Have a look

[12sec.5.3]

I just made these logs and my VE is still a bit off. Maybe bc I was logging in the rain yesterday with zero traction. Just applied the new cells from my map to my VE without smoothing. Tell me what you think. I'll work with this a bit but i'm gonna redo my sd log on dry road i guess. My AFR is way far off.

[odd boy]

At least you need to do complete log of 2nd and 3rd gear, from lower rpm.

What was the actual AFR?

[odd boy]

The VE looks rich, I would recommend fixing the open loop fuel table with a constant no. like 12.5 or 13, since your car needs fuel. Make it constant from low map to 85.

At high maps 90+ use same no as in the PE table.

Don't forget the actual AFR.

[12sec.5.3]

I let off of it bc actual AFR was like 13.3. I never adjusted the OL fuel table so it's should be stock. I think it's like 14.63 . Should I put the whole table to 12.5 or just the high map cells and put lower ones at 13 or something?

[12sec.5.3]

You're talking about B3605 right?

[odd boy]

yes i do.

For SD tune, it is much better when you make that table constant and keep changing VE table.

Is the 13.3 AFR at WOT or PT?

[12sec.5.3]

I'm not running SD.

[odd boy]

I thought ur running SD.

Well, what is the actual AFR at cruising and WOT?

Post the stock file too, also you need to log more to have a better look at LTFT and STFT?

[12sec.5.3]

Nope no SD. I will log more though and post it for u to see

[12sec.5.3]

I'm following the AutoVE tutorial so that's why the logs are in SD. Need to get this VE squared away. Any advice is welcomed...

[12sec.5.3]

I was able to go drive a while a make a good log.

[12sec.5.3]

This is the VE after copying and pasting from the log. I didn't smooth or blend it at all.

[12sec.5.3]

Smoothed out the VE table and applied to my tune and plugged the maf back in. Check this log and see wt u think.

[joecar]

Something seems wrong with your wideband, see all the spikes when the motor appears to be in almost steady state (3rd pic).

1st and 2nd pics show wideband AFR instant drop causing the big spike in the VE peak region.

[12sec.5.3]

Yea I do have a lot of spikes. My wb always did fluctuate like crazy but I thought that was bc the tuning was off a bit. I stays a little more steady now but still has upward and downward spikes. Would could go wrong with that though?

[ScarabEpic22]

Definitely need to smooth those HUGE spikes out to get a sane value, what filters are you using on transient throttle conditions? And what is your cell count for those cells? Shoot for 50+ before you apply any correction in that area.

[joecar]

Those spikes in the wideband AFR are periodic... is the wideband timing out and reseting itself, or does it have an intermittent power connection...

Erik said you could probably filter those spikes out by adding some clauses to the ben transient filter.

[12sec.5.3]

I'd like more info on building this filter

[joecar]

I played with the filter but I couldn't get it to filter out the spikes (it can filter an up or down transition, but can't seem to filter spikes)...

So, try this, in calc_pid.txt call damp() before assigning the value to the afr...

post your calc_pids.txt here so I can edit it.

It would be better if you could find out where those spikes are coming from...

[12sec.5.3]

Here u go...

[Mr. P.]

I played with the filter but I couldn't get it to...
Agreed, the log file analysis functionality in the software is rough at best, there's a few things in the data that will totally skew the results of applying a simple filter. I'm experimenting with something... :shock: we'll see how well I do with it, but I am using a MIDMEAN averaging to dismiss the outliers.

Mr. P.

[joecar]

Here u go...
12sec.5.3, here's the edited calc_pids.txt...

copy it to the user configuration folder and view the log file again.

You're using the AEM 30-4100, right...? {EXT.AD1}*2 + 10.0[/COLOR] ]


[I]Edit: I updated the calc_pids.txt file...

Look at the pic, green is damped AFR (no spikes), red is undamped (spikes).

You could maybe try setting the damp parameter to 5 and see what it looks like.

:)

[joecar]

Agreed, the log file analysis functionality in the software is rough at best, there's a few things in the data that will totally skew the results of applying a simple filter. I'm experimenting with something... :shock: we'll see how well I do with it, but I am using a MIDMEAN averaging to dismiss the outliers.

Mr. P.Keep us posted...:cheers: