source of noise on sensor signals

injb · Post by **injb** » Mon May 11, 2020 11:39 pm

I'm trying to track down the source of this noise in various sensor signals on my car. It's a 1989 Porsche 944 Turbo. The red trace in the screenshots is the throttle position sensor. It's working fine, but when the engine is running, there are weird pulses of high frequency noise that show up once per revolution. They get worse as load increases (the pulses get bigger and longer). I can see similar pulses of noise on other sensors. The green one is the airflow meter.

The blue trace is an ignition trigger pulse, and you can clearly see that the noise happens every second time, so once per rev.

Sometimes I don't get this noise at all. Any ideas on where I should look to track it down? What happens once per revolution that could cause this? Thanks!

injb · Post by **injb** » Thu May 14, 2020 1:37 pm

Anyone? I'll accept wild guesses at this point

Here's a few more screenshots - red is TPS and green is airflow meter (vane type). In one of them, the AFM signal is fine, a little noisy but not too bad. In the other one, it's all over the place - the noise is huge! Is it in any way possible/plausible that this noise is not really in the AFM signal and my scope/probe is just picking it up from something else in the engine? The scope, laptop and everything being probed were all inside the cabin - none of this was in the engine bay.

PicoKev · Post by **PicoKev** » Thu May 14, 2020 4:54 pm

On the wild guess front.......

Use of a non automotive scope.... the difference is not just the cost!
Use of unshielded leads

Battery charger on car
Failing components in the alternator
Failing ignition components
Crappy USB connection on the PC
Using a non Pico Blue USB Lead
Crappy power supply in PC
Using a laptop on a mains charger
Fluorescent lighting
Crappy LED underhood lighting

The list is long. And the above is only some of the possibilities!

There are several threads on this subject (The forum has a perfectly good search function).

This so called noise is all part and parcel of the underbonnet environment and you have to learn what to ignore and what to take notice of, it is all part of the learning curve and it is not without cost.

Kev.

injb · Post by **injb** » Thu May 14, 2020 5:32 pm

Thanks for the reply.

The usb lead and probes are the ones that came with the picoscope. The laptop was not plugged in. There was no battery charger on the car (just the alternator).

There's no fluorescent lighting or LED lighting. It was done outdoors in daylight. The hood was down. Also, the noises are obviously linked to the rotation of the engine, since they occur once per revolution at exactly the same angle every time.

I appreciate you saying "This so called noise is all part and parcel of the underbonnet environment and you have to learn what to ignore and what to take notice of, it is all part of the learning curve and it is not without cost", but how do I learn that?

Also, with the search suggestion: if that would help, then believe me I wouldn't be taking the time to grab all these screenshots and post lengthy questions. Do you have a suggestion of what I should search for to find more info on this problem?

Thanks!
John

Technician · Post by **Technician** » Thu May 14, 2020 7:21 pm

Do you need the engine running to test the TPS and Vane Air Flow meter?

With the ignition on manually test the sensors and see if noise is still present.

injb · Post by **injb** » Thu May 14, 2020 7:49 pm

Technician wrote: ↑
Thu May 14, 2020 7:21 pm
Do you need the engine running to test the TPS and Vane Air Flow meter?

With the ignition on manually test the sensors and see if noise is still present.

The sensors seem to work fine with the engine off - no noise. But I'm interested in what happens when the engine is running. If the sensors are fine but something is causing interference with the sensor signal and causing the computer to read incorrect values, I want to find out, and fix it.

The engine runs ok, but not perfectly. It's a little weak and hesitant on part throttle under load/low boost, and the AFR readings I'm getting from my WBO2 suggest that it's a little lean under those conditions. It's fine at idle and at wide open throttle, but between light cruising and WOT, I don't see it getting any richer than the mid/low 14s. Since I know the air flow meter is mainly used in part throttle on this car, I'm suspicious of it.

I've tested, cleaned, replaced and ruled out many things trying to track this down, including multiple ECUs and air flow meters, which I have as spares anyway. The car runs the same with all of them. Occasionally it'll fix itself without warning and run brilliantly for a short while. In fact, until it did that the first time, I didn't realize there was anything wrong. But there's no discernible pattern to it. So now I'm looking at raw signals in desperation after exhausting all the usual troubleshooting steps.

Maybe the noise I'm seeing here has something to with it, and maybe not. But I'd like to know what the cause of this noise is now regardless. What happens once every revolution at approximately TDC in a 4 cylinder engine that could be the source of this? It makes it very difficult to troubleshoot anything. For instance I wanted to graph my WBO2 output on the scope, but it's so noisy as to be basically unreadable. I showed screenshots to the WBO2 manufacturer and they said they had no idea, and I should send it to them for repair. But it's obvious that it's not the issue when the noise is present in everything.

I want to add: for those who say I'm using the wrong scope, I understand, and I know you're right. I might buy the automotive version at some point. For now though, I just want to see what progress I can make with this one. I'm fine with the risks of overvoltage etc. With that said, if you genuinely think that using a 2xxx series scope is the reason I'm seeing this noise, please tell me. Otherwise, just imagine I'm using the automotive version, and still getting these results - what's your advice in that case?

