Picoscope to simulate automotive sensors for testing

Thread Starter

Brian595

Joined Feb 11, 2020
6
Hi there,
I saw some videos online of folks using sensor simulators to help with automative diagnostics.
I sometimes use a USB powered PC based picoscope oscilloscope to look at communications and the one I have has an onboard signal/waveform generator.
In tracing a suspected airflow sensor issue today, I connected the Picoscope in place of the sensor and switched on the signal generator to produce the required waveform of a good sensor.
But as soon as I connect it to the two wires feeding the sensor it crashes the control module supplying it.

Is what I am trying to do fundamentally flawed? Perhaps because the Picoscope is powered by my computer and this being a separate supply to the truck I was working on? Any ideas? I would be very useful if I could emulate signals using the picoscope when tracing sensor faults.

Brian
 

Thread Starter

Brian595

Joined Feb 11, 2020
6
Here is another example from the shaft speed sensor from a generator. (Which is working normally)
It is a two wire hall type sensor. The voltage between both wires is 9.0V, the voltage between wire A and earth is 9.0V.
The attached waveform is recorded between wire A & B with the machine running at normal speed.
If I disconnect this sensor and replicate the waveform with the signal generator function the control panel reads no signal at all!
EE779C9C-FC49-441A-9C05-CE2C9D8088EA.jpeg
 

Alec_t

Joined Sep 17, 2013
11,141
You appear to have the scope AC-coupled. What does the trace look like when DC-coupled?
A 2-wire sensor should not have a terminal directly grounded. Exactly how have you connected the scope (and its power supply) to the vehicle's circuitry?
The sensor draws current (several mA) which varies with magnetic field strength, so your sig gen needs to provide a varying current source rather than a voltage source.
 

bwilliams60

Joined Nov 18, 2012
1,283
Just to add to that, can you tell us year/make/model of unit you are testing and engine size? Pico has a large waveform library available so you can match up what should be with what is. Pictures would be great to show us hookup.
 

Thread Starter

Brian595

Joined Feb 11, 2020
6
Thanks for the feedback guys.
In the example where I have attached the photo;
The machine is a Caterpillar C9 large generator which uses a CANbus communication system between control modules not unlike a car. One of these modules monitors shaft speed at one of the drive couplers via the two wire sensor.

To view the waveform: I have spiked the Picoscope into the each of the two wires of the sensor and I can see the waveform as per the photo above. This looks correct the and varies when the shaft speed increases or decreases. The Picoscope must be set for AC. If o set for DC I don’t see any waveform at all. All good so far!

The question is about trying to simulate this sensor (or any other one) using the Picoscope onboard signal generator. It would be a very flexible diagnostic tool.
What I was doing (and clearly didn’t work) was to simply leave the scope connected to the sensor circuit as described previously, swop the BNC to the Picoscope waveform output terminal and switch on the signal generator from the PC with the sensors signature waveform already set up. As soon as do this the machines control unit shows and error and the CANBus needs to restart.

Regarding how the Picoscope is connected to the machines power supply, it’s not! And this may be the start of the issue. It is powered from my laptop USB.
 

Alec_t

Joined Sep 17, 2013
11,141
What is the specification (voltage range, impedance, drive current capability) of the waveform output terminal?
If the scope is AC-coupled, what is the value of the coupling capacitor?
What is the lowest AC frequency the scope can respond to?
 

Thread Starter

Brian595

Joined Feb 11, 2020
6
You say there is 9V between the wires, but the scope waveform shows only 80mV pp ??
When I hook a DMM in DC volts there it shows 9.0V between the wires (I would have done this test when the machine was powered up but shaft not running)
 

Thread Starter

Brian595

Joined Feb 11, 2020
6

bwilliams60

Joined Nov 18, 2012
1,283
I am confused. You say you have a 2-wire Hall Effect sensor? Most Caterpillar equipment uses 2 wire variable reluctance sensors which produce an AC signal, and 3 -wire Hall Effect sensors which produce digital signals. Nothing on CAT uses 9V to my knowledge. It is either 5V, 8V (TPS) or 12.6V . So I will ask again, can you send a picture of what you are tapping into and the screen shot from the Picoscope that correlates with it.
Also, what is the Prefix from the ESN? I can llok at CAT schematics and we can help you figure this out but need more info.
Lastly, what airflow sensor are you referring to on this engine? Is this a truck or dozer or what exactly are we working on. PLease be specific. We are trying to help.
 

Alec_t

Joined Sep 17, 2013
11,141
From the datasheet I see the sig gen has an output impedance of 600Ω. Perhaps that is incompatible with the control module circuit that reads the sensor signal.
 
