wheelspin detecting boost controller?

[BAKES]

one thing i like to add there needs to be an arm or bypass switch so you can heat the tires up .

[90TurboVan]

^ But of course

Acceleration based would be difficult. What happens at part throttle or when traction doesn't allow peak acceleraton? You'd almost need a plot of acceleration versus TPS, boost, rpm for different surfaces. Maybe rate of change of acceleration? (Third derivative/jerk) If acceleration changes fairly quickly, the tires are probably spinning, but that might not be fast enough for real time control since you'd need a minimum of four speed sensor readings to calculate it.

A three sensor system (one in the back) would work just as well as a four. Turning could be compensated for with a steering angle sensor on the column, if its significant enough to be considered.

[streetvan1]

Most abs systems don't have the steering wheel angle sensor, unless they're equipped with stability control. If I were doing this, I'd hook up all 4 wheel speed sensors as mentioned above, and wire them to an abs system ECU from a vehicle that already has traction control.... whatever kind of car you want, as long as it has 4 wheel abs with traction control. This way, you would get a unit that would already be capable of reading the wheel speed sensor signals and interpreting them. It would also be programmed with parameters for allowable wheel speed differential, which would take care of the issue of differing wheel speeds around corners. This would save alot of time and effort. The other benefit to this is that most traction control systems have a disable switch already built in, so you could turn it off if you wanted to do a massive smokeshow or something awesome like that.

Then you could use the ECU's output the same way as a factory traction control, with slight differences. I would not use the hydraulic control unit (the part with the valves and whatnot), but only the ECU. The hydraulic control unit applies brakes at certain wheels to control wheelspin, but I would instead tap in to the output of the ECU which causes this action, and use that to actuate my boost controller. What sort of module was everyone else thinking of using?

Here's my issue with controlling boost from a stop: with an electric boost solenoid or the like, will it react fast enough to spool the turbo back up after recieving the OK signal from the system?? Or will this actually hurt performance?

[acannell]

the module everyone was thinking of is custom electronics and firmware designed by me or another EE type person on here

lets be clear here about our goals (or maybe just mine haha)..this is a drag only setup. that should make our easier.

the interesting thing about a system which compares change-in-rpms to vehicle change-in-velocity is that you could have a single accelerometer and its mounting does not have to be that accurate in 3 dimensions, because you dont care about the absolute accuracy of the acceleration, just its variation. this setup might actually be very robust and simple. and this system would not need wheel speed sensors at all. just vehicle acceleration and engine rpm.

and if the system is for drag racing only, and we keep the car straight on launch, then maybe we dont have to worry about how the system performs in turns or slight turns.

so to summarize, the system described above would be:

a single accelerometer mounted approximately straight ahead and in the center of the vehicle.

an rpm sensor (or just use the tach output on the coil)

heres how the electronics will determine wheel spin:

if there is 100% traction, the rate of change of vehicle velocity (acceleration) should be the same as the rate of change of engine rpm (crankshaft "acceleration"), since they are mechanically linked.

so the electronics just watch engine rpm and compute its rate of change, and watches vehicle acceleration and compares its rate of change. any difference indicates wheel spin.

the fun part is programming a control loop that takes the wheelspin and reduces boost appropriately.

boost will be controlled with the factory wastegate solenoid in the usual way (variable duty cycle PWM)

if you were deaf (couldnt HEAR wheelspin), this is how you would control wheelspin yourself. seat-of-the-pants acceleration versus engine speed.

using a digital scope, you could collect acceleration and engine speed data simultaneously and see what the situation really looks like.

if the system sees engine rpm decrease (you let your foot off the gas), it will also see vehicle acceleration decrease, so I think it should intrinsically handle off throttle/part throttle/jerk situations just fine.

you could add a bypass switch to disable the system for burnouts very easily.

thoughts?

[BAKES]

i just not sure that boost can be drop fast anough not to cause a lag time from when tire loose grip to were it is regained guess i can try and wire a waste gate sol with switch and pull a burn out and so how long it takes.

