Scratch built traction control

[tanderson1]

So i had planned on buying a RaceLogic system but was putting it off due to cost/build progress.

When i went to buy one online this week I found that they are no longer made/sold/supported. I hit the search engine and found that they were widely regarded as the best units with over a hundred configurable options.

I got to thinking this is quite an easy system to recreate.

First a little history on me, I write software and design hardware products for a living so i have the skills. I actively write software every day in 6 different computer languages. Last week i pitched 3 new product ideas at work, this week i demo'd working prototypes for two of them, makes a total of 8 new products in the last 30 days. The company i work for is fairy small so I generally work alone.

I don't plan on selling these units however i do plan on releasing everything for free (code, circuits, etc). Someone else could make them or i could create some instructions on how to build one yourself.

So now that we have that established i was looking for input on features

Here is what i was thinking so far.

• Hardware would be a fast DSP able to sample all of the ABS sensors thousands of times a second
• Wifi Hookup to a laptop to edit every setting possible (no hard coded settings)
• Easy hookup (plug the existing ECU cables into this unit, short cable hooks from unit to old ECU)
• Wires in line with each fuel injector, reading the signal from the ECU and replaying the same values to the fuel injector. (with sub millisecond latency)
• By removing fuel from one cylinder at a time throughout the firing order I can purposely cause a misfire without hurting the cat converter
• Tap into the RPM wire from the ecu to get the timings... in theory i could also get this data from the injectors
• Get a signal from the brake light switch (LAUNCH CONTROL :-))
• Wifi settings control with an HTML interface that would work from any laptop or phone.
• Some kind of interface in the car for selecting modes. (rain, mid, med, high)... I plan on using a full size LCD touch screen like a Tesla
• JSON interface so settings and real time data can be set or read by other devices. (like my Tesla style screen)
• Datalogging of ABS, RPM, etc data to SD card. (might not do this, JSON interface would allow a laptop to do this)
• Full throttle shift, Pressing in the clutch could automatically keep the RPM locked or lower the RPM quickly to the the target shift point. Throwing the car in gear and releasing the clutch would release the RPM limit.
• Wire for clutch peddle, used for full throttle shifting.
• Accelerometer, measure Gforces, vehicle angle, etc.
• Steering wheel angle sensor? I could determine if the cars drift was caused by wheel spin.

Configurable settings (some of these could be automatically learned but overridden with the laptop)

• Pulses per wheel rotation for each wheel
• Wheel dimensions front and rear (required to know if a wheel is slipping or just larger then the other wheel)
• % threshold of slip before acting for each mode
• Max drift angle (I could have the car hold a specific drift angle on an off ramp..) for each mode.
• A map of slip %, etc by vehicle speed for each mode (You could potentially have it let you hold a drift under 100km but not allow any slip above 100, or completely disable at slow speeds)
• maximum vehicle speed (i plan to set this to 999 but you could use it as a governor)
• max RPM (another governor to prevent floating)
• The ability to learn or input steering profiles (IE how much each wheel moves faster then the others when pulling a very tight turn to the left or right)
• Launch control settings by mode... (Hold RPM, traction profile for a fast takeoff with limited wheel spin/hop)
• Datalogging rate (how many samples per second)

This race logic page describes exactly how their system interrupted timings.
https://racelogic.support/05Traction...Setup_Software

I have read that wheel speed sensors will generally have a resistance of between 800 – 1.2 K ohms. To test the Corvette sensors i just need to jack the car up so a wheel can be spun freely then probe each side of each pair until a signal is found. Typically the signal will be a sinewave. (I have an Oscilloscope that will make short work of this)

Any features that i am missing here?
Anyone else interested in this?
I have only been thinking of this for a few hours so i'm sure there are more possibilities.

Once i get my GTM to start (hope to be ready for this very soon) i can start to work on a prototype. Step one is to create a board that datalogs the factory ABS sensors and gives me some data to work with, next up will be creating a circuit that allows me to disable fuel injector pulses from a laptop. I can then throw those two together and wrap an API and an interface around it.

I cant see this board costing more then $200 completed in an enclosure with wiring harnesses. I'm not sure how RaceLogic justified the insane price tag.