Had a couple more thoughts... I'll admit I don't know generators, but I do know in truck engines all sensors that monitor engine functions are wired directly to the engine ECU. My guess is the sensor is probably more for the generator side than the engine side unless it sends a stop engine command when the shaft breaks.

It may also be possible that the sensor may be monitoring RPM messages from the engine, and trying to alter the waveform it would expect to see is what is causing your problems. It shouldn't kill the CAN bus I wouldn't think unless seeing a "distorted" signal was deemed bad enough by the programmers to do it. Would expect to see more of an error message than a reboot.

Just an example I guess... I have seen a truck before that the speedometer would work with the sensor disconnected, but it would not shift into high range (13 speed). My guess is since it was an automatic transmission and knew what gear it was in and what the engine RPM was it would calculate the speed from that. Unlike a car most over the road automatics are actually just reworked manual gearboxes with a servo driven clutch and shift rails so the transmission input speed should always been the engine RPM and a simple calculation would give you road speed... Unless you are at a stop of course.
 
Also thought in theory AC should work the same no matter how it's connected or what it is connected to, but in the context of a sensor it may not. Instead of both sides of the circuit switching poles with a signal generator you would only have one side + to - , - to +, + to -, and so forth and the other at a fixed point (I'm assuming a bit on the signal generator).
 

Thread Starter

Brian595

Joined Feb 11, 2020
6
Guys much appreciated the feedback. Sorry for the lack of specifics, they were jobs that I was just at briefly. To take a firm scenario, that I can refer back to let me take a new example;
VW Touureg 3.0TDI 2017.
Trouble code for front left wheel speed sensor. I had a replacement WSS, swooped it and it solved the problem. But I’m thinking along the same lines, if I didn’t have one could we simulate the sensor with the picoscope? A wheel speed sensor might not be the best example because it is has no output in the quiescent state, but if you could just proceed with the concept for now.

Not unlike the CAT machine, there is a DC voltage between the the two wires about 9V. This doesn’t seem to change when the wheel rotates and it stays steady on the scope. The response seems to be visible when the scope is set for AC, and you can clearly see a square wave when the wheel rotates. Similarly to the CAT machine if I remove the sensor, switch my BNC to the signal generator output, the car immediately shows faults in the instrument cluster for traction control, TPMS, headlight range control and ABS.

See photos of the setup below.

343A5D92-E50B-4750-A977-B343E6A0852C.jpeg4B7B82A3-94DB-46EC-820F-861BD4FF9FB8.jpeg2F9F469F-9444-4FF0-98DE-5E8091FA7E9A.jpegA5E03C27-B676-4193-B6F4-5514B86FAEC4.jpegADD7B6E4-1A84-4835-A2C0-4DDA771BD037.jpegAbove when wheel rotated by hand.
 
This is interesting and I wish you luck. I'm at my knowledge limit in this area, but like the concept. I kind of based some comments off the differences pin wise on say the LM1815 which was the heart of the original GM HEI, the MAX 9924 family that I have played with a little, and how Megasquirt's VR circuit is set up for the distributor pickup.

It may be the 9 Volts you are seeing is a bias voltage for the sensor. In other words instead of + to - swings and such it's actually 8.5 - 9.5 volts so to speak (obviously much less from your scope pictures) with 9 volts being the center.

My only other question is how are you physically connecting your scope? Are you connecting your probe to one wire and your probe ground to the other wire? Have you tried just one wire of the sensor to your probe and the other to the vehicle ground?
 

bwilliams60

Joined Nov 18, 2012
1,283
Still baffled at how younare getting square wave signal out of 2 wire sensor. They are variable reluctance sine wave generators. And have no idea how you ate getting 9v on a signal generator line???
 
You know this is a three-wire setup with the Picoscope? It must be grounded to the vehicle, in addition to the probe connections to the two sensor leads. If your Picoscope and laptop is powered by AC mains adapter, this can cause common-mode noise problems that give confusing readings, and the car's VR sensor amplifier will not tolerate it.
I think a two-wire ABS speed sensor is variable-reluctance. Delphi makes Porsche, Audi, Volkswagen, Skoda ABS speed sensors and they seem to be around 1,100 ohm coil two-wire.
Modern Bosch ABS wheel-speed sensors are Hall-effect and have an IC in the sensor head. I think they need to be three-wire.

To further figure out the drama, float the laptop/Picoscope by running it on batteries, ground it to the car, set it to DC coupling (not AC), and see what you get.
 
Top