Thanks again for the input!

vasek · Post by **vasek** » Thu May 14, 2020 9:23 pm

Hi.
The noise that is present in the signals does not affect the problem with the engine operation in your example.
Such noises may be caused, as indicated above, the operation of the generator, drives the coils and injectors, fuel pump, electric motors, etc.
To more precisely, you need the file itself, not just images of fragments.
If there is a file, you can measure the noise frequency and relate it to the frequency of electronic components in the control engine.
In addition, the presence or absence of this type of noise depends largely on where the oscilloscope is connected to the ground.
.
The ECU does not receive such interference to the signal, since it uses the ground of the signal from the sensors.
You can also change the connection point of the oscilloscope to the hardware ground.
The signal will immediately become much cleaner.
You will calm down, and immediately start looking for the problem in another place (spark generation system, fuel supply, air boost stability, and so on).

.
If you connect the ground to the negative terminal of a battery, the noise can be very strong in amplitude.
If you connect the ground of the oscilloscope to the ground of the sensors, the signal will be noise-free.
For example,compare two signals that are taken from the same engine,but the ground is connected at different points.

injb · Post by **injb** » Thu May 14, 2020 9:42 pm

vasek wrote: ↑
Thu May 14, 2020 9:23 pm
Hi.
The noise that is present in the signals does not affect the problem with the engine operation in your example.
Such noises may be caused, as indicated above, the operation of the generator, drives the coils and injectors, fuel pump, electric motors, etc.
To more precisely, you need the file itself, not just images of fragments.
If there is a file, you can measure the noise frequency and relate it to the frequency of electronic components in the control engine.
In addition, the presence or absence of this type of noise depends largely on where the oscilloscope is connected to the mass.
.
The ECU does not receive such interference to the signal, since it uses the mass of the signal from the sensors.
You can also change the connection point of the oscilloscope to the hardware mass.
The signal will immediately become much cleaner.
You will calm down, and immediately start looking for the problem in another place (spark generation system, fuel supply, air boost stability, and so on).
.
If you connect the mass to the negative terminal of a battery, the noise can be very strong in amplitude.
If you connect the mass of the oscilloscope to the mass of the sensors, the signal will be noise-free.
For example,compare two signals that are taken from the same engine,but the mass is connected at different points.

Thanks. I'm not sure what you mean by "mass" though. Could you clarify that?

These traces were captured by backprobing the ECU connector inside the passenger compartment of the car. If there's a place I can probe the sensor that will give me a cleaner signal, I'll definitely try that, but where?

I don't believe it's fuel delivery or ignition causing my problem because then I'd expect it to be worse at wide open throttle, not better.

vasek · Post by **vasek** » Thu May 14, 2020 9:52 pm

injb wrote: ↑
Thu May 14, 2020 9:42 pm
I'm not sure what you mean by "mass" though. Could you clarify that?

Mass - ground (machine translation ).

Already corrected it.

injb wrote: ↑
Thu May 14, 2020 9:42 pm
These traces were captured by backprobing the ECU connector inside the passenger compartment of the car. If there's a place I can probe the sensor that will give me a cleaner signal, I'll definitely try that, but where?

Connect the oscilloscope ground to ECU sensor ground.

injb wrote: ↑
Thu May 14, 2020 9:42 pm
I don't believe it's fuel delivery or ignition causing my problem because then I'd expect it to be worse at wide open throttle, not better.

If you are sure that spark formation is all right, you need to look for the reason for sporadic lean mixture at a partial load.
And this - is not noise on the sensors signal.

injb · Post by **injb** » Thu May 14, 2020 11:14 pm

vasek wrote: ↑
Thu May 14, 2020 9:52 pm

injb wrote: ↑
Thu May 14, 2020 9:42 pm
I'm not sure what you mean by "mass" though. Could you clarify that?
Mass - ground (machine translation ).
Already corrected it.
injb wrote: ↑
Thu May 14, 2020 9:42 pm
These traces were captured by backprobing the ECU connector inside the passenger compartment of the car. If there's a place I can probe the sensor that will give me a cleaner signal, I'll definitely try that, but where?
Connect the oscilloscope ground to ECU sensor ground.
injb wrote: ↑
Thu May 14, 2020 9:42 pm
I don't believe it's fuel delivery or ignition causing my problem because then I'd expect it to be worse at wide open throttle, not better.
If you are sure that spark formation is all right, you need to look for the reason for sporadic lean mixture at a partial load.
And this - is not noise on the sensors signal.

OK maybe this is a stupid question, but I don't think I really understand what "sensor ground" means. This ECU has one common ground reference for all sensors (for all sensors that have a ground; some don't). These ground pins are all connected together (I just checked with an ohmmeter on a spare ECU). When I connected my scope, I connected it to one of these pins. I didn't think it mattered which one, since there is nothing between them except a few centimetres of copper trace.

There's one more pin on the ECU marked as "ground" that's not continuous with the others. So I should be using that one as my ground reference, instead of the one that the sensor is actually connected to?

PicoScopes
& kits

For vehicle manufacturers

EV kits and
accessories

Noise, Vibration and Balancing

Milliohm tester

Pressure sensors

Current clamps

Breakout
leads/boxes

Accessories

Automotive Forum