[tanderson1]

Aliexpress seems to have lots of ECU connectors for around $20 a set... Might take some time to find a set that matches the Chev ECU but i bet I can find it if i look hard enough.
I should need two male connectors and two female connectors. (80 pin)
http://www.aliexpress.com/item/Free-...08.3.46.3OEVlF


They have metal ECU cases for about $7.80
http://www.aliexpress.com/item/56Pin...00009.1.F9Zy3U

Main Processor should be something like this
http://www.nxp.com/products/microcon...re-dsp:MSC8151

It costs around $100.

Circuit board printing should be around $60/unit.

Add in another $20 or so in optic isolators, SSR's, Wire

so my guess is $207.80 or so for a completed unit before assembly. Someone could sell the unit for $500 with a very healthy profit.

[Harley818]

just saw this on the new posts.......
You should post this on the 818 forum. We have the same problem with the fact Racelogic is gone.
Maybe generate a generic system that can be configurable on the multiple FFR platforms.

I haven't started looking into this yet... but am thinking of adding it in the future on my 818S

[tanderson1]

cool. I may branch out to the 818 crowd.

for the LS1 connectors there appear to be a few options.

These guys are discussing an LS1 4 foot extension cable made by painless, and the option of Speartech making a custom extension for $100.
http://ls1tech.com/forums/conversion...s-ls1-pcm.html

I could ask Speartech for a custom cable set
1) LS1 harness
2) LS Crate harness
3) 818 Subaru harness.
4) Mustang for the Cobra builds
5) Other ford harnesses for the 33 ford.

The wire set would only need to be 6" long and brake out the required wires to a separate connector that would connect to the traction control box. This would allow me to use a smaller enclosure which is cheaper and easier to find a home for as well as allow for additional connectors to work with the rest of the factory five cars.

[billjr212]

Sounds cool. Good luck.

Note though that a lot of the features you discuss require the ability to brake individual wheels. Simply cutting power won't help with holding a drift angle. Also, if you want to work towards the stability control concept as opposed to just traction control, you will need the steering angle input (plus the ability to individually brake wheels).

Not trying to be a wet blanket, just making sure you don't set yourself up for endless frustration.

http://jalopnik.com/this-is-how-abs-...work-513807036

[VRaptor SpeedWorks, LLC]

Wow! Sounds like quite an undertaking.....and an incredible opportunity for someone to make a business out of this. Just think of what a small group we are as FFR kit buyers/builders.....and the guaranteed interest and demand for something like this due to the vacuum created by RaceLogic. With all of that said, I don't know that trying to develop this to mate to specific engine harnesses and ECU's is the way to go......like you are talking about getting the mating ECU connectors so it would be relatively "plug & play". One one hand, that would be really nice. On the other hand, it is going to be nearly impossible to do. On the "con" side, I can think of:

1. The connectors themselves are bulky and now we have to find a place for this big, bulky connector where there is very limited space.

2. I've built nothing but GTMs for the past 9 years. Some used the old C5 ECU. Some used the E38 ECU. Some used the E67 ECU. Some used a Mast Motorsports ECU. Then throw in all of the Subaru variants. IMO, trying to make this plug and play is going to severely limit the potential applications for this.

3. Cost. Lots of labor involved in creating this.....instead of the connectors of the harness plugging in to the ECU....now they plug into a separate connector and then we need one more connector to plug into the ECU. In the old C5 ECU's, there is 80 pins per connector times 2 connectors. That's 160 pins for the ECU connector and another 160 pins for the mating connector. That's 320 pins. That's a ton of labor.

4. Reliability. That's 320 pins.......and 320 new points of potential failure.

Yes, probably the biggest pain in the RaceLogic install is cutting into the engine harness to interrupt the signal to the injectors.....but IMO, it's still a better option than trying to build all of it into an "interrupter cable" that plugs into the ECU. That is also what made the RaceLogic "universal fit".....and universally popular.

We've installed a few of them here and it does (did) seem to be a very effective system. If you can reproduce the effectiveness in a similar system, I have no doubts that there will be plenty of people very interested in this. I can only wish I had the knowledge in my head to come up with a system like that! If you could come up with a complete system for $500, I doubt you could keep up with demand. I wish you the best of luck with this project!

