The 4 Levels of Drift Driving

As usual this is my personal opinion based on over 10 years of drifting and teaching various people with various setups. This article assumes the vehicle is reasonably well setup (locking diff, not blown shocks, properly sized/pressurized tires, decent alignment, etc). With that baseline, a driver should be able to move through these levels at varying speeds. Although, the order in which the inner milestones are accomplished can vary greatly.

Level 1(子): Donuts and Figure 8s

The basics of drifting is learning to start and maintain oversteer. The progression generally starts with learning how to break traction, then maintain the loss of traction, and finally how to control/modulate it.

The progression is generally:

  1. Full lock steering wheel (from a stand still) donuts while holding the wheel at lock
    1. Learning to modulate throttle to keep it going
  2. Full lock donuts (from a stand still) and then modulating steering to tighten and loosen the donut radius
    1. Learning steering inputs
  3. Donut with a running start (clutch kick or ebrake)
    1. Learning initiations
    2. Learning steering on initiation
  4. Figure 8s
    1. Learning weight transfer from side to side
    2. Learning steering in transition

Level 2(小): Linking Corners

Building upon the general concepts, the next steps are to gain new skills and polish what has already been learned.

These include (but are not limited to):

  • General initiations
    • Hand brake, power over, clutch kick, etc
  • Increasing radius corners
    • Line work
    • Hand brake extensions
    • Multiple clutch kicks (or more power)
  • Decreasing radius corners
    • Braking
    • Hand brake pull
    • Off throttle
  • Banks and other cambered corners
    • Left foot braking
    • Multiple hand brake pulls
    • Lots of techniques
  • Straights
    • Manji
    • Hand brake work
    • Multiple clutch kicks
  • Transitions between corners in general
    • Steering technique mid transition
    • Throttle inputs for various situations in transition

Level 3(中): Linking Courses

Once the technique tool bag is full, using all of them to reliably complete an entire course is the next goal. This level is mainly polishing Level 2 skills to the point where they can be reliably performed in succession with minimal corrections. Line work generally isn’t polished. The pinnacle of this is repeatably completing the course with no spins and no straightening.

Level 4(大): Tandem/Competitions

After courses can be linked reliably, the next step is to minimize errors and maximize speed/line/angle throughout the course. Competency can be attained by driving with other drivers and mimicking their techniques and lines. Also by attending competitions.

  • Line
    • In competition. Judges will designate the preferred line
  • Speed
    • The ability to complete the course quickly or catch up to another driver
    • “running away” from the following driver
  • Angle
    • Maximize angle without necessarily sacrificing speed
  • Proximity
    • Ability to get close to the lead driver without hitting them
Posted in Drifting, Theory | Leave a comment

Drift Tandem Hand Signals and Etiquette

When a driver finally steps up to tandem driving, it’s very intimidating. Learning how to lead and follow has much higher risks than regular solo runs. As a result, an introduction to the prestigious club of tandem drivers can ease the transition. Here is a few of the generally unwritten rules of tandem driving. (also see: Lone Star Drift’s Video for most of these rules in video format).

When two drivers meet up in a tandem line, the second driver generally holds up 2 fingers. This denotes a request to tandem with the two cars:


From there, the other driver will give a thumbs up or down. If it’s a thumbs up, the driver will then point at themselves then hold up a 1 or a 2. 1 meaning they lead, 2 meaning they will follow. From there, the first driver will thumbs up or down.

This interaction confirms how many drivers and who will be in what position. The drivers will then hold up the 2 fingers to tell the starting line corner worker that they are going to tandem. The run then proceeds normally.

Multi Car Tandems

When a track allows multi car tandems, things get a bit more difficult. A third driver can drive up next to the first two drivers and hold up three fingers.

Can I be third?

The drivers will then point to themselves or others and hold up fingers for which position each person will be in and confirm/change how many drivers are in the group. Again, they will communicate this to the starting line corner worker. Disagreement (EG: no, we are just gonna go two) is done with a neck side to side hand gesture:

Nah man

The corner worker will then hold up the number of drivers in that run. This conveys to the drivers how many people will be in the run and allow disagreement if necessary. This can continue for however many drivers there are.

