Mantis Bug Tracker

View Issue Details Jump to Notes ] Print ]
IDProjectCategoryView StatusDate SubmittedLast Update
0000753FirmwareGeneral Featurespublic2012-10-28 19:342013-06-12 17:41
Assigned ToFred 
PrioritynormalSeveritymajorReproducibilityhave not tried
PlatformOSOS Version
Product Version 
Target VersionFixed in Version 
Summary0000753: Engine Stalled and would not restart
DescriptionI was competing at an autocross last weekend.
The engine had been running all morning without any problems. I have my FreeEMS connected to a separate power switch and I had left it on all day. When I can back from lunch I started the car as normal and backed it up a few feet and it stalled and would not restart. I power cycled the FreeEMS and the car started again.

The engine stalled a number of time again. Always at idle or moving at slow speed while waiting in pit lane. When racing I did not detect any hesitation or miss fire.

I managed to capture a log of when it stalled and when I was trying to restart it before I power cycled it. After a power cycle, it always started up no problem.

I am using hash 9125f48f...

The only change to the code I made was setting my disable Threshold to RPM(6800) and reenable Threshold to RPM(6700)
TagsNo tags attached.
Issue TypeBug
Risk of Breakagemedium
Attached Filesgz file icon logs.tar.gz [^] (161,827 bytes) 2012-10-28 19:34

- Relationships
duplicate of 0000247assignedFred Make MAP sensor configuration both more robust and more flexible. 
has duplicate 0000665closedem_knaps Reset's while rpm signal is present cause non funtioning ign and inj outputs 

-  Notes
User avatar (0002370)
Fred (administrator)
2012-10-28 19:45

Possible duplicate of 0000665

Will review logs ASAP, which may not be for a week or two, sadly. If you can help diagnose, that would be great.
User avatar (0002625)
Fred (administrator)
2013-04-10 12:50

Spudmn, I had an idea about how to stop this affecting you, if it still is, if it has happened again:

 - Init a uint16 variable to 32768
 - In the scheduler increment each time you schedule a pin
 - In the output code decrement each time you fire a pin

If it ever gets higher or lower than X margin away from 32768, send special log packet and reset the device.

This will catch some firmware flaws, but not hardware issues where the output is being setup correctly and simply not physically firing. It won't catch the case where it's setup incorrectly, fires an interrupt, but doesn't toggle the pin.
User avatar (0002680)
Fred (administrator)
2013-06-12 16:41
edited on: 2013-06-12 16:42

Spudmn, you stated:

"I am using hash 9125f48f...

The only change to the code I made was setting my disable Threshold to RPM(6800) and reenable Threshold to RPM(6700)"

However without a non-dev commit hash in the log file there's nothing to say you didn't also change something else and simply forget. Specifically, the MAP sensor config. If you happened to be using a Honda 183kPa sensor as discussed and posted by you here: [^]

Then I know the cause of your issue and can resolve it with a one-line change. I don't know the cause of your CPU reset that likely triggered the bug, though. But that's a hardware issue anyway, and nothing I can do about it.

User avatar (0002681)
Fred (administrator)
2013-06-12 16:42

Reopen if freescale MAP sensor was in use.
User avatar (0002683)
Fred (administrator)
2013-06-12 17:35

Just checked the log files, guess what? Boost cut! This is finally closed for the right reason! Yay! I hate not understanding something...

Copyright © 2000 - 2011 MantisBT Group
Powered by Mantis Bugtracker