Impregacy's Jaguar 0.6 build

[Impregacy]

Cool, thanks Fred.

All seems quiet on those fronts:



I guess the next step for me is to install a MAP sensor, 2 LS1 coils, and pray my trigger offset is close enough

[Fred]

+1, DO IT! It's been ages since any new engines ran... :-D I won't even help when i get #2 revived, either.

[Fred]

Also, push up your changes to github on a named branch and keep committing the changes as you go and before you build/try firmware. Good advice! :-) See this:

viewtopic.php?f=54&t=2323

The second and third to last items are key.

[Impregacy]

Okay Fred,

I will try to figure out that software development git voodoo, bare with me though, I'm a mechanical engineer who is winging it. Do you commit from elcipse--right clicking on the modified file?

[Fred]

Use a terminal.

git status < most important command, period. This tells you what you messed with. It tells you what's laying around unknown to git. It tells you what you have "staged" (told git you're pretty happy with and are nearly ready to commit). And it tells you more or less where you are in the commit tree.

git diff and/or git difftool (setup kdiff3 with prompt = false. These show you exactly what you did so you can make it clean and pretty and nice. Minimal change, only what you need to and should do. No extra white space or other cruft, for eg.

git add x y z < use this to stage files you're happy with after diffing them.

one git status shows all the things you want to commit staged and no other files changed and no other stuff laying around (use the tmp dir for this stuff if necessary) then you can simply do this:

git commit and/or git commit -m "omg my first commit evar" < this makes it "permanent" and gives it a cryptographic name.

Then you need to push it up, get github's help to help you with this.

Fred.

[Impregacy]

Thanks Fred, just in time for me to start configuring.

Installed a GM 1 bar MAP sensor and some truck coils.

[Fred]

Neat! :-) Make sure you use the latest dev branch commit as your base. And post youtube vids of your attempts to start and/or your startup :-D

[Impregacy]

Took a whack at the configuration code and pushed it to git. Always wanted to tell people I stayed up late coding--ever since that hacker movie came out in the 90s.


After rebuilding the carby and other things, I finally got the mower engine running right yesterday and can begin tinkering on a good base condition.

With the engine running and the Jaguar only reading RPM the signal was very dirty--it claims to retain sync now that I have a pot across VR+/- and adjusting it to about 7Kohm, but the error bars on the speed reading are about +/- 1000 rpm. I will post a screenshot of the log later.

Is there a way to slow down the code so that it is averaging more events --the average does looks about correct for the engine speed? What does the toothtolerance setting do?

[Fred]

Get a heavier flywheel! :-D Seriously, you're looking at instantaneous RPM. Averaging is NOT the solution. Someone incapable of designing an ECU thought it was, and implemented it, but it serves you no purpose because you're actually only using a subset of those numbers. IE, ignore it.

That setting helps with sync issues, if you have them during starting:

viewtopic.php?f=54&t=2369

During running sync issues are something you need to solve with wiring/topology. Post bz2 log files as well as/instead of screen shots. I like screen shots if they illustrate your issue, but it sounds like you have none?

Datalog stuff, I think I showed you this before: viewtopic.php?f=54&t=2323

Fred.

[Impregacy]

Okay it may not be an issue, dirty may have been the wrong word. Here is the screeny please let me know if it looks to be an issue:



My initial thinking was the rate at which the engine speed varies would confuse the EMS. I thought the frantic nature of the input would make the outputs frantic too so some smoothing would calm things down-(garbage in garbage out).

But I also probably don't have a grasp on how fast the maths are happening and maybe shouldn't be worried about it.

edit: I'm not sure it's the giant flywheel/cooling fan/driveshaft/alternator/trigger wheel monstrosity, but I have a theory that the unsupported length of the plastic crank position sensor may be allowing the sensor tip to flop around a bit caused by engine vibrations.