Off Road Traction Control

[robotteck]

I plan to build an off road traction control system roughly based on this link.
http://blog.chryslerllc.com/blog.do;jsessionid=F83B4BCDB38454BD1A43D066882C5EAE?id=270&p=entry

My truck has wheel sensors on each rear wheel to monitor speed. I will have an arduino monitor the wheel speed of each wheel, and when the percentage of change reaches a certain tolerance, the arduino will intervene the faster spinning wheel by applying brake force to the e-brake cable.

I have seen a lot of tutorials to build a window motor servo.
http://www.fieroaddiction.com/servo.html

The servo will apply the needed force to operate the brake cable.

The servo I need for the project needs to be waterproof. My thoughts were, if I used the pot to monitor degrees of motion, then the water proofing portion of this task would be much more difficult.

And, I only need the servo to be in a center position, 90 degrees left, or 90 degrees right. I had an idea of using magnetic switches, similar to home alarm systems. The switches are sealed, and assumably waterproof.

If I had a switch for each position of the servo arm(3, left, right, and center), would I be able to make this work with an arduino?

[cr0sh]

A window motor or wiper motor may or may not have the power needed; plus there's the possibility of stripping gears. I would actually use linear actuators with built in feedback potentiometers (whether you'll want to use one or two is up to you). Such actuators are more robust, can be found in waterproof versions, and generally also have built-in end-stop switches. They are driven just like a DC motor; in theory, you could even apply the idea of the "servo-conversion" to one, as long as it has a feedback potentiometer (or you use an external one).

Just make sure you get the proper one - some have different ratings for pull vs. push (extend/retract), some use a pulse output for position sensing (so you count pulses to determine where it is positioned), some are designed so that they can be "back-driven" (you may or may not want that?) - and of course, you need to determine how fast the extension/retraction will need to occur.

They also aren't cheap, but then again, you are talking about making this a part of your vehicle - you -do not- want to skimp on something like this (especially when it comes to speed controls on wheels); or at least I would think you wouldn't want to...

Here's a dealer of such devices, but look around for others - there are many manufacturers (and several have ones designed for automotive purposes - but generally for things like opening hoods or trunks; apps of that nature):

http://www.firgelliauto.com/

[robotteck]

I am unclear about how much torque a window motor can produce. Although, I do know these motors are used for robotics competitions because they are cheap, strong, and easily turned into servos. I saw a build on a simulator chair where these motors were capable of moving an average sized man plus the weight of the chair on demand.

http://www.x-simulator.de/forum/my-simple-sim-motion-chair-t1667.html

The cost effectiveness was definitely a challenge to prove that this could be done without a lot of funding, but more ingenuity. A rear locker for a vehicle could cost $500 to $800. On the lower end of the spectrum, a linear actuator could cost $100. Two actuators, plus the electronics, may start to get closer to the cost of a locker.

My original idea was to piggy back a 4 channel anti lock brake system, unfortunately the vehicle has a 3 channel ABS. This would not allow the individual braking of the rear wheels

Snooping around the web, I stumbled upon the window motor servo. They are roughly $15 - $20 and don't take too much effort to get running. Thanks for your information. I was unaware there was an off the self solution to this project.

The traction control system operates at speeds of 0.5 to maybe 5 mph. I do not think I would be skimping out by using these motors, with the exception to the extra step needed to waterproof them. I would have to factor that in, how much cost and effort it would take to waterproof a motor of that kind. As far as precision is concerned, I will not need precise movements.

I do have experience on the locker used on the Toyota Tacoma. It uses a window like motor, and is exposed to watery environments. Actually, they were poorly designed for water, and many malfunctioned shortly after use. I had to refurbish and reseal the one I owned and it did not take too much effort.

[mykiscool]

One thing that you should definitely do is have a differential. This will spread out the traction between the two wheels (or four if you put on on the front also) and give the non slipping wheel more torque.

[robotteck]

The vehicle is a Toyota 4Runner, and it does have a differential.

[elsp1991]

Wow I also consider to build a e-lsd, never thought that someone have the same idea with me 

I just made a search and found out some pretty strong linear actuator but was too expensive(250E).

