|Vehicle details:||Tesla Roadster|
|Author:||Steve Smith / Rona Eriksen|
Upon driving the vehicle at normal road speeds the ABS and traction control (TC) warning lights illuminated. The symptoms were present at every drive, regardless of temperature and road conditions, and had been present for over a week. The warning lights lit upon ignition on and remained on until the ignition was turned off again.
The above symptom was most certainly evident along with additional drivability issues:
Given that this is a battery electric vehicle, the acceleration can only be described as phenomenal due to the electric motor producing maximum torque from zero rpm. Traction control is therefore essential for a successful launch of this vehicle from rest.
Perhaps not so obvious was the loss of regenerative braking, which is prohibited once any ABS and TC faults are present. For those not familiar with high-performing electric cars, such as the Tesla Roadster, regenerative braking is a part of the unique driving experience.
For example: When driving the vehicle hard with regenerative braking active (i.e. no fault), lifting off the accelerator momentarily prior to cornering provides sufficient regenerative braking to enable a successful passage through the corner without applying the hydraulic brakes. When driving the vehicle in traffic, lifting off the accelerator would allow you to virtually stop without any hydraulic brake intervention.
Finally, without regenerative braking, the vehicle suffered a drop of approximately 25% in range, leading to a need to charge the vehicle more frequently and limiting the customer to shorter journeys.
The customer interview confirmed a full service history with no accessories installed, and highlighted that the front ABS sensor on the left-hand side had been replaced recently. We verified the customer complaint and confirmed the vehicle’s ID and specification.
The basic inspection confirmed that the visible ABS connections were secure and that the wheel/tire combination had the correct specifications for the vehicle.
We also inspected tire wear, brand, orientation, condition and pressure and found them all to be in order. This is a fundamental inspection that is often overlooked when diagnosing ABS and TC faults.
The ABS controller monitors the rate of acceleration and deceleration of the wheel speed and signals to determine if brake pressure intervention is required. Wheel speed signals vary depending on tire circumference, and in scenarios where tire tread depths, styles, and manufacturer differ, there is potential for tire circumferences to vary by 15 mm or more. This can result in differing wheel speeds during straight-line driving. Such a variation in wheel speed signals will have an adverse effect on brake control intervention.
In a perfect world, all installed tires should be the same make and specification, and within the same construction date (stamped on the tire wall). In reality, this is near impossible with vehicles of a certain age. We can, however, physically measure tire circumference and tread depths to confirm that the deviation between tires is within the manufacturer’s specification and eliminate these variables from our diagnosis.
At this point, a vehicle scan would have been the order of the day but none of our available scan tools could communicate with this vehicle. Faced with this dilemma, we had to use our experience and knowledge to figure out how to measure the ABS sensors for the correct output.
Prior to the speed sensor inspection, all available service data confirmed that there were no outstanding known issues relevant to the customer’s symptoms or to the wheel speed sensors and circuits.
We referenced the description and operation data to gain in-depth knowledge of the component functions. This knowledge is imperative when diagnosing any system, and highlights the need for continued research and training.
Armed with this knowledge, we confirmed that the wheel speed sensors fitted to this vehicle fall under the active design type, presenting a measurement challenge to technicians as the output can vary from 5 to 15 mA. As a quick test, the AC coupling feature of PicoScope can be used to determine the presence of a signal from the ABS sensor, but beware; using this technique alone can lead to errors.
I want to divert away from this case study momentarily, and refer to a Mini Cooper with an ABS fault that tripped me during the diagnostic process. To cut a long story short, there was no speed signal from the front left wheel speed sensor. The sensor had been replaced with a reputable aftermarket component along with a wheel bearing containing a magnetic pickup ring. I used PicoScope to confirm that a speed signal was present (using AC coupling) at both the sensor and the ABS controller, however, there was still no road speed with relevant fault code for ABS sensor.
The waveform below was taken from the Mini’s front left wheel speed sensor. On the surface, everything looks fine. We can clearly see the speed signal formed in the power supply (AC and DC coupled) along with the current flow through the sensor via a x10 current multiplier. It was only when I compared the current flow through the front right wheel speed sensor that the alarm bells started to ring.