Drift Nirvana Specific

Shenandoah bridge course and Jefferson south course have long and short courses. Asking the other driver which they prefer is done by a short or long gesture with hands to denote short or long respectively

General Etiquette

  • Unless there’s a reason, rotate who leads and follows
  • As a lead(ing) car, when spinning out or having issues, try to get out of the way (off course if necessary). Preferably on the inside of the corner
  • When spinning, if getting off course or out of the way isn’t possible, stay still so that other cars can predictably get around
  • Passing is generally frowned upon
  • The corner worker has the final say in how many drivers are in a tandem group
  • Every driver is responsible for their own car and their own damage
Posted in Drifting | Leave a comment

E46 can system and motor swaps

When swapping motors, the factory DME is no longer receiving signals. As a result, the cluster panel won’t show rpms, coolant temps, and lights like the oil pressure dummy light and the CEL. The E46 chassis also uses a ton of can bus connections between modules to double check VINs and pass data for other modules to use. Despite having can bus technology, the cluster panel and other modules still use analog/pwm signals for various instruments. For instance, the speedometer is a pwm signal from the ASC/DSC module. With my M3 rear end swap, this sensor no longer communicates with the ASC/DSC system so the cluster doesn’t receive that signal. But also, the can bus signal gets a ASC/DSC not working dummy light command from that module to the cluster.

To maintain most of the lights and VIN checks, I kept my factory DME installed but severed the can bus connection to the DME to the cluster. I don’t use the ASC/DSC system so the arduino uno just sends a signal to the cluster that it’s as ok as it can be.

I still have my “carduino” freematics module that plugs into the factory LS2 GTO PCM via the OBD2 port using an ELM327 chip. The ELM327 is just a generic chip used in most off the shelf OBD2 autozone/advanced auto parts stores to connect to most OBD2 systems. This reads things like coolant temp and RPM that is then fed to the arduino uno to send to the cluster. The carduino also has a GPS module to replace the speed input to the cluster via an analog output as well.

Once things start working, I think the cluster can bus has access to things like brake pressure and steering angle that I can send back to the carduino to log in the SD card. Having 2 arduinos isn’t really necessary but giving the cluster its own unit makes sense. Also it saves me time of mashing together all the components into one unit. The general idea is simple, but the execution has some nuances and isn’t always as straight forward as the diagram suggests.

Posted in Uncategorized | Leave a comment

LS PCM to E46 Can Bus – Part 3

There isn’t a huge amount to update from a technology perspective but there is much worth talking about. I did my best to read in the rpm readings from the PCM supplied wire but I just couldn’t figure out how to receive the PWM signal properly. Without an oscilloscope, I’m basically flying blind. I would love to figure it out one day but it isn’t super necessary.

From here, I want to use my old carduino datalogger to feed in speed/rpm/temps into the cluster arduino as a slave. The basic serial/uart connection should be good enough to communicate between the two. I did find some bluetooth uart solutions but I don’t think it’s necessary. The hope is to make the carduino a removable unit that sits around my rear view mirror and the cluster arduino permanently sits behind the cluster.

The carduino will use its ELM327 and pull rpm and coolant temps from the PCM via the OBD2 interface (It will likely be logging rpm/clt/iat/throttle only). It will also gather speed from the GPS module. All of this data will be sent via a 5 character “commands” to the slave/cluster arduino. Begin and end bits, an identifier, and 2 characters for the data. I’ll post more as the code gets developed.

This functionality should be basically the same as the Thaniel module. I hope to eventually incorporate things like the IAT Into the MPG gauge. And maybe hooking up the CEL so it actually works. I can also turn on and off warning lights for my own purposes in lieu of their factory uses. I don’t know what I would use them for, but it sounds interesting.

Sorry, no cool graphics today. I hope to have something soon.

Posted in Drifting, Technology | Leave a comment

LS PCM to E46 Can Bus – Part 2

Turns out the e46 cluster won’t tween between values. It just shakes between the two values. So I severed the can bus connection from the chassis and made sure important features still worked. It turns out the e46 cluster is a combination of can bus, k bus, and analog inputs. The can bus inputs seem to only be a direct DME input. Maybe some ASC/DSC.

At this point, the tachometer works and the general can system has been setup. By reading this blog, I found out that the speedometer is actually a 12v output from the cluster that can be conditionally grounded to output a proper speed. A 2N2222 transistor with a 1k inline resistor in cahoots with the arduino tone operation go the speedometer working correctly.

Cluster with only can bus inputs and speedometer

The can bus managed to get most of the nasty lights out. I know the seat belt is done via K bus but I’m not sure about the red brake light. That’s likely the parking brake analog input. Also, the temp gauge is centered as a proof of concept. I’m not sure I will actually hook it up.

Next steps are to read rpm inputs from the LS PCM’s output and probably the start of the arduino to arduino communications to gather ELM and GPS data to feed to the cluster.