[Aceflo]

I wish I was knowledgeable to provide any input into this idea. However, I will say that I'm interested in adding traction control to my GTM.

[tanderson1]

Sounds cool. Good luck.

Note though that a lot of the features you discuss require the ability to brake individual wheels. Simply cutting power won't help with holding a drift angle. Also, if you want to work towards the stability control concept as opposed to just traction control, you will need the steering angle input (plus the ability to individually brake wheels).

Not trying to be a wet blanket, just making sure you don't set yourself up for endless frustration.

http://jalopnik.com/this-is-how-abs-...work-513807036

So i went out to have a drink with the guys from work today and this became quite the discussion. I have never used brakes when manually entering a drift, i have never used brakes to maintain a drift. I have always used throttle control to manually maintain the drift. However there were individuals at the table that completely agree with you. So i did a quick search on line.... this site http://www.enginebasics.com/Advanced...20Control.html it has the following statement "What’s really entertaining is setting up the traction control for a certain slip angle and then showing your friends a perfect drift through a parking lot. Its no wonder they outlaw systems like these for D1 drifting and racing. " So it would seem that the Race Logic box was capable of holding a drift angle as well.

I'm not saying your wrong as i could find lots of systems online that accomplish the same task via slip disks on each wheel or individual brake line controls... Time will tell, I should be able to write the software hit a parking lot and see if i can dial it in. :-).

you are talking about getting the mating ECU connectors so it would be relatively "plug & play". One one hand, that would be really nice. On the other hand, it is going to be nearly impossible to do. On the "con" side, I can think of:

I see your point. The idea would be to provide both options. A wiring harness would be an available option however a 24" pigtail connector could also be provided for those who want to splice. (splicing would also save the user $80 or so)

I wish I was knowledgeable to provide any input into this idea. However, I will say that I'm interested in adding traction control to my GTM.

I don't think your the only one :-)
are there any specific options you would be looking for?
How do you want to adjust complicated settings? would an IPAD/laptop style interface work?
How do you want to adjust the easy settings like changing traction modes or enabling launch control? Do you want a small module you could mount on the dash?

[tanderson1]

So first up is creating a basic circuit for reading ABS sensors. ( we will try to make this as generic as possible to support a wide range of vehicles )

So the ABS Sensors in most vehicles appear to be Variable reluctance sensors. (I originally thought that they were hall effects sensors but it turns out I was wrong)

VR sensors need waveform shaping for their output to be digitally readable. The normal output of a VR sensor is an analog signal, shaped much like a sine wave. The frequency and amplitude of the analog signal is proportional to the target's velocity. This waveform needs to be squared up, and flattened off by a comparator like electronic chip to be digitally readable. While discrete VR sensor interface circuits can be implemented, the semiconductor industry also offers integrated solutions.

Examples are the MAX9924 to MAX9927 VR sensor interface IC from Maxim Integrated products.

http://datasheets.maximintegrated.co...24-MAX9927.pdf

The Max9924-7 chip runs $3 if i buy them one off. Price drops to $1.20 if you order them 2500 at a time.

I am still trying to work out the best way to handle the incoming and outgoing signals. I am leaning toward using a small $5 process like the Cortex-M4 for reading the incoming ABS sensors and performing the basic math to determine what action should be taken. I would then have a separate $20 FPGA that will sit between the fuel injectors and the ECU. The FPGA will have the heavy lifting power to perform the required math and keep up with reading signals, performing the required math, and writing the signals all within microseconds. The final processor is a $3 ESP32 (these chips are not available yet but should fit the bill perfectly), the ESP chip will provide the user interface via a webpage using wifi. The chip will also interface with an SD card for datalogging and interface to the other chips for transmitting setting changes.

One item I haven't had time to consider is the ability to field upgrade the firmware, 3 processors makes that very challenging, especially 3 different types of processors that were each chosen for there specific capabilities. I'm not sure if this is a feature i need, units could be sent back to someone with the proper equipment for reprogramming if a software bug was found or a new feature was required..
For now i will focus on the Cortex M4 part of the equation and interfacing that with the abs sensors. I can layer on the rest as i go.