The waveform below is taken from the Mini’s front right wheel speed sensor:
Once again, everything looks fine on the surface, but a keen-eyed reader would have spotted that the current flow measurement from each sensor differs (I did not spot this initially).
By using the reference waveform and scaling features in PicoScope, I could overlay both the front left wheel speed sensor (magenta) and the front right wheel speed sensor (black) and see the immediate difference.
The horizontal blue line indicates the theoretical crossing point used by the ABS controller to calculate the frequency (approximately 6 mA). Each time the current level rises above 6 mA and then falls below 6 mA, the ABS controller measures the time taken between each of these events and calculates the frequency, providing the wheel speed. The signal in the magenta waveform never crosses the theoretical crossing point as the current never falls below 7.5 mA, and so the ABS controller simply cannot calculate the frequency (hence no speed signal recorded via the scan tool for the left front wheel and the generation of the applicable fault code).
My first thought here was to blame the measurement technique (which is good practice) as we must question our approach, connections, etc., before we can confirm that the anomaly is real. I confirmed the current clamp’s zero point and the number of windings of the break out lead (x10) through the clamp, so the anomaly proved to be both real and relevant.
The image below demonstrates the operating principle of a current multiplier, where 5 mA is fed through a coil of wire wound 10 times through the jaws of the current clamp.
The cure for the Mini was to replace the aftermarket speed sensor for an original speed sensor. The wheel speed values returned and the fault was cleared. Unfortunately, the vehicle was reassembled and handed over to the customer before I could return to the workshop for a final capture.
The message from the experience above is to use AC coupling with caution as the signal fluctuations and ripples are centered on approximately zero volts, regardless of the fundamental signal level (removing all bias voltages). This does not mean that AC coupling should not be used, in this instance, it serves as an initial and quick measurement to confirm the activity from the ABS sensor only. Here we can confirm that the power, ground, signal and pickup are functional.
However, to confirm the ABS sensor and circuit, it is important to measure the current flow while comparing it with the offending sensor signal level on a known good (LH front with RH front).
For more on the Mini ABS fix see: https://www.picoauto.com/support/topic19041.html
Back to the Tesla case study:
With the vehicle set to ready mode (equivalent to engine on), in Neutral, and raised off the ground with the wheels free, rotating each wheel in turn resulted in the TC light flashing on the instrument panel. We stumbled across this feature during the diagnostic process. We can only assume that this is the detection of wheel slip, or some form of confirmation feedback, to enable rapid location of offending wheel speed sensors. All wheels, except the right-hand rear wheel speed sensor, pulsed the TC warning light.
Our initial test was to confirm power, ground and signal from the ABS sensor.
By using Pico’s 30 A current clamp (TA234) with a current multiplier and the relevant Pico breakout leads (PQ070), we could measure the expected low-switching current.
By using the low pass filtering feature in PicoScope, we removed noise from the signal to improve the resolution of the measured switching current at the wheel speed sensor.
With the ABS sensor circuit interrupted, we also seized the opportunity to monitor the sensor ground return line and the power supply (AC and DC coupled).
The AC coupled measurement was originally chosen to provide improved vertical resolution of the DC voltage ripple (speed signal), but further analysis post-capture helped to provide additional information.
NB. The connection between the sensor (Channel A) is reserved for the 4225 and 4425 models of PicoScope which allow the ground level to float ± 30 V from battery ground. This connection method is fine to allow inspection of the ABS sensor power and ground (using a single channel) when the wheel is spun by hand (low frequency only).
The waveform below confirms that not only is there no output when the wheel is spun by hand but also that:
Possible causes based on symptoms and evaluations made above:
The action plan is based on initial evaluation, experience, and accessibility.
Based on the following, we can, with a degree of confidence, accept the integrity of the wheel speed sensor circuit to be normal:
Had the ABS control unit detected a short to ground (through excessive current flow), the scope would have detected a momentary 12 V supply (at ignition on) accompanied with a rush of current followed by 0 V as the control unit shut down to protect the system.
The minimal current flow via the sensor (approximately 9 mA) with power and ground present on Channel A suggests there is no short to ground.
A short to ground would minimize the characteristic noise captured on Channel B (AC coupled).