Posted in Uncategorized | Leave a comment

LS PCM to E46 Can Bus

Been over a year since my last post. I have the LS2 installed in my E46 and the basics are in and running. The current issue is that the cluster rpm/speedo/temp gauge are all disabled since the stock BMW DME is installed but not hooked up to an engine. I decided to leave the DME in to maintain as much functionality as possible. The speedometer is normally read by the left rear wheel speed sensor but I did an M3 E46 rear swap and the sensor is different.

Looking around the internet lead me to the Thaniel module (facebook page is “can bus module”). After talking with them, the module won’t fit my needs. It uses an ELM327 to connect to the LS PCM and I need that for my carduino for data logging. Also, my factory DME is still giving the cluster signals so they will be fighting each other.

Of course this lead me to creating my own ardunio can bus module. I’m using an arduino uno and the sparkfun can bus board. Like every other can bus arduino module, it uses an MCP2515 on SPI. The sparkfun library didn’t work with any of my tests so I eventually found this sample code that has a generic MCP2515 “library”.

The LS1 PCM has a 5v output for RPM signal that I’m planning to feed into the arduino, then to the cluster can bus. The stock DME should be sending conflicting data so I’m hoping the cluster just tweens between the two values. So 3,000 rpms might show up as 1,500 rpms on the cluster. Otherwise, I have to sever the can bus from the DME to the cluster and troubleshoot what is lost from not having that connection.

Otherwise, the speedometer can be done via GPS. Theoretically, the coolant temp can be done through a 5v sensor or fed from my carduino. Regardless, I plan to run an aftermarket gauge to get a number. The stock coolant gauge is pretty vague.

Posted in Technology | Leave a comment

Evaluating Angle Kits

Usual disclaimer. I am not an engineer. Nor do I know everything about angle kits. This is a reflection of my current level of knowledge about angle kits. 

For most chassis, angle kits are narrowed down to either cut knuckles or Wisefab. In those cases, competition drivers will likely choose Wisefab and people with limited budgets will likely go with cut knuckles. In the BMW world (e30/e36/e46) options vary wildly and picking the right option is more than money limited. So if you’re in the position to actually pick your angle options, here are somethings to think about.

  1. Adjustable top hats

Angle kits generally come with longer control arms. The problem with longer control arms is that they give massive amounts of negative camber (-6 or more). General aftermarket camber plates will not give enough adjustment to dial this out. Some kits (wisefab pictured above) come with the option of fixed or adjustable camber plates. The camber plate on the left gives so much adjustment that the shock tower will have to be modified. The camber plates pictured on the right solve the same issue but will not be able to give camber adjustments since it’s just a fixed plate. It’s worth more to get the full adjuster plate. It is also worth noting that FD legal plates must leave the shock center point within the factory bolt holes in the shape of a circle.

2. Ackerman Adjustment

This one isn’t as much of a deal to anyone. Most kits these days 0 out ackerman and give no option to adjust it. If you’re a driver just getting into things, having the ability to try out various settings is a very good thing. For instance, the SLR kit for BMWs has 4 or more adjustment settings to go from 0 to close to stock ackerman.


3. Mechanical Trail (Caster)

There are 2(ish) factors that determine a car’s overall caster. The first is the actual caster in degrees the overall suspension components are angled to (what’s on your alignment sheet) and the second is the trail built into the physical knuckle via the control arm’s attachment point. The simplest way to explain trail is the picture above. It is measured by the distance between the wheel’s axis of rotation (blue) and the wheel’s physical attachment to the rest of the suspension (red). The farther apart, the more trail.

Tilting the suspension (caster) can add/remove this effect to an extent. But that will affect camber curves. The reason for wanting trail is that it helps the car’s self steer (IE wheel return). Shopping cart casters have no caster angle but the front wheels still have good self steer due to the mechanical caster.

Above is the SLR adapter bracket (an old version) that utilizes the factory control arm attachment point at the knuckle to maintain trail.

 This is the wisefab adapter. Notice that the yellow line is the actual control arm attachment point. This setup will rely on more caster angle to make up for the lack of mechanical trail. (Also notice the lack of ackerman adjustment)

4.  Actual Caster Adjustment

Most kits these days provide some form of caster adjustment. Either at the top had or at a lower control arm attachment point. But not every one does. Even then, the lower control arm adjuster is more useful than the top hat. It can also help relocate the wheel in the wheel well after the control arm changes its original location. This is one of the overlook ways to dialing in more self steer or other steering characteristics.

