Setra 517 HD Clutch position sensor

[flaurits]

Hi.

I am working on this Setra 517HD bus. Its sets error codes regarding the position of the clutch. It measures via a hall sensor how far the release bearing moves and then calculates the travel length.

With KOEO good looking signal nice sqaure waves.
But when KOER it has alot of noise on the signal ground, and signal drops out, Therefore the travel position gets miss calculated by the tcm. Is my guess.
So far i have tried removing the belts for all alternators. And disabled injectors on at the time with scantool.

Clutch and release bearing is all replaced with new genuine parts. (Dosnt know if this is original problem or built in with new component)
How can i locate and error like this more accurate?
Scope is connected in all four in/outputs on sensor, and i am using the sensor ground for the sensor too all my leads, is this wrong? Should i connect all my lead grounds too battery instead?

Best regards Laurits.

[Rfmotors1]

Hi flaurits,
Here is my explanation:
First how such sensor/device works. (I may be wrong as I have never seen such machine)
If it is what I think, many such applications are in passenger cars. There is a PWM signal from the ECU which is desired position of the device, then the device carry out the task and reports back the actual position via another PWM signal. Now how such signal is mostly created.

The side that is receiving the PWM value is keeping the wire "high", or let’s say providing a bias voltage on the wire which is 5V and the sender device is pulling the voltage to "low" or we can say ground. (clutch bearing is receiving the desired position and ECU receiving actual position)

I think the Blue waveform is the actual clutch position PWM and therefore the ECU is providing the bias voltage, keeps the wire high, on 5 Volts. Then the clutch is pulling the voltage down to ground, to low in order to create the PWM signal.

There is drawing (attached pic) how the signal should be if all is OK and the only explanation is the bias voltage 5V is interrupted.
But not fully interrupted, there is sparking like pattern and the voltage does not go fully down to ground (red arrows). This is happening when there is bad connection between the backpined needle (assuming clutch bearing connector) and the ECU socket (correctly the internal ECU driver).
So I would watch the scope screen, set longer timebase and wiggle the wire and connectors between the clutch bearing and ECU. then I would expect the wiggle match the "sparking" pattern once I am on the affected area. More spikes when I play the poor contact location.
Also see this post: topic23308.html
Regards,
Roman

[ben.martins]

Hi flaurits,

I agree with Roman here. Given you power and ground to the sensor appear to be good apart from some slight interference, the main issue is on channel A where the signal is being really corrupted. If the ECU is looking for the output either duty or frequency, it has no way of calculating this due to the level of corruption in the rising and falling edges of channel A.

As Roman as advised, checking for poor connection is a must. In order to rule out any other issues, it might be worth measuring the same wires at the ECU and carrying out a wiggle test, depending on accessibility. Suggest also carrying out a pin drag test to make sure there is suitable contact between the sensor and the terminal in the connector.

Keep us posted with how you get on.

Kind regards

Ben

[flaurits]

Hi Ben AND ROMAN thank you for looking at My case.

I do understand your explanation. But this is not the case.

I AM sorry for the lack of information but the information about this system is vague. I only have information regarding the signals from the oe scan tool, yet still vague.
This release bearing have two sliding potentiometer it Would seem Theres some kind of electronic build inside making the pwm signals.
The release bearing is mowed by compressed air, no actuators in the bearing itself besides the position sensors.
I have to out going signals, a feed and ground.
When doing clutch calibration KOEO Theres no dropouts in the signal. Tapping, moving, "bending" dosnt effects the signal at All.

That's what makes we worry about the noise on the sensor ground.

Best regards Laurits
Sitting here thinking i cannot say if its the tcm providing the 5v Bias or the sensor.
My next move on monday is gonna check if its Bias from tcm by unplugging sensor.

[Rfmotors1]

I would try all what Ben advised and also unplug the sensor as you suggest, to see if the signals will both go high.

If the bearing is not an active actuator itself but a sensor only, and as you said there are two potentiometers, it sounds like a double signal sensor, as back up like design where the PWM signals are opposite phase. This could give a checksum like control to CM, sensor 1 plus sensor 2 must be 100%

That would be opposite PWM signal, and in that case both signals should go high when you unplug the sensor.

I would suggest you also capture entire cycle, as you said the system is able to run basic setting or clutch calibration.
To capture entire cycle, you may want to set timebase to 2 seconds per division (gives you 20 sec per screen) and sampling rate 500k.
Then capture the clutch calibration and post the file here.
Also with the same setting applied, another file when the system failing. (math channels will be possible to fully utilise then)
Regards,
Roman

[flaurits]

I Will look into on monday, thank you.

[flaurits]

I would try all what Ben advised and also unplug the sensor as you suggest, to see if the signals will both go high.

If the bearing is not an active actuator itself but a sensor only, and as you said there are two potentiometers, it sounds like a double signal sensor, as back up like design where the PWM signals are opposite phase. This could give a checksum like control to CM, sensor 1 plus sensor 2 must be 100%

That would be opposite PWM signal, and in that case both signals should go high when you unplug the sensor.

I would suggest you also capture entire cycle, as you said the system is able to run basic setting or clutch calibration.
To capture entire cycle, you may want to set timebase to 2 seconds per division (gives you 20 sec per screen) and sampling rate 500k.
Then capture the clutch calibration and post the file here.
Also with the same setting applied, another file when the system failing. (math channels will be possible to fully utilise then)
Regards,
Roman

Sorry Guys, had a fight with a VW transporter t6, with a bad nox sensor, clogged injector and empty adblue tank, so didnt get time to measure. What a pain.

Tried getting a wiring scheme but because of mercedes buying the business we useally get wiring for setra from. They Are not allowed to share the service info.

The mechanic working on has decided to try and swap the position sensor so he started removing the transmission again. (8 hour job)
I did check pin fitment, and had visual of the wiring while he was wrencing and checkes out good.

If it fixes it, ill still try and get a capture while performing the clutch relarn command.

Best regards.

[flaurits]

So, i never got a last capture.

The wiring from the release bearing is ziptied to the housing. It broke on test drive. Those dropouts is the sensor wiring shorting to each other.

Best regards
Laurits

[Rfmotors1]

Hi Laurits,
Thank you for updating us here, it helps to all of us with our scoping adventures.

Best Regards,
Roman