J-series Acura/Honda

[Peter]

So I booted back into Ubuntu, and it won't build the code now. YAY for clock skew, and endless problems.

Jan 26 08:36:05 MST 2012"\" -D OPERATING_SYSTEM=\"Linux\" -D BASE_FILE_NAME=\"initialisers/TunableConfig2\" -E initialisers/TunableConfig2.c > preprocessedc/initialisers/TunableConfig2.pp.c
make: warning: Clock skew detected. Your build may be incomplete.

I'm not sure, but I think that means it's time for a different computer.

[Fred]

No, it probably means that ubuntu and winblows are fighting with the clock and the files were in the future.

cd out of src and rm -r src then git checkout src then try again. (EDIT: well, cp src src.new;rm -r src;mv src.new src might be better, if you have some changes)

windows keeps the clock in local time, *nix keeps it in GMT, I think, or something like that. Ubuntu could be fucking with it using NTP as could winblows.

Fred.

[Fred]

I just added the 6-2 stuff to the standard build and checked my notes on missing to actual ratios and it is a valid combo so should work, assuming minus 2 setups work at all, which I should get to testing here tomorrow, or maybe even tonight. At least this way you won't have as much hacking to do. If it turns out that you can make it work with even or no parity, then I'll migrate that hard coded config to configurable config and add a build key word for you to use so that your config is in the code base permanently. At your service! :-)

[Peter]

SUCCESS!!! It's giving me a good RPM reading at idle. I assume MegaTunix tell me if it loses sync? It seems to be working just fine with odd parity. Next step? Solder the wiring the harness into the sensors, and calibrate them. I assume I want to pull the 5V off my board not off of the OEM ecu. Documentation of my nonsense coming soon.

cd out of src and rm -r src then git checkout src then try again. (EDIT: well, cp src src.new;rm -r src;mv src.new src might be better, if you have some changes)

I didn't try this, but I'm running Ubuntu and Windows on separate hard drives, and I don't plug them both into the computer at the same time. So the sources and builds are both completely separate, and this looks like to that you're creating a new src for Ubuntu to use. At any rate I found the cheapest laptop I could, and put the Ubuntu hard drive in it. Now I don't have to run 60ft of usb cable from my desktop out to the car. Even though I did for this test, because I like sitting on the couch.

[Fred]

Congratulations, what changed? And where is my data log of it behaving itself?? This is enough to run, btw! Is the RPM correct? No one has tested the Cam missing tooth stuff or missing tooth = greater than 1 before, not even me. Hence keen for a datalog of it. You can enable bin logging in mtx if you cant make something else work. Long USB is usually not a good idea, but if it's working, what the hell, I guess.

What do you mean re 5v? The CPU (r5 mod) and sensors should share the same supply otherwise they'll be noisy and outright wrong.

MegaTunix has a counter for sync losses and status thingies that flash red when it loses sync and sync loss reason will go non-zero, etc. It's fairly comprehensive. I need to add some code in the next few days to make it a bit more safe, but it should be OK, already, at least :-)

Fred.

[Peter]

What changed? Hum? I'd have to say, winblows is no longer blowing all over the project. Sometimes the device manager would say COM10, and uBug12x would connect on COM2 and say unable to open 10. Advise to new users: Use Ubuntu! The RPM seems to be really close to what the tach is displaying.

Long USB is usually not a good idea, but if it's working, what the hell, I guess.

Lol I know. My other laptop has an old Pentium 2 that's so slow it pains me to use it. So I already had it all setup to connect to my desktop.
For the 5V I was thinking that they're already being powered by the oem ecu, and it would probably be a bad idea to leave them that way.
Megatunix crashes or something as soon as it's done saving the internal log. Kinda strange.

Video of MegaTunix: http://youtu.be/BdT660YIzPI
EDIT: Forgot the log

[Fred]

I have just realised that you can't run it in sequential yet due to the low tooth count. I will help you configure it in semi, though. The work to allow sequential with this simple setup is coming, but it's not super high priority. It should run fairly well with the exception of tip in of the throttle as is, though. You're going to be feature hungry, I can feel it. Please be patient with me :-)

[Fred]

I've not tested the CSV stuff in OLV for a while and it appears to be not working very well... I'll take a look tomorrow, or get Ben to. Sorry about this...

[Fred]

BTW, when I asked about RPM, I expected it to be right or out by a factor of 2 or something equally wildly wrong, not a bit wrong. It should be pretty exact at low revs and a bit too coarse at higher revs. Tested to 8k on my truck, though, so it doesn't seem to matter much. It's in the issue tracker too... :-)

[Fred]

First line of those logs has a bogus value in it, that causes OLV to choke. It's SO slow to load the bigger ones, though, that I'm going to gice up soon. The smaller one showed no losses of sync and something resembling good RPM, though. Cool :-)