Speed sensors / aftermarket traction control

[Shifty]

Hey guys

I'm setting up a traction control system. In short the system compares the speed at the 'powered' end of the car with the 'rolling' end. As such the traction control ECU requires a square-wave speed sensor on each axle.

As I understand my AE82 should have a factory speed sensor triggered by the speedo. Was hoping someone could please confirm:

1. that this is the case (pretty sure as I have a 4AGE pinout diagram showing a 'SPD' input)
2. what colour the wire is likely to be on my AE82 Twin Cam, and where I can trace it from without pulling the whole dash apart (as I only have a pinout diagram & the plug is gone I have no way to trace it without stripping the loom out from behind the dash).
3. that it is definitely square wave (I assume so - assume it triggers a voltage once per revolution).

The other thing I was wanting to get is advice on how to set up a speed sensor for the rear axle. Again, this needs to be square-wave signal, not a sine-wave... but it can be triggered by the disc, hub, axle, whatever.

Thinking about the rear of the car there is basically no rear axle, just a rear hub, so I assume I'll need to trigger off the disc or wheel somehow. Any recommendations on this would be great as without a second speed sensor the show doesn't even start

[CrUZida]

If the speedo is like the MX73 you'll see 2 screw near the centre at the very rear (where the speedo cable plugs in). One of these will go to earth, the other will go the SPD output.

The SPD output is 4 pulses to earth per rev of the speedo cable.

I don't know how compatible the AE82 is with newer corolla's but if you can find a rear hub with ABS thats your best option. Cast into the hub will be teeth that the sensor will pick up on.

[Hiro]

The other thing I was wanting to get is advice on how to set up a speed sensor for the rear axle. Again, this needs to be square-wave signal, not a sine-wave... but it can be triggered by the disc, hub, axle, whatever.

Thinking about the rear of the car there is basically no rear axle, just a rear hub, so I assume I'll need to trigger off the disc or wheel somehow. Any recommendations on this would be great as without a second speed sensor the show doesn't even start

When I was involved with the FSAE car, we designed a traction control system using a "chopper wheel" (basically just a toothed disc) and a hall-effect sensor pointing at the cutouts (from memory it was a Honeywell GT1), this gave a square-wave output every time one of the teeth past the sensor - this got fed into a microcontroller which then determined road speed and compared it to the indicated speed (ie gearbox output), if the difference was more than a set percentage (indicating loss of traction) the microcontroller output a signal to the ECU's flat-shift function which cut ignition to reduce power and restore traction.

[-GT-]

Yeah Hiro that's the plan, obviously you don't need the microcontroller, just feed the sensor output straight to the ECU. The issue is installing the wheel and sensor to a closed rear hub on a fwd - not really an axle accessable to work with.

I've looked into this a little, even have some AE111 ABS rear hubs to play with - too different to try and install them in an AE92/3, and the ABS wheel is off the back of the bearing, right in the centre of the hub with a chamber machined in and the sensors in the hub too. Can remove the sensors and the wheel, but not really straightforward to machine my existing hubs to accept them. Is on hold at the moment.

Edit: There's some pics in my ride thread from when i did my rear hub/brakes conversion if you need a better idea of what i'm trying to describe.

Edit2: Here's a couple which show what i mean with the bearing - don't have any of the hub setup uploaded at the moment. You can imagine how that sits with the wheel in the guts of the hub - i agree the only real options are off the brake disc, or maybe the wheel;

[Hiro]

Yeah Hiro that's the plan, obviously you don't need the microcontroller, just feed the sensor output straight to the ECU. The issue is installing the wheel and sensor to a closed rear hub on a fwd - not really an axle accessable to work with.

Our sensors were only on the front wheels of the FSAE car (as they were deemed to always be going at true road speed (with an over-ride for when the brakes are locked up)), they ran a closed hub (no axle actually), the chopper wheel slid over the hub/axle (I called it the CHAXLE, was one of the key points of my design) and was locked down by a BFFO locknut, and the sensor bolted onto the upright and looked axially through the chopper wheel.

