Speed sensor loosing signal

[TobyHarris]

Hi,

In order to make use of traction control facility on my ECU I have installed two hall effect speed sensors, one on the rear undriven wheel and one for the front driven wheel.

The problem I am having is that the ECU (Motec M4) looses the signal from the front wheel speed sensor consistantly at 20 mph. I placed an oscilliscope on the output of the sensor and can see that the signal is not being lost, though its not the square shape I was expecting.

Several examples of the signal taken from 'scope can be found
HERE and HERE

details of the sensor I am using can be found HERE

Does the orientation of the sensor make a difference ? it currently lies at 90 degree to the drawing.

The sensor ring / cog looks like THIS though it has been cleaned up on a lathe. Dimensions exceed those indicated in the drawing above.

As far as we can tell there is no run out between sensor and the attached cog. Distance between cog and sensor is in the region of 0.5mm.

Any ideas would be gratefully received
Thanks

 

[Comcokid]

Although the signal is not square, this is probably not a problem. I an not familiar with the Motec M4, but here are the questions you should try to investigate.

What signal conditioning is on the ECU input for the Front wheel sensor? The rear wheel sensor? In other words, does the ECU input square up the signal properly, and does it have some hysterisis/Schmitt trigger input characteristics (so it doesn't multiple-pulse on the slower rising edge). There must be some contitioning as the signal you show has a 7 volt amplitude.

Why does the ECU loose the front wheel sensor and not the rear? (I assume both use the same kind of sensor and tooth).

Does the ECU input drive a interrupt or is it a sampled input? (Can the ECU miss an input because it only samples the input). Is the input edge or level triggered. If interrupt driven, how fast can it handle interrupts? (The third possibility other than interrupt or sampled input is that it goes to a counter input which the processor regurarly reads to determine speed.)

 

[carnage1]

Yes orientation matters. straight up and down you have a much more instintanious edge. the sensing area is a three dimensional cone extneding from the end of the sensor tip. as the tooth aproches this cone you slowly ramp up and once it passes you instantly drop. with the cone pointed to the center of the cog it sees nothing and then suddenly there is a tooth under it then suddenly it is gone. match the orientation and everything should get happy fast.

 

[VetNutJim]

Most Microprocessors use a Schmitt Trigger on the inputs that are connected to 'real world' sensors.
Therefore the signal from the sensor does not have to be shaped as it will be squared by the Schmitt trigger on the input itself.
If you are losing the signal from the sensor perhaps the wiring or plug connections from the sensor are defective or have some fault.

 

[Fabrico]

I would be careful "cleaning up" the cog wheel on a lathe. This could easily round off the signal. I would compare wave patterns or swap wires with a working corner to check for this.

Is the ECU set for the same tooth count on all four corners?