My own conclusion is that the car, driven at a high level (based on the G-G graph) is remarkably neutral even in the dynamic environment of this race track.

The next test was a local autocross. This one held at a 3/8 mile paved banked oval. The course was set with a large number of slalom gates and some interesting off-camber turns. Proximity of concrete would not meet normal autocross rules, but the course sure was fun. The image below is generated from a .kml file exported from the AIM DL logged data and opened in Google Earth.



The corresponding track map viewed in the AIM analysis software assigns corner numbers:



The speed vs. distance graph and the G-G plot give a sense of the course:





Clearly this is a highly dynamic course with no steady state sections. I was interested in graphing the front wheel US/OS function I developed from the track environment to see what that it showed in the autocross environment:



Not surprisingly slip angles are much higher in the autocross environment, with correspondingly increased understeer and oversteer angles. Oversteer is pronounced at the quick turn from the start, the two pin turns. Elsewhere, understeer predominates but is never excessive. This handling made the car comfortable to drive at the limit, despite the proximity of concrete walls. The strict accuracy of the understeer function in such a dynamic environment is questionable, but it will likely be a useful tool for looking at the effect of setup changes.

To answer Ginger_Extract’s question about e-diff activity at an autoslalom, I have graphed the data from the autocross. The activity seems relatively modest, but this event was on very smooth pavement.



What is next is another track day at another track, and I want to have a more careful look at the steady state US gradient with some more skid pad testing, using some of the lessons learned. Also if I can find some other AIM DL equipped vehicles that are logging steering angle, I would like to find out how they compare with respect to steady state and on-track US / OS.

[The_Dude]

Not to change topic off your data logging, but have you modeled the suspension to come up with your ideal ride height and camber/caster/toe settings? I didn't see CG or roll center geometry to try and minimize roll moment arm. I would assume someone with you attention to detail and data analysis to have performed this before arriving at your targets and then adjusting with testing.

I've long wanted to be able to easily enter this data without the use of a 3D CMM arm, but the string and plumb method on paper seemed a bridge too far, even for the engineer in me and I generally just try experimenting and using lap times and calibrated butt analysis (CBA) to try and arrive at the ideal geometry settings. Problem is I'm not a professional driver and I'm often guilty of adjusting to many variables simultaneously to get a true A-B-A comparison.

At any rate, carry on - phenomenal thread and superb attention to detail. Thanks for taking the time to post all of this. I'm sure I will gain from this knowledge as I start to modify my 135 suspension (KW V2, 235/40/18 front and 255/35/18 PSS rears are only mods so far - Cut rear fenders and heavily rolled fronts, plus shaved the fender to bumper mounting bolt up front)

[fe1rx]

Originally Posted by The_Dude

Not to change topic off your data logging, but have you modeled the suspension to come up with your ideal ride height and camber/caster/toe settings? I didn't see CG or roll center geometry to try and minimize roll moment arm. I would assume someone with you attention to detail and data analysis to have performed this before arriving at your targets and then adjusting with testing.

I've long wanted to be able to easily enter this data without the use of a 3D CMM arm, but the string and plumb method on paper seemed a bridge too far, even for the engineer in me and I generally just try experimenting and using lap times and calibrated butt analysis (CBA) to try and arrive at the ideal geometry settings. Problem is I'm not a professional driver and I'm often guilty of adjusting to many variables simultaneously to get a true A-B-A comparison.

At any rate, carry on - phenomenal thread and superb attention to detail. Thanks for taking the time to post all of this. I'm sure I will gain from this knowledge as I start to modify my 135 suspension (KW V2, 235/40/18 front and 255/35/18 PSS rears are only mods so far - Cut rear fenders and heavily rolled fronts, plus shaved the fender to bumper mounting bolt up front)

1) Ride height on the 135i is already low in the real world of speed bumps and curbs so I didn’t want to lower it so much that the car scraped anywhere it didn’t previously. Also you lose camber gain on a strut front suspension if you lower too much. Given that any one change has a knock-on effect, a favorable camber gain curve with a higher cg is probably more beneficial than a less favourable camber gain curve and a lower cg. This was the motivation for my lowering decision.
2) Camber was chosen based on personal experience and general rules of thumb. 1 degree more negative camber at the front vs. the back is pretty standard practice, and -3 to -3.5° front camber is pretty normal for a for a track setting.
3) CG height is what it is. I haven’t measured mine yet, although there is a fairly simple method for doing so with reasonable precision.
4) Roll center heights are also fixed by the suspension geometry, athough they move with heave and roll. Calculating the static front roll center height is easy for a strut suspension (it does vary with static camber and ride height). On our cars it is close to the ground, which is why the front bar is much stiffer than the rear.
5) The rear roll center height is well above the ground, which is the other half of the explanation as to why the rear bar is so flimsy. I don’t really have enough geometrical information from my plumb bob and measuring tape method to pin it down. I wish I did.
6) My intent was not to arrive at an optimal setup by modeling the suspension before modifying it, but to model the suspension so that I knew what I was starting with, and so might have some clue as to what I might want to try changing, and to understand what effect that change might have. Any changes must necessarily be motivated by actual measured results. For that reason I tend to change nothing at the track until I can consistently drive the track at a high level with a fixed setup. There is no point in confusing one’s own learning curve with a change to the setup.
7) There is a parameter called the “total lateral load transfer distribution” which describes how the vehicle roll stiffness (due to the tires, springs, motion ratios, anti-roll bars, cg height and roll center heights) act to resist body roll in corners and in doing so transfer weight to the outside tires. This parameter has also been called the “magic number”, and as a starting point it should be 5% higher than the static front weight distribution. This ensures that the basic response of the vehicle to steady state cornering is understeer, but not too much so. If you have been following my posts, you will know that I know many of these parameters, but not the cg height (which I simply haven’t got around to), or the rear roll center height (which is difficult to pin down with enough precision to be useful).
8) Optimum G has a spreadsheet that can be filled in to estimate the magic number of your vehicle. Attempting to fill that in is informative because you quickly realize how much data you are lacking. Also, the calculation is already an approximation so if the data used is not very accurate, the result is suspect. One of the reasons for my post is to fill in some of those gaps for others. You can get the spreadsheet on line and playing with it is interesting to get some feel for how sensitive the magic number is to changes in the various parameters.
9) I did not change my anti-roll bars for one very simple reason. Without knowing my “magic number” I would be completely shooting in the dark as to whether it will balance or unbalance my car. From actual tests of the car’s neutrality, I am now quite sure that a stiffer front bar would not help MY setup. I believe (and I am still trying to pin this down with data) that my car is (by blind luck) very close to neutral in the linear range (where the magic number applies), which means my magic number is greater than my static front weight distribution (52%) but not by much. That, of course, is ideal.

[Kgolf31]

Lowering the front more though helps front end grip though...yes?

[PeteA]

Nice write up. I have a similar setup to yourself. Ohlins R&T with swift springs, thrust sheets and vorshlag camber plates. 7" 60NM/mm front spings, 9" 120NM/mm rear springs.

Sitting on APEX ARC-8's with Ventus RS-3 rubber. 235/40/18 fronts 265/35/18 rears. No clearance or rubbing issues on a ride height 10mm lower than stock. Was toying with the idea of 245/35/18 up front as a 20mm tyre ratio gap is a more neutral setup. I'm not 100% sure if it will clear the strut. The perch is a non-issue for me.

Will probably be pretty tight.



Given all your research, what alignment settings would you recommend as a dual duty street/track setup?