M800 dbw noise

[BluByU]

I was curious if anyone else using DBW gets noise in the signal from the pedal (M800 using the dbw feature on the ecu)? I rewired my project car's throttle due to small blips that were showing up in one of the (Driver Request... pedal) inputs. I reran the wires to the pedal as well as the wires to the throttle body, using new shielded tefzel, brand new connectors on each end, brand new pedal and throttle body, verified the pinouts about 50 times. While the engine is off it works perfectly (has always been that way), but as soon as the engine is running I can see a blip of about 10% about 2-3 times a second. As you can imagine that really screws with the idle, and you can feel it while driving, and you can see it in the logs. TPD will blip, but TPD2 will stay where it should be (0% during idle), TP will blip to 10%, TP2 will follow TPD2 perfectly or the idle request perfectly.

[Holmz]

An oscilloscope would be helpful for this.

Can float the ground at the pedal?
If so you could run all the wires back as a twisted pairs... And ground them at the ECU end.

If the noise is not present with the engine off, then I would start by trying to find out what components are associated with the noise.
Alternator, coils, fuel pump, injectors, etc.
It is either:
1) Your 12V carrying noise on the line.
2) The ground carrying noise
3) Magnetic fields from other wires nearby inducing currents in the pedal wires. (twisted pairs help here)
4) It may be possible for a bad ground to have a ramp on the signal wire (differential pair), which then may break down. A good ground is important, but a bit of a concept as it is never absolute.

#1 and #2 get measured together and relative to each other. If you are not seeing 2-3 Hz or other spikes on the main voltage, then I would move to the individual wires and check all those relative to ground.
I think that you will need a scope.

[BluByU]

I'm running the gounds all the way back to the ecu's appropriate 0V. I really need to bite the bullet and pick up a pico-scope or something similar.

[MarkMc]

Unplug the ECU and using a multi meter check for continuity from the 0v wire in the loom to the chassis. 0v wires should ONLY be continuous with the chassis through the ECU. I have seen people use factory looms and not take notice of which "sensor grounds" are tied to the chassis directly somewhere and which ones only go back to the ECU. earlier Nissan looms are particularly bad for this. If you measure continuity and there is no ECU connected you will need to find what part of your 0v wiring is connected to chassis earth, it could even be the other 0v pin (B15, B16) as they are internally connected.

[BluByU]

So with the ECU disconnected?:
From 0V Eng on the pedal to chassis ground, should be no continuity
From 0V Aux on the pedal to chassis ground, should be no continuity
From 0V Eng on the pedal to 0V Aux on the pedal, should be no continuity, but when ecu is connected should be continuity

[MarkMc]

Correct. If you unplug the ECU and go to the sensor connector, engine temp, air temp, MAP, TPS, pedals or anything there should be no continuity to the chassis.

[BluByU]

Thank goodness I made good wiring diagrams. Tested from 0V Aux on the pedal and yep it was short to chassis. Checked at the 0V Aux distribution block I made, bingo found the culprit, the check engine light signal I was sending to my Aim dash. I'm betting that analog ground on the dash is sharing with something else like my diff temp sensor.

I'll have to give it a test tomorrow. Keeping my fingers crossed.

[BluByU]

Short of disconnecting all possible non-integral sensors, I still have the blip in the signal and I'm at a loss to why. If it works fine when the engine is not running, what could be injecting or interfering with the AV1 input? The wires to the ecu are flawless and shielded with no tight bends so I doubt they are getting RF of that magnitude but only intermittantly. High speed logging of the 5V and 8V busses looks pretty stable +-0.05V when running.

Any other ideas?

[BluByU]

Had a chance to log 3 different situations (Full, partial, and idle)
At idle, TPD and TP both blip about 10% while TPD2 and TP2 stay where they should be
While accelerating, it looks like some type of differential signal is happening
While at full throttle it looks like the reverse of at idle where TPD2 and TP2 are maxed and TPD and TP dip down.
So the partial/accel log appears to point out the problem in all situations, if you average the two signals it is fairly flat. Now I just need to figure out what would cause that situation.

[Holmz]

In the middle screen shot the "train of pulses" are ~0.8 seconds apart (or 1.25 Hz).
The individual pulses are ~0.1 seconds apart (10-Hz).

If you did an FFT of the throttle position you will see the exact frequencies and can see if they are the same for WOT, cruise and idle.
Then you have see what happens at 0.8 and 0.1 seconds (1.25 Hz and 10 Hz). I think that could help.

I would also 'up the logging rate' on that as there could be more happening at higher frequencies.

Does it do this with the ECU on but the engine not running?

Also you could probably also hook the 12v to an input and then a different ground to a high speed input and see if they are moving around.