[Martin]

Just posting so I am subscribed to this thread. I have the race logic system for my Roadster (not yet installed, but soon).

As an aside, a lot of fuel injection control systems do traction control as well now. e.g. the FAST XFI 2.0.

Martin

[Buzz Skyline]

I have an 818s, and I am extremely interested. I'm a bit overbooked now, but I will be happy to help in any way I can , once I have a little more time. I have a some electronics and programming hobbyist experience.

I look forward to following your progress.

-Buzz

[07FIREBLADE]

The 818 crowd needs this. So I'm interested and also work at iWire so we could help on the 818 side with configuration needs or info if needed. The few harnesses that we built for customers that wanted to incorporate the race logic into we simply made a jumper harness from injectors to race logic Ecu to its abs sensors and power ground leads. The systems were removable if needed. A simple task if using the proper tools and the better approach so the product is more universal.

[tanderson1]

The 818 crowd needs this. So I'm interested and also work at iWire so we could help on the 818 side with configuration needs or info if needed. The few harnesses that we built for customers that wanted to incorporate the race logic into we simply made a jumper harness from injectors to race logic Ecu to its abs sensors and power ground leads. The systems were removable if needed. A simple task if using the proper tools and the better approach so the product is more universal.

That might be a more logical option... I like it, I will let you know when i get to the wiring side of the equation.


For now i am focusing on what processors fit the application (both now and in the future where someone might want it to work on a V12 engine, or something that redlines at 10K)
I was considering a cortex chip for reading the inputs and potentially even interrupting the fuel injectors

I need to read more about the different types of injectors that are available and how how they are controlled. Timing is VERY important, if the timings are off the fuel wont burn properly and will ruin a very expensive catalytic converter or turbo. This is the main reason i was looking at FPGA, the FPGA would be there for the sole purpose of maintaining those input/output timings under all conditions.

Its quite possible that the Cortex M4 could keep up if we keep the logic short and simple.
I might pick up a wide band 02 sensor to test with.

8000RPM is about the highest RPM i think we would need to worry about. 8000 RPM/60 seconds = 135 rotations per second * 8 fuel injectors for one rotation = 1067 injector pulses per second.

if we need to move to 10,000RPM or to 12 fuel injectors the number goes up to 1800 pulses per second.

so essentially we need to read from 8 digital inputs, decide what action to take and perform those actions at least 1070 times a second. I will need to look up what the Cortex pin cycle times are, ie how fast can we read from the pin and write to another pin cycle around do to it all again I could just pick one up write some simple code to see how many times i can get it to sample/write and loop.

but maybe i am overthinking this whole arrangement. My first thought was to take the current injector pulse as input, analyze it in real time and then creating my own signal output that would control the real injector. The danger though is if the timings fell out of sync by even the smallest amount it could cause unburnt fuel to head down the exhaust pipe which could cause damage. Instead we could listen to the fuel injector signal to analyse the timings. We then just need to interrupt specific fuel injector signals using the same timings

Here is my favorite logic so far.

1. Boot up with the car in listening mode.
2. Wait a few seconds until the unit sees a smooth idle
3. Record the firing order of the injector pulses over 10 RPM (this also teaches us how many injectors we are looking at)
4. Sit and wait for ABS data
5. If the ABS data says that we are slipping over our threshold matrix it starts working
6. Light up an indicator on the box to let us know that it kicked in
7. perform one of 3 different sequences that block injectors depending on the severity that we need to reduce slip (Exactly how the Race Logic unit would).
8. loop back to step 4

This is how we would block injectors (based on logic taken from the Race Logic website)
sequences.jpgsequences2.jpg

[Roger Reid]

What a great project. You spend $10 grand on body and paint. One slip up and you paint all over again. I like the TC for the safety as well as track worthy. I have the Racelogic on my recently completed GTM and it works great. I also have an Audi with a V10. So a 10 cylinder option would be interesting.