So i tried a windscreen wiper motor from a bmw and was ultra strong, I used this to operate the clutch of an R6 yamaha engine for an FSAE car. It was able to pull about 20kG in a 4cm lever with a very good mosfet H-bridge @12V(PC PSU). So a total pull of about <8cm which is too much for e-brake cable, potential a smaller arm can used to gain some force. In my project i build a small drum out of a pipe in order to have steady force an unlimited travel cause the 20kg was not enough so with 2cm drum diameter i manage to pull the clutch in less than 1 second and with a single turn a travel of Pi*D = 6,...cm which was enough

If you plan to use two motor you can help the window motor with a spring not torsional.

Do you know how to pick the signal of the VR sensor without affect the ABS reading i dont want ABS to fail in any case. Maybe a high impedance input did not affected but I am not an electronic guy.
Also there are some IC that coverts the frequency to voltage but it needs zero crossing, so they are suitable for VR sensor in the other hand doesnt work with hall effect.

[robotteck]

Hello elsp1991,

Using a window motor as a servo is a great solution to this system. In the mean time, I had the idea of piggy backing the ABS system it self. Instead of a linear actuator, or motor controlling the brake, why not use the ABS pump it self. Potentially the ABS pump has valves, that close off the break pedal to individual brake lines, and then applies brake line pressure using the ABS pump (you need a 4 channel/valved pump).

Great attention to detail with the VR sensor. I never thought about that, however, I do agree that tapping into the VR signal could create extra draw that would alter the signal into the ABS system. At the moment, the only thing I could think to do, would be to mount an additional sensors. The traction control system I intend to use was only for off road use, and would not be using, and/or disabling the ABS during use. Here is a notion, the brake pedal signal could disable the traction control system during use, thus working independently from the traction control (set relays/logic gates to toggle signal feed into TCS). With that notion, I do not believe the ABS actually works until you hit the pedal, because it does not randomly brake while driving (excluding any vehicle with on board TCS traction control system).

Back to linear or window motors mounted to a floating axle. The unsprung weight of the floating axle causes a lot of G-force on what ever is mounted to it. I would be cautious to sensitive, or precious electronics being mounted to this type of environment (don't forget about water).

Also, I'm pretty stoked that at least one other person is interested in this topic.

[elsp1991]

I think ABS doesn't have pump to generate presurized brake fluid, pressure created from your foot. Mostly Abs just have a release valve(pressure valve) per channel and open it in case of lock in order to reduce pressure on the caliper and the most recent units (EBD) change dynamicly the pressure distribution.
ESP in the other hand has a built in pressure pump to create the pressure to engage the brakes even if the driver doesnt press brake pedal.

For me isnt possible to customize (piggy backing) the abs unit cause it has 3-channels. And to be honest i dont think its is possible, also the electric drive is better cause it has proportional effect and can control the braking torgue better (maybe PID) instead of pulsing on-off.

I found that tapping into the VR sensor doesnt affect ABS. Racelogic aftermarket traction control do this succesfully, no problems reported.


So becoming possible, i just found some chips for VR sensor interfacing (LM1815, MAX9927), actually is just a conditioner circuit.

NCV7001 chip offer 4 VR sensor support
Found this also http://brickems.com/brickrpm/

[dewy721]

I got this... (I'm a diesel mechanic, ABS specialist.)

Some older (junked) school buses had early version abs that required a pump that could be pulled off separately for service (which never needed it). One of these pumps would give you AMPLE braking power. They're made to run on 12v systems, even weatherproof. All you would need is two hydraulic T-way solenoids to re-route each rear wheel cylinder to either the vehicle factory system or the pump, as per arduino command. The arduino can get is traction control signals directly from the wheel sensors.

Here's how:
First, the anatomy of and ABS wheel sensor. - Most common wheel sensors are actually just a magnet with a coil wrapped around it. This sensor when placed close enough to a rotating sensor ring will distort the magnetic field of the sensor. This causes a signal to be generated but only when the wheel turns.

Second, the Arduino. - Having inputs tapped from the wheel sensor, the arduino looks for one input to go silent. (Stuck wheel, other wheel spinning wildly.) The Arduino then fires up the pump, activates the solenoid for the spinning wheel and clamps the brake until the other wheel rotates at a similar speed. (Usually within 10% tolerance.)

Tada. Basic traction control the way industry does it!
Advanced traction control is just more signal evaluation between the all four wheels plus throttle reduction, ABS is just traction control applied when the brake lights are on.

Hope it helps.

[dewy721]

Oh, by the way. Open differential vs. locker diffs.