5. Back spacing

Quite a few angle kits offset the angle blocks from the knuckle itself. In my experience, this is to gain more realestate in order to properly locate the joints/nuts. The problem with this is that the point of articulation at the control arm is now farther away from the center of the wheels rotation. I don’t know 100% of the other down sides but at minimum, to maintain scrub radius, the wheel would need to be mounted the same amount inboard. Which, with some coilovers, may not be physically possible.

6. Roll center correction

Most kits offer some kind of roll center correction. The problem is that some offset roll center on the control arm but don’t do the same for the tie rod. At rest, suspension arms should be near parallel to the ground. If both control arm and tie rod aren’t on similar planes, suspension compression will cause uneven changes in geometry over travel. This can cause uneven steering feeling.

** Other considerations

  • Spherical bearings over balljoints
  • Parts availability for sections of the kit in the event of an impact
  • Welding/fabrication necessary to fit the kit on
Posted in Uncategorized | Leave a comment

FD 2018 Technical Rulebook Overview/changes

For some reason, FD’s yearly technical rulebook isn’t readily publicly available. Regardless, I found a copy and looked through it. As usual, the formatting is so wildly different from last year that you can’t just do a straight text comparison. They do highlight major changes in red so that’s nice. It is annoying that, in this not publicly available document, there are advertisements. There are also some typos or accidental duplications, but overall the document is well put together in comparison to last year’s.

There are several additions that seem to just close loopholes not exploited yet (no pressurized refueling allowed) or closing loopholes someone was maybe exposed for (you can have 2 batteries but they have to be connected and in use).

Notable changes are:

12v must be available for FD cameras/telemetry where the rear view mirror mounts. Seems like FD will soon be taking in car telemetry and record runs through their cameras. Later in the document it specifies provisions about data collection. This could be an interesting game changer in the future but looks like it likely won’t be used this year.

FD will be impounding all podium finishers immediately after end of competition and they will be checked for weight and tire size compliance. I suppose this will be quick to perform but why the sudden change? I feel like there’s drama here I don’t know about.

Technical manager is now called competition director. Although I have no idea why this is a change. Kevin got a promotion?

Bumper structures must be magnetic steel. This is one supposedly due to JR’s aluminum crash support system to save weight. I’m not sure why it was outlawed. Maybe just due to cost savings for other drivers.

Wing standoffs no longer allowed at specific tracks. Unless it’s trunk mounted. This appears to keep wings from hitting walls and shattering. Tracks include “but not limited to Long Beach, Orlando, Wall, Las Vegas, and Seattle”

If your vehicle has chassis tearing or subframe stud problems please contact the Formula Drift Technical Department. This isn’t new, I just like how there’s an exception for the E36/E46 chassis.

Brake systems may be biased only front to rear. No brake bias may be used in a side to side configuration. Again this isn’t new but I have heard of drivers biasing side to side in order to not lock up one of the front tires when lightly left foot braking without turning on the brake lights.

Wheel paint/tire stickers. This is likely related to the bullet above. Perhaps this will also help show ebrake drags too. I can see how many people will hate this change but I personally like it.

Wheel tethers are recommended on the front and rear wheel for 2018, but may be required for 2019. I like how they hint that it might become mandatory. I haven’t looked into ways of tethering wheels but if they aren’t too obtrusive, it should be a simple change.

Pro and Pro2 will have different patches for 2018. Because we need to remind Pro2 guys that they aren’t Pro1 drivers?

Tires sponsor section is completely blank. It’s definitely TBD but it’s odd that there isn’t anything listed at all. I don’t recall this in previous years.

Pro2 specifics:

Minimum 2900lbs, mandatory 260mm max width tire, and 18 inch wheels. The 260 width rule is fine. I find it a little odd that they are mandating an 18 inch wheel. This will level the playing field and most drivers would be on an 18 anyway.

Tires must be used unmodified, as supplied by an Official Tire Supplier. Filing, buffing, or any other disguising of tire sidewall is prohibited. This is a Pro2 specific rule. Maybe drivers wind up having to use a tire they don’t like and are sanding off the logos of the manufacturers? That’s a lot of effort to go through to hide what tire you’re running. The unmodifying tire rule (like sticky compounds or bead lockers) is still around for both Pro and Pro2; this is simply on top of it.

Changes that I think are odd, newly specific, or weird:

Air jacks are now explicitly prohibited. I don’t know of anyone currently using them and they would add extra weight. But I can see why this is now explicitly prohibited.

