ok.....ive got some questions...

what exaclty does the Auto VE do?


will it take a datalog file, compare desired vs. actual AFR, do the math, and output a new VE table to approximate new pulsewidths?


ive used MLV to tune my buell and this is how it worked, and it worked fantastic.



if not, does the actual AFR need to be tweaked real time to match desired value? then saved as a new bin?


hope this makes sense.


ps- TEC3 on VW 1.8t & LC1 WB02. TEC is receiving good AFR signal from 02 and calibrated.

In WinTEC3, the AutoVE function takes either a datalog or live data from the unit and converts EGO corrections into changes for the VE table. The changes are recorded and your are given an option to save the changes to the current or a new file once you stop the log is done or you stop the process. If you choose to use a datalog, be sure to use a very long one with lots of driving conditions in it.

WinTEC4 does the same, but only works from live data.

well, i did a 30-45 minute log, loaded it in the autoVE but no values changed.

there are places where the afr is far from the desired table.

Good timing with this topic! I'll add my experiences as well (not trying to hijack):

My car runs great with only the MIT, POT and UAP numbers plugged in, but I know I have to tune the VE table to unleash the full potential. Today, I was trying the auto-tune feature today and have mixed feelings about it after seeing some of the results. I turned on EGO correction and left the default values in each entry box. I then entered the desired AFRs into the AFR table (13:1 at 100kpa to 14.6:1 at 55kpa). After driving around for a bit, the map was leaned out so much at low load/low RPM that the car wouldn't accelerate. It is most definitely going leaner than the desired 14.6:1 AFR I had entered. Should I use a larger sample size next time? Maybe my WB sensor is bad?

What method and EGO parameters do you all use with autotune?

Jon, Is there any future consideration for being able to use AutoVE with datalogs in WinTEC4? This would be very helpful for tuning at WOT when you can't spend much time inside a cell to auto calibrate.

It's on my list. Unfortunately, it's low on my list.

That said, there isn't really any difference between running AutoVE from live data vs. running from a datalog. The data is treated the same and you still have to hit a cell just as many times in a datalog as with live data to make a difference. In other words, you'll have to drive just as much datalogging as you would directly running AutoVE to get the same effect. The only advantage to a datalog is that you can use an internal log to do AutoVE. But that has limited usefulness because the internal logs are rarely long enough to have much of a useful effect. This is why AutoVE from a datalog was never implemented in WinTEC4.

Theoretically, wouldn't a short but structured run through the gears at WOT still provide a beneficial datalog? I realize the limitation of a short datalog, but I have found that a well thought out session will provide decent results. For example, when I want to adjust the VE table for WOT, I will datalog like I would on a chassis dyno. I'll start the log as soon as I short shift 4th gear and have my TPS at max voltage (WOT). I shut the log off after I'm well past my shift RPM. This yields consistent information between logs so I can make accurate VE changes. All this is a tedious manual process.

My desire would be to allow AutoVE in realtime or datalog to use less than 5 data points. It's very hard to tune in the WOT VE cells when I have to hit them 5 times to make a correction. Aside from the safety aspect of driving at WOT on the street, it's also dangerous on the engine if I'm mistakenly lean in a cell(s) and have to hit it 5 times for a correction. The only way to solve this is to guess what VE values would ensure all cells are overly rich. I would just like a more precise way of tuning at WOT and I think allowing less data points for a correction would help. Just my .02C

You can force your short data log to work but you will be circumventing the safety of the minimum sample size. Be warned. This method will fool AutoVE to act as if only one sample is needed to make a change by giving it the same data five times in a row. Some changes to the VE table will likely be bad. After doing this I always compare the new and old VE tables and modify everything that I think moved to far. I then carry the trends I see to neighboring cells.

Internally your log file is a text file. Open it in a text editor like notepad. Select and copy all the data, leaving the header info alone. Duplicate/paste it in five times and save. Then run it thru AutoVE normally.

Great idea, but I think this would only work with WinTEC3. WinTEC4 does not allow inputting a datalog in to AutoVE. I'm hoping for this functionality to be added someday.