Locker diffs when facing slippery conditions maintain the proper gear ratio, whereas open diffs divide the output ratio between the wheels. So if an open diff gets a wheel stuck, the other sides responds by spinning twice as fast.

For your average SUV this isn't a problem. But for professional rock crawlers with $1400+ tires and power management systems. It is. For them old-school military locker diffs are the only solution vs something expensive going BANG!

[robotteck]

Dewy,

Great information.

I do know the ABS system uses valves, and a reservoir to close off the brake line and accommodate in coming pedal pressure.

Dewy give us more information about this pump you speak of, what it's called and where I could get one?

The last thing you mentioned was, "ABS is just traction control when the brakes are on." Would that mean if one could activate the system with out the brakes applied, would it work like a traction control system?

I got this... (I'm a diesel mechanic, ABS specialist.)

Some older (junked) school buses had early version abs that required a pump that could be pulled off separately for service (which never needed it). One of these pumps would give you AMPLE braking power. They're made to run on 12v systems, even weatherproof. All you would need is two hydraulic T-way solenoids to re-route each rear wheel cylinder to either the vehicle factory system or the pump, as per arduino command. The arduino can get is traction control signals directly from the wheel sensors.

Here's how:
First, the anatomy of and ABS wheel sensor. - Most common wheel sensors are actually just a magnet with a coil wrapped around it. This sensor when placed close enough to a rotating sensor ring will distort the magnetic field of the sensor. This causes a signal to be generated but only when the wheel turns.

Second, the Arduino. - Having inputs tapped from the wheel sensor, the arduino looks for one input to go silent. (Stuck wheel, other wheel spinning wildly.) The Arduino then fires up the pump, activates the solenoid for the spinning wheel and clamps the brake until the other wheel rotates at a similar speed. (Usually within 10% tolerance.)

Tada. Basic traction control the way industry does it!
Advanced traction control is just more signal evaluation between the all four wheels plus throttle reduction, ABS is just traction control applied when the brake lights are on.

Hope it helps.

[dewy721]

Older, school buses have them, the ones that have a "DT466" emblem on the hood cowl, I'm going to hazard a guess of around 1987~1992.  It was an option for medium duty dump trucks and moving vans, but it was required by us law to have ABS on school buses by then. So buses would be the best salvage target of opportunity.

Where to find em'? Seek out the local school bus company, they usually decommission then sell off all the tired buses after 250k~350k miles. If they have a pump installed, a simple stomp on the brake pedal should reveal a muted whirring noise from under the hood. (Can't remember, but I don't even think you need a key in the ignition to test it.)

As far as reversing ABS as into traction control, if I read correctly you said you have independent front and a single rear hydraulic circuit ABS. Known as 4s3m (4-sensors, 3-modulator valves). You could traction control your front wheels (if 4x4), and use the rear circuit for traction balance between front and back.

*Wait* you 'should' be able to forget the pump. Your factory ABS should have one internally. Without it your foot would hit the floor during an "ABS event" and dump all pressure and thus fail to stop at all. But the school buses should have the same ABS wheel sensor wiring connectors, even ABS extension cables along the frame to get all the way back to the rear wheels. Use these industry standard connectors to plug your ABS/Traction hack circuit into the factory harness without cutting any wiring. They're called an "ABS wheel sensor extension cable" and are also available from Bendix. You'll need one per wheel, each one should have both plug genders on them.

Activating the system when you "want" it to come on: I suppose you 'could'  diode isolate the brake light input wire to the ABS controller and route the wheel sensors through an Arduino and then whenever your brake lights are off the Arduino cross-wires the wheel sensors and fakes the brake light signal, and only when in 4-Low. That way the system pulls double duty, using only electronics and only when your not already using the brakes to stop with.

[elsp1991]

ok i found even more people searching for that and it seems that someone do that even better. read this
http://forums.lr4x4.com/index.php?showtopic=69398

[dewy721]

In that post they mentioned the brand name "Wabco", the ABS training I got (for semi-trucks) was from Meritor/Wabco. Their docs are open to the public. But yeah, just swapping sensor inputs and faking a brake signal will do the job, provided the axle in question has independent solenoids.

[robotteck]

That is a great link, Thanks for all the input. I hope to make more head way in the ABS solenoid area, because that is the biggest mystery for me. A friend of mine mentioned I could mimic the wheel speed sensor signal to provoke an ABS event. Thinking about that... I could data log the event to kind of piece together how this thing works.