The grey area here, therefore, is the condition of the ABS pickup ring, as access is near impossible without removing the ABS sensor which is an integral part of the hub assembly. This, you might say, is a blessing in disguise as the purchase of a hub includes the ABS sensor and pickup ring and so there is no need to dig any deeper.
With the new hub assembly installed the waveform below confirms the repair.
The Math channel included above (black) divides the acquired values of Channel C by 10 to determine the correct current value flowing via the ABS wheel speed sensor.
Unfortunately, with no Tesla scan tool available to erase the ABS fault code, we had to do a number of road-tests before the system would accept the wheel speed signal to be back online. Once this state had been confirmed by the ABS controller, all warning lights were extinguished and TC and regenerative braking was restored.
Right Rear Wheel Hub Assembly (Inc. Wheel Speed Sensor & pickup ring)
I think it is safe to conclude that Pico current clamps remain essential accessories to any diagnostic toolset. When included with PicoScope hardware and software, the case study above demonstrates the further potential of what is often overlooked as a tool of choice during diagnosis. Without the relevant current clamp, how can we qualify MRE wheel speed sensor operation and circuit integrity?
By its nature, the current clamp has to be the most non-intrusive diagnostic tool that is capable of capturing events not seen when measuring voltage, while secretly confirming voltage and ground and that the component functions correctly based on Ohms Law.
You could argue that if the current flow is correct, the voltage, ground and component are functioning normally, saving you two precious scope channels for other measurements.
When a hub assembly comes complete with a wheel speed sensor and pick-up, how can we qualify the integrity of the ABS magnetic pole style pick-up ring with vehicles utilising typical stand-alone wheel speed sensors?
Assuming the pick-up ring is accessible (and they rarely are) then the magnetic field finders available via the bearing manufacturers are beautifully simple, yet effective.
In a scenario where the wheel speed sensor is functioning but the pick-up ring magnetic field is broken or reduced in field strength, a speed signal output of some form would be evident. However, the signal may be sporadic, deformed or sequentially non-uniform if damaged in a single location about its circumference (and hopefully accompanied with an implausible signal code via the ABS controller).
An alternative way to see this would be to use the Pico TA330 Keyless Entry Detector, which by its operational characteristics becomes excited about a magnetic field! With the wheel speed sensor removed, insert the TA330 Keyless Entry Detector into the sensor aperture so as to bring the sensor tip into close proximity to the pick-up ring.
Using the scaling and filtering feature of PicoScope, select a 50 mV input range, increase the scaling by a factor of 10, and activate the default Low Pass filter of 1 kHz. Now rotate the wheel assembly (paying close attention to hands and fingers in relation to the brake caliper and suspension components!).
The keyless entry detector will generate a sufficient sine wave in order to serve as an indication of the pick-up ring condition and orientation, where a new wheel bearing may have been installed. The waveform below demonstrates this exact technique:
Notice how the keyless entry detector behaves in exactly the same fashion as an inductive wheel speed sensor, responding to changes in the magnetic field about the tip of the sensor. The conclusion here is that we can confirm the pick-up ring to be installed correctly, with sufficient magnetic field strength and uniform in structure.
Channel B utilises an Optical sensor to determine one complete wheel revolution. The following case study highlights the use of the Pico Optical Sensor: https://www.picoauto.com/library/case-studies/subaru-with-incorrect-abs-operation
March 05 2018
Hello Rafael and thank you for the feedback.
You are correct, theoretical crossing point was exactly that (theoretical). The only basis we have for the 6 mA value was the fact that this value could not be reached by the LH wheel speed sensor
I guess you could say that back-to-back testing highlighted a difference between sensors and this is our theory. The value could have been slightly higher or lower and we wish we could find out from the designers etc. as knowing the true value would be a piece of gold from a diagnostic point of view.
Regarding your question asking what is the crossing point for the Tesla…unfortunately, we don’t know. The Tesla demonstrated no output from the speed sensor (unlike the Mini). Since, once the new wheel sensor was installed, all was well, we did not need to find the crossing point.
February 28 2018
Steve and Rona, thank you for a great post and detective work. I have a question… where/how did you find out that Mini uses 6ma as the cross reference point of the signal to measure frequency?. And, do you know what is Tesla’s reference point?