M800 dbw noise

[BluByU]

Oddly it does not do it when the engine isn't running, smooth as butter with no blips. I borrowed a hardware scope from work in hopes that I'll be able to try it out this weekend.

[Scott@FP]

In standard wire wound or solid variable resistors (TPS in particular) this is the classic failure mode, at the positions where the sensor spends a lot of time, it wears out, and you see fast momentary spikes. Try swapping pedal assemblies.

[BluByU]

Just tried a new pedal and throttle body during lunch and it has the exact same behaviour. I upped the logging to 200hz and it literally looks like the %open on the pedal diverges from TPD and TPD2 every .3 to .8 seconds. So when the position of both TPD and TPD2 should be increasing in position % value they are not, one goes up and one goes down. Looking at their associated AV inputs when the blip occurs the voltages are moving in the same direction when they should move opposite. I verified that by creating a maths function to add AV1+AV5 for pedal, then AV3+AV4 for the throttle body, and during non blips where the original signals are correct and not flat the math shows a flat line, but during blips the voltages both go up and show a spike in the summed voltages. It doesn't appear to change based on battery voltage, rpm, any area of the actual pedal position, engine temp, manifold pressure, 0 or 2.5 or 3.3 or 5 or 7 or 8 or 9 volt references, cpu usage, nor is it a whole or fractional multiplier of the sync/ref, doesn't matter if it is moving or not. One side effect is that it occasionally trips the error state since the signals are too far from their expected and the ecu has to be reset, and I was able to catch one of those events. The engine still runs but only at 2500 rpms, and the blips were still there eventhough the ecu was not responding to my pedal inputs. As soon as you shut the engine off though, everything is as smooth as butter.

The one thing I haven't tried was to disconnect the O2 sensor to see if the "heater" circuit is somehow causing those blips.

[BluByU]

I just pulled some of my logs from 2007-2010, completely different wiring, different ecu position, different sensors, different firmware, same model but different pedal and throttle body, and I see the same blips but only about 20% the size of the current ones. It looks like the blips have grown over the years. What is the chance that this could be ECU related?

[Holmz]

I just pulled some of my logs from 2007-2010, completely different wiring, different ecu position, different sensors, different firmware, same model but different pedal and throttle body, and I see the same blips but only about 20% the size of the current ones. It looks like the blips have grown over the years. What is the chance that this could be ECU related?

Damn near zero - If the ECU was working with the engine off, then I cannot envision how it is not related to the engine running.

If the hardware scope shows the blips on the sensor wire then it can only be the ECU related if the ECU is feeding voltage to the DBW pedal. But for two different ECUs to do that is unlikely.

You could look at the voltage going to the pedal and see if there are spikes there.

It is either:
1) voltage spikes going to the pedal
2) voltage spikes at the pedal
3) voltage spikes coming from the pedal.

In the last case it is possible that the voltage is induced from a magnetic spike. This would be from a high current wire that is close to the pedal wires, and would be on a high impedance wire. I am assuming that the voltage at the pedal is low impedance which would be hard to induce a voltage in that wire (?).

Since it only does this while running I see a few options:
1) get the scope out and test the voltage like crazy (every wire you can think of, and particularly the ones close to the DBW wires).
2) Start measuring current like crazy (every wire you can think of, and particularly the ones close to the DBW wires).
3) Start pulling fuses from the extra gear that is running to see if you can find what makes the condition go away, when it is unpowered.

Or- What clicks on and off?
Does the fuel pump run at a constant current, or does its current pulse?

Is the return wire from the pedal a twisted pair (Twisted with a ground that is attached to something at both ends)?
If not I would twist it up and ground each end. A shielded wire may work for RF, but magnetic shielding is a different beast all together.

[Herrubermensch]

Ever get this resolved? What was the cause?