ManiacLachy wrote:But back up a bit, explain this:
Cus wrote:I ran my latest log (without adjustest AFR targets) over my latest MSQ (with the adjusted AFR targets)
I don't understand. You can compare a log to a tune?
You sure can!
Sorta...
Not really.
This is what I meant, the whole process, skipping exactly no steps.
_________________________________________________________________________________
Make sure the logging in TunerStudio (TS) is logging every available field, all of them.
Set your power management on your laptop to blank the screen instead of hibernating when you close the lid while on battery, plug it into the car, open TS, press CTRL-L ("Start logging" hotkey) give the logfile a name, put the laptop on the passenger seat, closed and upside down, with the USB pointed towards the front of the car. This bit is important, if you rail around a corner and the computer slides into something on the USB side, it'll break the USB port, so point it off the seat (guess how I know that!
). Upside down laptops have more efficient cooling than right-side-up laptops, and TS uses a bit of CPU power on my machine, so it gets hot.
Go driving for as long as the laptop battery will let you.
CTRL-B will stop logging - try and do this before the laptop goes flat. If the power management on your laptop isn't up to scratch (or the battery goes flat faster than your OS is anticipating, you may get a truncated logfile, which is all kinds of useless)
Open the log in MegaLogViewer (MLV) - if you have the licensed version, you have access to unlimited log sizes (40Mb takes a while to process though) if it's unregistered, you'll only get about 3 minutes worth of log analysis. (5000 records)
In MLV go to the View option, and make sure "Tuning console" is selected.
It should look like this:
Then, press "Open Tune" on the right, and find your CurrentTune.MSQ
Then press "VE Analyze" over on the right, and you'll get a screen like the following:
Lambda Signal Delay is a bit of voodoo I'm yet to fully jive with, but basically lower numbers tune high-RPM cells better, higher numbers are more accurate in the low-RPM cells. I usually leave it at 1 or 2 and don't worry about it.
Set your CLT filter to just above wherever your WUE turns off in TS. I don't think this is entirely required, but it motivates me to warm the car up first, and I'm more likely to be tuning against a heat-soaked car, which is how the car runs most of the time.
Once you're happy, press "Run Analysis" and you'll get something like this:
Press OK. If you put your mouse over the red and blue numbers on the left side, it'll tell you how many hits a cell had (how often the ECU was in that cell) and what the initial value was, so you can get an idea of how much of a change is about to be made. It's always goof to check this to make sure you're not about to use a tune where it's pulled 20% out of a cell or something silly. I've never see a change that big, but double-checking is alot cheaper than rebuilding an engine
Once you've had a look around the proposed new table, press Accept Table, then Save Tune. (or Save As)
If you have TS running in the background (I always do) it will ask you if you want to load the changes in CurrentTune.MSQ. Say yes. Then let it update on the car, then go way back to the start, and take a new log of some driving.
Once you've got a new log, go back to MLV, and load your log.
You'll want to be viewing RPM/VSS/MAP/TPS on one graph, and AFR/AFR Target/AFR Error in another graph. (I have Duty Cycle in there too, and you don't need VSS, but it makes it a boatload easier to find relevant parts of the logs for review)
So, in the above log you can see I'm cruising just below 100km/h, my AFR Target is 16.0, my
actual AFR is 15.2 and my AFR Error is -0.8 which means at this exact point in the log, I'm richer than I want to be. (you don't really need the error field, but it makes it easier)
That's the comparison I was talking about
From here, you can VE analyse again, and it'll be closer to the target (Mine is so far off because I still haven't dialled it right in after changing my AFR targets, I have a 40 minute log to run, but haven't)
The other point I was making, in a round-about fashion, is you only ever want to run VE Analyze on log files that were generated by the CurrentTune.MSQ that is running on the car.
Once you've done a VE Analyse and a "Save Tune" don't analyse any old logs against that tune (all logs are single-shot, basically) - if you've collected a bunch of logs from say, going to the shop, then some highway cruising, then maccas drivethrough, then some twisties, and they're all separate logs from the
same CurrentTune.MSQ you can hold down CTRL and click all of the relevant logs when you're opening the logs in MLV, it will smoosh them all together, and you can create a big log out of a few short ones. The more data you can analyse in one hit, the more accurate it tends to be.
But once you've done VE Analyse and saved the tune and put it on the ECU, you need to get new logs otherwise it'll be analyzing incorrect data, and you'll have an incorrect VE table, and that's how I ended up a full point of AFR leaner than my target.