[-GT-]

Not so easy to install a wheel over the hub in this case (Shifty's AE82 will be practically identical to mine above), and there's no axle or shaft to speak of. Brake disc hat sits up against the hub face (to the bottom in the above pics) but the hat extends the disc back up over the hub so the only real option is to install a wheel or something around the middle of the hub, which ends up inside the disc - not a heap of room but possible?

Solid disc too so can't read the vanes on the outer edge.

[Hiro]

Not so easy to install a wheel over the hub in this case (Shifty's AE82 will be practically identical to mine above), and there's no axle or shaft to speak of. Brake disc hat sits up against the hub face (to the bottom in the above pics) but the hat extends the disc back up over the hub so the only real option is to install a wheel or something around the middle of the hub, which ends up inside the disc - not a heap of room but possible?

Solid disc too so can't read the vanes on the outer edge.

I've seen other FSAE ones which read the bolt-heads on the brake hat too

But yeah, I know the hassles of trying to get something in place, we actually had the wheel mounted inboard of the strut (the chaxle went all the way through the upright, and the locknut was used to lock the whole thing together as well as hold the chopper wheel on, plenty more room that way.

Haven't stuck my head under a Mac-strut rear-end for a while (last time would have been when I changed the pads on the 102), but does anything project through the strut at all facilitating an inboard mount?

[-GT-]

Nope, nothing. I started eyeing off the heads of the wheel studs but if you're getting in under the disc you might as well just install something (balanced) on the hub.

Next complication: I regularly lift inside rears off the ground on the track, and stop them completely, which complicates reading even an averaged value across the rears.

I can maybe setup the traction control to not work above a road speed below which i'm not going to lift wheels (and tbh not really need traction control), but that kinda defeats the purpose of the learning exercise. And if i maybe get around to having enough power at some point to actually need traction control i'm back at square one.

I also think the adaptronic has a lower speed limit for traction control (not upper), and an upper for launch control, to separate the two functions.

Sorry to hijack so hard Shifty, just figure it's all relevant to your thread.

[big_zop]

Wouldnt you only be lifting wheels on decelerations so wouldnt really be needing traction control? Maybe you can get some logic in regards to reading both inputs where you take the average and ignore any values below a certain value or at the very least any zero values.

Any thought on machining slots into the outer edge of the disc and mounting the sensor on the dust shield. Shouldnt need to be too deep or expensive and you can get away with one or two depending on resolution required and balancing the disc - and add more as more resolution is needed.

Other ideas put forward by our FSAE team has been to use a fixed part of the wheel itself, like a wheel stud or nut, or fix a small magnet to the wheel. I never got to see them go through with this, but would have been much easier with the small diameter wheels over quite large hubs.

Edit: on the traction and launch control wouldnt it be safe to assume that launch would only activate for an event starting from zero - that way once a condition broke that event (say any deceleration, brake application) then it would go into traction control mode until the launch pre-recs were met. My understanding with the way launch control was set up on the year i was in the FSAE team (on a M800 - turbo CBR600 F4) was that you let the rear wheels reach a certain road speed before you cut ignition to hold it there rather than cutting it completely or completely elimitating wheelspin - so it allowed a certain slip percentage of the rear wheels over the front which gave more boost by preventing it bogging. All the actual values needed to be worked out in testing so while this may suit our car (and a MR drive at that), may not suit a FWD one or even one without turbo, etc. Not sure how much this made a difference to our car or if it was even run as there was limited testing.

[Hiro]

Nope, nothing. I started eyeing off the heads of the wheel studs but if you're getting in under the disc you might as well just install something (balanced) on the hub.

Next complication: I regularly lift inside rears off the ground on the track, and stop them completely, which complicates reading even an averaged value across the rears.