[speeduphoria]

Quote:

Originally Posted by acannell so to summarize, the system described above would be: a single accelerometer mounted approximately straight ahead and in the center of the vehicle. an rpm sensor (or just use the tach output on the coil) heres how the electronics will determine wheel spin: if there is 100% traction, the rate of change of vehicle velocity (acceleration) should be the same as the rate of change of engine rpm (crankshaft "acceleration"), since they are mechanically linked. so the electronics just watch engine rpm and compute its rate of change, and watches vehicle acceleration and compares its rate of change. any difference indicates wheel spin.

Well different gears will create different acceleration so tying that to rpm doesn't see like it will work well

and like others said a drag only setup will be easier to create and operate effectively then a TCS traction control type setup.

The next thing is that retarding the ignition timing is much faster then spooling and un spooling the turbo. So if you lost traction it would cut boost then have to built back up. Its basically a multi stage boost control that is automatic depending upon speed.

For simplicity using a speed based boost setup that is conservative would work well. Having rpm switches triggered from the speed sensor operate a separate solenoid and G valve works

[acannell]

I *think* that the acceleration based system will be immune to what gear the car is in, because its looking at *rate of change* and not the absolute numbers. So in 2nd the rate of change is smaller than in first, but as long as its the same rate of change as the engine rpm, there is no wheelspin.

*i think*

think about it

And even if it wouldnt work in multiple gears, the system could detect the sudden abscence of acceleration when you shift gears and interpret that as a gear change and use the next "gear map". Or it could be wired to the throttle position sensor and interpret zero-throttle as a gear change. Or a clutch pedal switch. Etc..

The turbo will not be spooled and unspooled. Boost will be actively *adjusted* real-time by modulating the wastegate solenoid. It will not be on-off.

Of course there are traction systems that use timing retard like you say. I just want to use boost control

I think this project should be classified as "possibly inferior to other methods of launch control, but possibly superior to all of them, and definitely fun to work on"

Im doing this for the sake of doing it, it may prove to be less effective, too complex, too expensive, etc... but thats not the point here...at the moment...for me.

[streetvan1]

You say that the calculations are the "fun" part, but the module I'm describing already has those calculations built in, using the differential of all 4 wheel speeds, which is the most accurate way of determining wheelspin. The programming you're describing will only work in first gear, and will be alot more difficult to set up. You'll need some sort of system to disable it once you shift gears.

The other issue I have with RPM vs Acceleration is that you'll have slippage, whether you're using a manual or automatic. With a manual, you'll be slipping the clutch to get a hard launch. In an automatic, your converter will not stall until a certain RPM. Now, with an automatic, you may be able to predict with some degree of accuracy just how much slip you will have, but with a manual, it will be different every time. How will your system determine whether the slippage occurs in the driveline or the wheels?

Also, you'll need to add both front wheel speed sensors anyway. The optical sensor you've described may work, but you'll need several lines per rotation. The abs sensors use several teeth, so that they can operate at low speeds, which is going to be the main time for this system to activate. Besides, who wants to fabricate a bunch of optical sensors, and paint lines on the axles carefully, and keep them clean, when you can just grab some spindles from the junkyard and modify the backing plates?

[acannell]

Like I said in the above description, the acceleration method can handle multiple gears by either detecting the sudden drop in acceleration during gearshifts, or by using a clutch switch or sensing the throttle being slammed shut with the TPS sensor.

I think the interaction between an acceleration system and a slipping clutch is something that needs to be tried out in order to see how it can be handled or if it is a problem at all.

The acceleration method requires no wheel speed sensors.

An optical sensor like i described could easily have hundreds of ticks on it per revolution, all you need to do is print them.

You dont paint the axles. You laser print the lines on stickers and then just wrap the axles with them.

Swapping in the axles may require making custom sensor mounting brackets. Or it may be easy.

I want to fabricate lots of optical sensors. And I like writing code and making circuits too.

The neat thing about the acceleration system is that it could conceivably be plug and play and require no mechanical parts. This would make it easy to sell as a kit.