[fuggles]

You might consider one of the car pc boards already assembled. I have seen a few options with ARM processors, CAN interfaces, display ports ect. Having a built in CAN interface maybe easier to manage all of the data you want. ARM based ecosystem is extensive as you probably know.

http://developer.toradex.com/product-selector

[K3FY]

Here is a pricing point for you to consider:

I originally used a GM E38 for my LS2. This past summer I switched to MegaSquirt 3 Pro. A very nice waterproof design. The switch over was a lot of work.


I now have the aftermarket ECU used for A/C Control, Traction Control, PWM Wet Nitrous, A/F Safety Shutdown, Fuel Pressure Monitor & Shutdown, Gear Determination, Fan Control, Can Bus Broadcasting, Scaled Speedometer Output and built in SD or Connected PC data logging.

This is strictly Acceleration Traction Control, no braking.
In addition to the standard ECU inputs/outputs I have wired the following
VSS1 Rear in,
VSS2 Front in,
TC (traction control) on-off switch input
TC Slip % Adjustment Potentiometer input (analog)
TC active LED output
Nitrous on-off switch input
PWM – Nitrous Solenoid Out (6 amp)
PWM - Fuel Solenoid Out (6 amp)
Gear output to block nitrous in first gear
Fuel Pressure input (analog)
A/F Warning LED output
The Ecu and 8’ wiring harness with the ECU Connectors already installed was $1199.00.
It uses a 16-bit MC9S12XEP100 @ 50MHZ (+100MHZ XGATE coprocessor)
I added a Dual Channel VR signal conditioning board for $40. Fuel Pressure Sensor for $40. I used existing antilock VR sensors for both VSS’s.
Source code is publicly available.
Currently running firmware 1.3.4

Remember no braking, handling or suspension control.

It has worked out great but it takes time and patience.

The greatest benefit is the Turner Studio Software. Much easier than the HP Tuners, fraction of the price and includes auto tuning.

Mark
K3FY.com

[crash]

k3fy...Cool website. I would have liked to see the drag car though.

[tanderson1]

It uses a 16-bit MC9S12XEP100 @ 50MHZ (+100MHZ XGATE coprocessor)

That processor is about $7 in quantities of one and is far slower then the $5 processor i was looking at. It is however marketed as a rugged automotive processor however I can find nothing in the instructions that indicates that it would be any more reliable then the Cortex chip.

I do like the idea of being able to control Nitrous.. I could potentially have that as a feature, for now i will focus on the basic traction control features. Both processors have enough IO to handle Nitrous or a few other features.

I am heading out to CES next week. They have a whole hall of automotive toys, I figure i might get some ideas there. (My work pays for me to go, may as well take advantage of that)

[TouchStone]

I am heading out to CES next week. They have a whole hall of automotive toys, I figure i might get some ideas there. (My work pays for me to go, may as well take advantage of that)

Our manufacturer will be at CES "Jetta Company Limited". They might have our product on display http://www.robokindrobots.com/ check it out if you get a chance.

[tanderson1]

Our manufacturer will be at CES "Jetta Company Limited". They might have our product on display http://www.robokindrobots.com/ check it out if you get a chance.

I will have to check that out. The video on the website is pretty cool.

This project is moving along. Touchstone and i have been sending messages back and forth around processors (which i am certain will be the Cortex M4) and development boards that we can start building protypes with, today i came out with the first draft of the main processor pinout. If we want to support the possibility of a V12 engine we will use every pin that the processor has to offer which isn't a bad thing.

I think i also have a method worked out for flashing new firmware onto the device using USB which will flash both processors that the device will have at the same time.

[BigGoals]

I will have to check that out. The video on the website is pretty cool.

This project is moving along. Touchstone and i have been sending messages back and forth around processors (which i am certain will be the Cortex M4) and development boards that we can start building protypes with, today i came out with the first draft of the main processor pinout. If we want to support the possibility of a V12 engine we will use every pin that the processor has to offer which isn't a bad thing.

I think i also have a method worked out for flashing new firmware onto the device using USB which will flash both processors that the device will have at the same time.

Hello, I am curious if you have made any progress?

Also, I have only scratched the surface in regards to how the car is built, so excuse my ignorance, but what keeps builders from transplanting the entire corvette wiring harness/computer/ABS module/steering column(including position sensor)/ etc. so that you can inherit the traction control/active handling features?

