MAF Scaling - Open/Closed Loop Fueling info

[Grip Ronin]

Originally Posted by vgi

from those errors (sorry before I was looking at my phone and couldn't see the msgs) that you did NOT select those column. In that version you must highlight (select) the column from the list.

ohhh i see what your saying now. i completely skipped over that part everytime.

[ML]

Originally Posted by Kodename47

Here's a nice easy spreadsheet that I threw together for those with very uneven MAF curves:
https://dl.dropboxusercontent.com/u/...omparison.xlsx


It's designed so that you can paste in your MAF curve and compare differing smoothing factors. It compares your curve against the OEM curve and the smoothing is based on that. It's a little different to the one applied in the java tool as it doesn't take into account your previous MAF scaling, just the original one. The idea is that if you've done multiple re-scales that you can see how it compares to stock.


You can input an auto smoothing function out of the 5 available or create your own curve by copy and paste by picking and choosing what you want.

Thanks for this, apparently just what I needed. My LTFT are steady around -1.6 and 0 now.

[steve99]

Originally Posted by Kodename47

Here's a nice easy spreadsheet that I threw together for those with very uneven MAF curves:
https://dl.dropboxusercontent.com/u/...omparison.xlsx


It's designed so that you can paste in your MAF curve and compare differing smoothing factors. It compares your curve against the OEM curve and the smoothing is based on that. It's a little different to the one applied in the java tool as it doesn't take into account your previous MAF scaling, just the original one. The idea is that if you've done multiple re-scales that you can see how it compares to stock.


You can input an auto smoothing function out of the 5 available or create your own curve by copy and paste by picking and choosing what you want.

I will link this to original post thanks

[solidONE]

Originally Posted by Kodename47

I can't work it out, it must be due to the behavior of air at those temperatures.

As for zeroing out the IAT comp, just do it while doing the MAF and then put them back as stock would be my recommendation.

Originally Posted by Kodename47

The IAT is a load compensation table, not MAF.

There are 2 ignition timing vs IAT tables, one that changes timing depending on IAT and the other determines the % of that applied depending on RPM and load.

Are you sure it's a load compensation table and not MAF g/s output compensation? Positively sure?

Here the CL error result of a MAF scale I dialed for 68 degrees ran at around 80 IAT's.

And here is the compensation table and part of this maf scale for reference:


As you can see The areas with 0 or the least amount of compensation remains the most accurate in the higher temps. The areas with more compensation resulted in error in fueling. Areas with + comp resulted in negative fuel trim, and the area with - comp ended up with positive fuel trim due to error.

I'm not typically this anal, but this is messing with my head....

It seems that the compensation should be lessened if not zeroed all together. Thoughts?

[Kodename47]

Originally Posted by solidONE

Are you sure it's a load compensation table and not MAF g/s output compensation? Positively sure?

It seems that the compensation should be lessened if not zeroed all together. Thoughts?

ECUtek doesn't clarify, it's under the Engine Load header so I assumed it was load calculation based. Their definition is: Compensation factor for Mass airflow against intake air temperature. So you could be right.


As I've said before, I would 0 it out for calibration purposes and then stick it back in. The table has obviously been set like that for a reason, likely to do with airflow behaviour at certain temperatures. I would try as hard as possible to calibrate the MAF as close to 20C as possible.

[solidONE]

Originally Posted by Kodename47

ECUtek doesn't clarify, it's under the Engine Load header so I assumed it was load calculation based. Their definition is: Compensation factor for Mass airflow against intake air temperature. So you could be right.


As I've said before, I would 0 it out for calibration purposes and then stick it back in. The table has obviously been set like that for a reason, likely to do with airflow behaviour at certain temperatures. I would try as hard as possible to calibrate the MAF as close to 20C as possible.

So I went ahead and used my 68* f scale and changed the temp values above 68 by adding 15* to each of the columns. (So 104 is now at 119*) and the results are looking very good. It may be a little off where the compensation goes from neg to positive. I believe that has a lot to do with how the maf scale was modified from the stock values throwing the compensation table off a bit.

All this logging and fiddling with the maf scale is finally gonna pay off. Fuk yeah!

[solidONE]

Here's my initial results with my MAF68 with adjusted MAF scale temp compensation table.

New adjusted values on temp compensation table



Results of initial CL log


Not sure if you can tell by looking at the little dots, but at above 1.84v or 11.20 g/rev, the fuel trim tend to go to negative and below those voltages the fuel trim tend to be more in the positive. This is exactly opposite of what I got running my MAF68 with the original temperature compensation table. This tells me that with the changes I made above, there isn't enough compensation as opposed to too much compensation with the stock compensation table. So I need to dial back the values a tiny bit to be closer to ideal. (maybe +10* above 68* columns instead of +15*)This log was taken with around 90*f IATs.

Also, not sure what's going on with that little lump at 2.88v or at about 49 g/rev. This little bump appears in both the CL and OL corrections.

[Turdinator]

I have been thinking about modifying Shiv's tune for doing MAF scale logs. Something like set the fuel maps to 12:1 everywhere and run only port or direct injection. I am sure I have seen this talked about somewhere but I can't find where. What should I be aware of with making this map? Will it be more prone to knock? Will I run out of injector flow?

[Kodename47]

Originally Posted by Turdinator

I have been thinking about modifying Shiv's tune for doing MAF scale logs. Something like set the fuel maps to 12:1 everywhere and run only port or direct injection. I am sure I have seen this talked about somewhere but I can't find where. What should I be aware of with making this map? Will it be more prone to knock? Will I run out of injector flow?

Look at your AF sensor scaling and choose a AFR value that is as close to one of the values as possible. The leaner the better without going too lean, you can run lean for a few pulls just make sure you reduce the timing.


Changing the PI and DI ratio, I do that purely in CL and leave the stock values above 5.2k for safety. There is a spreadsheet I posted up earlier in this thread that will help.

[steve99]

Originally Posted by Turdinator

I have been thinking about modifying Shiv's tune for doing MAF scale logs. Something like set the fuel maps to 12:1 everywhere and run only port or direct injection. I am sure I have seen this talked about somewhere but I can't find where. What should I be aware of with making this map? Will it be more prone to knock? Will I run out of injector flow?

In the first post of this thread there is link to jamesm screencast on injector scaling be worth a watch

[solidONE]

Just a little update on my temp comp tweaks, if anyone is interested.

Here is the OL errors in 68 IAT's. Not much correction going on as you can see. Aside from the strange lump at 2.88v.


Here is the same MAF scale with the adjusted temp comp with 90 IAT :


Leaned out in higher temps. Guess I got more tweaking to do. Need more + compensation at the higher MAF voltages. :/

[s2d4]

Originally Posted by solidONE

Here is the OL errors in 68 IAT's. Not much correction going on as you can see. Aside from the strange lump at 2.88v.

CL to OL transition, I've seen this on mine as well as a couple of others so far.
There are also instances of the car going slightly lean during the transition which may explain that lump.

[Kodename47]

Originally Posted by solidONE

Here is the OL errors in 68 IAT's. Not much correction going on as you can see. Aside from the strange lump at 2.88v.

What max/min error % are you using? As the rest of your curve is good and consistent you could probably use a value of 5-7.5. If you look down the run data, what was the max error at high rpm on the 90F run. Its likely that the value you're using is too big.

[solidONE]

Originally Posted by Kodename47

What max/min error % are you using? As the rest of your curve is good and consistent you could probably use a value of 5-7.5. If you look down the run data, what was the max error at high rpm on the 90F run. Its likely that the value you're using is too big.

I was using the default 10%. I will try again with 5%