[speeduphoria]

well I was just mentioning that its best to keep boost so it doesnt have to drop to regain traction. So that means either setting it up so its conservative or using timing retard instead.

once the wheels spin your going to have to drop boost quite a bit to regain traction(depends on how much wheel spin is considered "wheel spin" by the device, so overall how precise it is), then it will have to spool back up, so there will be some time loss if your talking about 1/4 mile times.

I dont have any concrete data but I can tell you that many racers prefer using timing to limit power on the starting line while keeping boost "up" instead of limiting boost. This mostly applies to hard core racers with Huge turbos that are laggy in general. For your application this is not as much of an issue.

Also I dont think the torque converter slip is even or linear as described, normally they "lock up" more on the top end and it depends on the converter and torque the engine produces.

Overall there is not going to be 1 setting that works great for all, just like custom cal's everyone's setup is a little different and requires different "stuff"

Like others have said I think its best to use the ABS parts to get the readings, its a bolt on affair if you find a doner car to get parts from.

[90TurboVan]

Don't make it any more complicated than it already is. You'll have a hard enough time getting a control loop to control boost implemented without playing around with complex detection methods. Three wheel speed sensors, and you're done. From there its a simple window compare depending on how much slip you want at the front to know if boost needs to go down or not. The goal is slip, not spin, meaning that the tires have not lost traction yet but are about to.

The problem with pulling timing is that you'd need an aftermarket box or a heavily modified ECU to do it. If you have the module "learn" a boost curve, you eliminate the overboost situations as it approaches the optimal boost for that traction situation.

And as I've said, optical = bad. They're great for something that won't be living underneath a car, but once you have to clean them off after every rain and the rings start to degrade, you'll wish you'd gone VR or Hall.

[streetvan1]

It would be cool to have a kit, and a purpose built unit, but an accelerometer or RPM unit that monitors TPS and clutch would simply be alot of time and effort for a much less accurate system. Even torque converter lockup is unpredictable. Wheel speed is the way to go. All the factory systems use wheel speed, and there are teams of people who develop those all day long for years at a time, so all the software has been developed and works the way it's intended to, and the sensors and modules are reliable. Modifying one of these boards for your use still requires alot of effort and skill, but would be alot easier than starting from scratch, and atempting to collect data under various conditions, etcetera. There will be enough R&D to do with the boost controller, plus you'll have to make some kind of pulse width modulation unit for the solenoid, if that's what you're planning on controlling it with.

For drag strip use, the optical system with 3 wheels might work, but on the street or for road racing or autocross, you'll want to monitor all 4 wheels and have the cornering differential programmed in, and you'd want to avoid the optical sensors unless you lived in the desert.

I think this is a neat idea, and I'd love to see some pictures and stuff of your unit in progress and when it's done. I'm not trying to rain on your parade, just trying to point possible issues, and suggest easier means of doing this.

[acannell]

no rain felt at all! this is what this thread is for!

lets be clear here, detecting gear shifts is not complicated. you add a clutch switch or look at the tps sensor. that is very easy.

does the 3 wheel system avoid clutch slippage issues? it does..and thats nice. I think that may make it worth pursuing over the accelerometer version, since it makes it more likely to WORK.

Anybody know exactly what the mechanical problems are with putting the tone ring axles, hubs and sensors on a 90 daytona would be?

[crazymadbastard]

I have a 92 gtc that came with abs I am sure the daytonas came with abs too. The sensors would be a bolt on with knuckles and axles allowing a direct swap. I will try to look at my rears and see if/how the sensors are installed. The fronts have a tone wheel built onto the axle and the sensor is mounted on the knuckle/bearing holder.

[acannell]

okay so basically this means that if you swap the knuckles too there are no sensor mounting issues in front, as far as we know. makes sense.crazymadbastard any info you can get on the rear sensor setup would be much appreciated!

this link is interesting..its about the abs system:

ABS antilock brake issues and repairs

and it says that bendix 10 came on:

90-93 C body (Dynasty/New Yorker)
90-93 Y body (Imperial)
91-92 BB body (Premier/Monaco - Renault designs)
91-93 S body (Minivans - Caravan/Voyager)