Regardless, I like to build things and would be interested in any information you share about how to make/program one of these systems- I also have some programming background and definitely think it would be a neat project.

[VRaptor SpeedWorks, LLC]

There are a few people who have done a complete electrical system transplant, I believe. The biggest factor against doing this is that it's just a ton of work. The other factor is that you'd be stuck running a donor LS1 or LS6 engine unless you buy a new crate engine and "convert" it back to the old 24-tooth crank sensor so you can use the old C5 ECU with the new engine. Since the engine is in the back instead of the front, there are tons of wires that would have to be lengthened to reach where they need to go in order to retain all of the components needed. Then you have the issue of fitting all of those extra components into the GTM....which was not designed for them. Again, yes it can be done and some have done it, but it is lots of additional work.

[BigGoals]

There are a few people who have done a complete electrical system transplant, I believe. The biggest factor against doing this is that it's just a ton of work. The other factor is that you'd be stuck running a donor LS1 or LS6 engine unless you buy a new crate engine and "convert" it back to the old 24-tooth crank sensor so you can use the old C5 ECU with the new engine. Since the engine is in the back instead of the front, there are tons of wires that would have to be lengthened to reach where they need to go in order to retain all of the components needed. Then you have the issue of fitting all of those extra components into the GTM....which was not designed for them. Again, yes it can be done and some have done it, but it is lots of additional work.

That's understandable, I didn't think about wiring lengths with the rear mounted engine, and engine/sensor compatibility. Do you happen to know of any build threads where people have done it?

[mikeinatlanta]

Why not just use one of the other systems still on the market? I'm a fan of the racetronics throttle control, however, with throttle by wire it gets a bit tougher to implement.

[UnhipPopano]

I will have to check that out. The video on the website is pretty cool.

This project is moving along. Touchstone and i have been sending messages back and forth around processors (which i am certain will be the Cortex M4) and development boards that we can start building protypes with, today i came out with the first draft of the main processor pinout. If we want to support the possibility of a V12 engine we will use every pin that the processor has to offer which isn't a bad thing.

I think i also have a method worked out for flashing new firmware onto the device using USB which will flash both processors that the device will have at the same time.

As traction control needs to tie together Brakes, Steering, Engine and Telemetry, why not use a system off of a late model car that has a working system?

For the 818, Subaru implemented Traction Control in 2009. It should be possible to use the electrical system off of a 2009 donor, and get Traction Control.

The next evolutionary step is to have CAN Bus devices that are either interchangeable or able to use one of the analog outputs to an ECS for a working system.

[tanderson1]

For me i would want something tweakable. adjust thresholds, integrate accelerometers, create my own launch control etc.

Yes there are some on the market but they are very expensive and lack the ability to tweak everything.

and it will be fun to build.

(I am waiting to get my car to the point where it will start and run down the street before i start on this system. I should be at that point in the next 15 days or so)

[mikeinatlanta]

For me i would want something tweakable. adjust thresholds, integrate accelerometers, create my own launch control etc.

Yes there are some on the market but they are very expensive and lack the ability to tweak everything.

and it will be fun to build.

(I am waiting to get my car to the point where it will start and run down the street before i start on this system. I should be at that point in the next 15 days or so)

Now I get it. Basically an electronic version of when people don't get my thinking on some of my aero and sheet metal mods. You are operating levels beyond most if you are seriously considering designing your own system and I look forward to seeing progress reports. I still won't fully get it, but it will still be great to keep an eye on. Please keep it simple enough for us old farts who understand electron flow but are a little ignorant when it comes to ones and zeros. For me, I'll stick with simple things like ram air systems and diffusers.

[fletch]

I know this thread is way out of date, but I’m curious if this project ever came to fruition. I sincerely hope so since it seems there are no other options for traction control. If I’m wrong here, someone please correct me.

[Shoeless]

I'm not sure if this was ever accomplished, but I am going with traction control via my AEM Infinity EMS. This will replace the stock engine computer and give me tons of functionality. Not to mention I'll be able to tune every aspect of how my GTM will operate. Traction control was at the top of my list of must haves for my build and I did tons of research into this before going this route, but found exactly what I was looking for with AEM.