REF spike at high RPM

[xinlu]

Hi guys,

I’m currently working on the engine for our FSAE race car and I have run into a number of issues with getting it to run properly. Any help would be much appreciated. I have attached our current tune and one of the log files I have obtained to demonstrate the symptoms:

MotecLog.ld

Engine Log

(269.64 KiB) Downloaded 667 times

ECUTune.e35

Current Tune In Progress

(72 KiB) Downloaded 644 times

Here is our current setup:

- 2004 KTM 525 EXC single cylinder stock engine
- Custom Electronic Fuel Injection with 12-1 REF trigger wheel using stock magnetic REF sensor and a Hall Effect SYNC sensor mounted on the valve cover measuring the opening and closing of the intake valve as our CAM tooth.
- Running a pretty old MOTEC m400 ECU
- Running E85 fuel
- Custom intake, exhaust and loom

The engine successfully starts so there are no ref/sync issues at cranking. I got it to idle pretty well around 2500 RPM with a 0.95 lambda and no misfiring at this low rpm state. The issue is whenever the engine RPM increases to about 5000 RPM and above, the engine stutters all over the place, misfires and constantly loses REF/SYNC. Eventually, the engine will just die.



My first question is if a single cylinder is meant to lose REF/SYNC during operation at any point? I understand that a single cylinder is very unbalanced and it may be difficult for the ECU to continuously maintain synchronisation. In my case, it seems that every time it loses REF/SYNC the engine seems to suffer.
Secondly, looking at the engine log, the RPM is constant at idle, however any throttle input and the RPM starts spiking up and down, sometimes to 20,000 RPM and sometimes to -20,000 RPM which is obviously impossible. Thinking this may be due to noise, we tried to shield the ref and sync wires, however noticed that the 0V Eng B15 and B16 pins dedicated to ref and sync sensor grounds, and shielding, are both not showing continuity. It seems our ECU has a few blown internal fuses? At the moment everything is connected to essentially the same ground which I know is not the right convention. Using an oscilloscope, I couldn’t find any noticeable noise in either sensors. Is there a workaround for these pins? How do these ground pins differ to the common ground?

Thirdly, looking at the log, fuel injection duty seems to be really whacky, jumping to zero all the time. Fuel pressure is stable around 3 Bar so I’m pretty sure we’re getting fuel at all times so why is MOTEC telling the injector to shut off?

Lastly, could these issues possibly arise from simply a bad tune? Or do the current symptoms not make sense? I have spent quite a while playing around with the tune but seem to have no improvements in the higher RPM range.

Thanks so much for your help, if there is any info I have missed, please ask.

[David Ferguson]

You should run a Ref/Sync capture near the RPM where the problem occurs. Upload the .RSC2 file so we can take a look.

Tell us about the trigger wheel you machined. What is the size of the trigger teeth relative to the size of the pole element in the VR sensor?

I notice that your Ref Filter is set to maximum at all Engine Speeds, read the help (F1 key) for that table, you will find that you probably need a smaller filter value at higher RPM to avoid missing valid pulses (there is an example described of a 12-tooth wheel, take note). Again the tooth width is part of this equation, and a Ref/Sync Capture would answer a lot of questions.

Another unusual thing is to have a non-zero value for the Ref Trigger Voltage. Are you sure that is correct?