Skid plates are now allowed but can only cover what’s necessary. That’s nice.

In cabin adjustments are not explicitly not allowed. Except brake biasing front to rear but that isn’t allowed between runs or during runs.

Helmet visors must be down during runs. That makes sense but why is it written in the rules all of a sudden?

Posted in Uncategorized | Leave a comment

Case for the E46 330i/330ci

In an attempt to not make this a blog, I’ll keep this short. I crashed the E30 on a friendly private day, parted the car, kept the LS2, and now I’m in an E46 330. This posting will go over my thought process on choosing this chassis.

E46 Vs other models:
At the time of writing, possible drift chassis are still limited. With my primary concern being financial, anything with an initial cost over $5-7k was immediately removed from thought. 370zs, FRS/BRZs, newer mustangs, etc. Also, running costs were also taken into account. M3s/M5s can be found with dead engines but replacement ‘M’ specific parts can be very expensive. Even S14s/S13s are getting older and receive a heavy “drift tax”. Staying on the financial aspect, any chassis in the 90s and before were removed as parts are getting harder to find and prices for used parts are going up.

With all of those depressing conditions in place, the next logical step is to look into a vehicle’s merit as a possible drifter. You can see the drift car master list I used here. Due to weight, the G35/G37 were removed as possibilities, Mustangs in my price range were removed due to my dislike of solid rear axles, and miatas were removed due to the their short wheel base and my physical inability to fit in them.

With that, the field was narrowed down to RX8s, 350zs, and E46s. The RX8 was ruled out due to the LS swap difficulty and my inability to work on rotary engines. I was heavily debating between the E46 and 350z but ultimately picked the E46 because I could get one in 4 door and parts were far cheaper than the 350z. On top of that, locally, 350z are more common and I wanted to try to be different.

E46s vs other E46s
Within the E46 model, there are several models – 323/325/328/330/M3. Immediately the M3 is removed from my consideration due to cost. Here is a comparison between the remaining models.

Above is a comparison of power to weight between all E46 models and E36s. Ironically, even the 330 has a worse power to weight than E36 M3s but I hate E36s. The 323/325 have very lower power to weights but the 328 is also a contender. But due to the 323/328’s MS42 DME/M52B2xTU engines, their tunability is limited and the emissions system is much more complicated/fragile than the 325/328, the 328 was also removed from consideration.

Above is the brake rotor diameter x thickness in mm for E46s and E36s. The non M E46s have all the same front knuckle but very different front brake options. The rear of the 323 is particularly small.

The 330 specifically is a special E46 model, the brakes are so large that the factory equipped wheels at any trim level starts at 17 inches to clear the calipers. The rear has larger axles than the 323/325/328 and therefore a larger bearing, different differential, and trailing arm. Spec E46 specifically uses this model for these reasons. The 330 still uses non M parts for the subframe and many other parts so this keeps running costs down.

The case against the E46
The E46 is not without issue. The frame where the rear subframe mounts to (even on M3s) are known to crack. Reinforcements are a must. Differential options are limited. M3s have a very oddly locking LSDs and there is no other LSD option for the E46 from the factory. This also is coupled with more limited ratio selection in comparison to previous generations of 3 series. The rear suspension design is still trailing-arm-like with a very heavy main “arm” positioned in such a way that toe out on deflection is a major issue.

For my personal situation, the 330 sedan fits my needs. I consider this chassis and 350zs to be the next popular drift chassis. Both are capable, cheap enough, and available.

Posted in Uncategorized | Leave a comment

Yoshi’s Drift Box: Day 11

To follow up on the last posting, I managed to get the output to Dashware to work correctly. Turns out the GPS isn’t a reliable source of timing.

I wound up having a few issues with finding and syncing datalogs to video. The Gopro was numbering runs by the run in its own order while the logger just continuously took several laps until I turned off the unit. This made linking up what went where difficult.

To solve this, the logger now has a “split” button where it closes the current log and starts a new one. This way, you can format the Gopro to start the file naming from 001 and the logger will do the same. Each run, the Gopro will increment and you can hit the “split” button on the Carduino to keep things in line. When splitting, the screen will flash white so it will help sync the seconds in Dashware. On top of this, the logger now always displays the file ID so you could flash gang signs at the camera or write down the file ID somewhere.

With this addition, the Carduino’s gui is in shambles so I’d like to rewrite it somehow. I’d also like to start coding the drift scoring mechanism. Although that requires the secondary arduino to be setup and logging things like angle and steering angle position.

Posted in Drifting, Technology | 5 Comments