Engine speed fault

[Juan]

Hi everybody,
I'm having some issues with the reference speed sensor. It's a magnetic sensor with a 60-2 wheel.
As you can see in the image (I've also attached the .ld file), cycle lock is missed. There was some noise at high rpm before, and the fault started when reaching engine speed limit.



The M130 was configured using a Magneti Marelli magnetic sensor, but in this case I was using other brand that is a replacement for the Marelli. I have been seing some Rejected pulse at high rpm, that I think it could be solved increasing hysteresis, but honestly, haven't tried it yet.

Most of the times, the engines works well (eventhough we keep seeing Rejected pulse) but after some heats cycle lock is missed.
We have so many problems with speed sensors in this racing series that cars have 1 back up speed sensor (wired to the ECU through a switch, also switching the map).

I have made an Input capture and the signal looks OK.

I will apreciatte any advice in how to minimize signal failure.

[David Ferguson]

Can you tell us more about the engine? I'm guessing this is a partner firmware package since I've never seen the "Engine Speed Redundant Pin Diagnostic" channel in the GPx pkgs. What Engine Speed Reference mode are you using?
What is the source of the Engine Synchronization? In the log file, the Engine Synchronisation Min/Max voltage never changes...

Can you post an archive of your package? Since the error first happens at the rev-limiter. I'm going to guess that your engine has a very lightweight drive train and slows quickly with the limiter active. Perhaps the value of Engine Speed Reference Wide Pitch Threshold should be larger, and the value of Engine Speed Reference Narrow Pitch Threshold should be smaller.

You say you are using a magnetic sensor, what is the hysteresis at 7500 RPM -- I think it should be about 5-7V.

There could also just be an issue with your sensor, it's gap, or the runout of your trigger wheel. At 44.4s in your provided log file, it certainly looks like there is an issue with extra pulses, right after noise has been reported in Engine Speed Reference Diagnostic. I have had an issue were car-to-pit radio transmission interfered with wheelspeed sensors that looked like that. Could there be similar interference happening here? How is the Engine Speed Reference sensor shielded?

[Stephen Dean]

Hi Dave,

Correct, it's a custom Firmware for the Argentinian Touring Car series. This engine uses the Mitsubishi 4G63T Reference Mode and has Hall sensors for both Ref and Sync. The Debounce is longer than I would normally use for a Hall at 120us, but for the number of teeth on this pattern, it probably isn't an issue.

[Juan]

Hi Dave & Steve,
As Steve said, it's a custom package (GP Lite), but reference mode is Crankshaft Two Missing Two Stroke, set as a 3 cilinder engine and driving 3 double coils. No syncrinization sensor and magnetic sensor are use for Ref.
Engine are 6 cylinders engine, reving 9k and lkightweight drivetrain. Maybe Steve is referring to another car series here in Argentina where they use a 2.0L 4 cylinder engines.

At 7500rpm I have 2V, I will try using between 5-7V.

Gap is set between 0.6-0.8 and runout was checked. The sensor is wired with a shielded and meshed 2-wire.

I think that it´s a configuration issue, as when I use a Marelli sensor we hardly have a fault, but when using another brand, even it´s a replacement, we do have problems after a couple of heats.

[Sean]

Juan, based on what you've said, this seems like a faulty or poor quality sensor replacement to me. If it's a true replacement sensor, why would the configuration suddenly be way wrong? Have you put a brand new Marelli sensor in to try it? What brand of replacement sensor have you been trying, is it a different oem high quality sensor (Bosch, Delphi, Honeywell etc), or an aftermarket copy?

The other thing not ecu related that I have seen cause this is worn terminals in the connector. With most oem connectors, the pins/terminals can fret in race car environments (which can also leave a dusty residue). Maybe the terminals in the connector are worn and the new sensor has pins of a slightly different size. The car I work on mostly, we have flyleads potted on to all the critical engine sensors to get away from the connector wearing problems.

Bear in mind I'm no Motec savant, but these things are what I'd go to first if I was at the track and had these problems.

[Juan]

Hi Sean,
I have tried Marelli sensors and they work much better (still have some sporadic faults), and the replacement that I have tried it is not a oem sensor, but an aftermarket replacement.
Could be possible that this aftermarket sensors have a slightly difference in the magnetic field or the coil, and oem ECU can withstand that difference? If this is the case, M1 should be able also, but configuration should be changed.

As you said, connectors and wiring was the first I looked at, but everything was fine.

Could heat affect the sensors? I have not measured engine bay temperature but surely it is higher than a normal car.

I am now using marelli sensors, but shouldn´t I be able to use any brand or model?
How would the setup process for a new sensor be?

[Sean]

I have tried Marelli sensors and they work much better (still have some sporadic faults), and the replacement that I have tried it is not a oem sensor, but an aftermarket replacement.

Juan, I'd recommend running the genuine Marelli sensors, and then work on eliminating the sporadic errors. Think about how much track time, or an engine costs, a reliable sensor is a must have for best results.

"Aftermarket replacement" sensor can be any quality level, from good to very poor. I've seen knock sensors sold with reputable knock monitoring kits that don't detect knock. Also seen Delco style map sensors that have given map traces that look like a seismograph trace. I honestly can't say I've seen an "aftermarket replacement" sensor that has given the same performance as a genuine quality brand name sensor.

[Juan]

Juan, I'd recommend running the genuine Marelli sensors, and then work on eliminating the sporadic errors. Think about how much track time, or an engine costs, a reliable sensor is a must have for best results.

Agree with you 100%, but the problems is that I am not being able to get them in Argentina.

So, I am thinking on using another sensor (high quality) that I can easily get here.

In order to do that, the steps for setting up a new sensor would be:
- sensor mounting: check gap and runout
- shielded wiring from the sensor to the ecu
- Input capture while cranking: higher voltage divided by 2 for the hysteresis. Define rising/falling edge
- Wide Pitch Threshold: approximately 170-180% for lightweight powertrain
- Narrow Pitch Threshold: approx 30-40% for lightweight powertrain
- Blank ratio: 10%
Speed pin
- Threshold: 0V
- Hysteresis: starting with the voltage defined in input capture, up to half of the max voltage seen at high rpm
- Debounce: ??

If you thing that this should be a new post just let me know and I will create it.

[David Ferguson]

I would try doing an input capture when running as well. If you can repeat the problem when free-reving in the shop, then I would try to get an input capture of the problem. Share those input captures with us here (not a screen shot, the actual file so we can zoom in, look around, etc.

[Juan]

I am attaching 3 Capture inputs at differents rpm.
Unfortunately, I could not capture while failing.
Next week I will be testing the cars in the dyno, so I will be able to take more of them.