I can maybe setup the traction control to not work above a road speed below which i'm not going to lift wheels (and tbh not really need traction control), but that kinda defeats the purpose of the learning exercise. And if i maybe get around to having enough power at some point to actually need traction control i'm back at square one.

I also think the adaptronic has a lower speed limit for traction control (not upper), and an upper for launch control, to separate the two functions.

Sorry to hijack so hard Shifty, just figure it's all relevant to your thread.

Wouldnt you only be lifting wheels on decelerations so wouldnt really be needing traction control? Maybe you can get some logic in regards to reading both inputs where you take the average and ignore any values below a certain value or at the very least any zero values.

Any thought on machining slots into the outer edge of the disc and mounting the sensor on the dust shield. Shouldnt need to be too deep or expensive and you can get away with one or two depending on resolution required and balancing the disc - and add more as more resolution is needed.

Other ideas put forward by our FSAE team has been to use a fixed part of the wheel itself, like a wheel stud or nut, or fix a small magnet to the wheel. I never got to see them go through with this, but would have been much easier with the small diameter wheels over quite large hubs.

Edit: on the traction and launch control wouldnt it be safe to assume that launch would only activate for an event starting from zero - that way once a condition broke that event (say any deceleration, brake application) then it would go into traction control mode until the launch pre-recs were met. My understanding with the way launch control was set up on the year i was in the FSAE team (on a M800 - turbo CBR600 F4) was that you let the rear wheels reach a certain road speed before you cut ignition to hold it there rather than cutting it completely or completely elimitating wheelspin - so it allowed a certain slip percentage of the rear wheels over the front which gave more boost by preventing it bogging. All the actual values needed to be worked out in testing so while this may suit our car (and a MR drive at that), may not suit a FWD one or even one without turbo, etc. Not sure how much this made a difference to our car or if it was even run as there was limited testing.

We had launch control and would have used the same system for traction control, but it never got fully implemented whilst I was there. Basically just a case of having two condition-based programs - one for launch and one for traction (automatically selected above a certain speed). Traction control program would also have automatically disabled when the brakes were applied (just use a brake light switch as an override) to stop it from happening under lockup


I'll try and find my photos/models of the system we had (damn, shouldn't have deleted those Pro/E models the other day )

[Viper]

Im looking at doing something 'simular', probably, except with ABS and TCS.

Best wheels to get might be from a Falcon, BA/BF, etc.
On the rear end, they use a simple pressed steel or alu. 'toothed' wheel which then can be hooked upto a basic hall effect sensor. These can easily be made to fit behind hubs or discs.
You could then also go another step, and use the BA/BF sensor which is a Bosch DF11 sensor. They are abit more advanced then a basic hall effect sensor and give a perfect 'rectangle' output, plus its all there anyway

[petergoudie]

Don't worry so much about finding a device that will give a square wave from the begining. Do a google search with something like 'convert sine wave to square wave'. There should be plenty of electronic devices (integrated circuits or similar) that can do this as all they do is clip the signal or conduct at a certain height up the sine wave. Their output converts the input to square and cleans up the signal in the process.

[DX20VT]

We have to sort this quite often at work fitting a Halda/Terratrip to the local road surveyer's vehicles, ( that they nchange every 2 years)

Some of the modern cars you just cannot get a speed signal out of, Mazda's especially.

Anyway, ABS sensors will not give a square wave, and a signal converter is just another point of innaccuracy.

We usually fit a Proximity sw or sender to the hub and sense off a wheel stud.
sometimes you may even have to extend the stud out the back by adding to it.
Works well once you get it all in there, is accurate and reliable.

As for your problem of lifting wheels,
program it to read the highest speed of the two non driven wheels, as you wont be lifting both at the same time.

Interested to see how you get on.

Matt

[hemi twofifteen turbo]

an easy way to turn a sine into a square is to use a bosch ignition module, not sure what part number it is but the HEI module as used on Holdens (VC?VH-VK) and XE/XF Fords of Magna's TR? and earlier. and a whole heap of others.